OPINION AND ORDER
SHIRA A. SCHEINDLIN, District Judge:I. INTRODUCTION 556
*554II. EXECUTIVE SUMMARY................................................557
III. APPLICABLE LAW.....................................................563
A. Monell Liability.....................................................563
B. Stops, Frisks, and Searches Under the Fourth Amendment..............565
1. The Definition of a Stop..........................................565
2. Stops Must Be Based on Reasonable Suspicion.....................567
3. Protective Frisks for Weapons....................................568
4. Searching into Clothing for Weapons..............................569
5. De Bour and the Fourth Amendment..............................569
C. Equal Protection Under the Fourteenth Amendment...................570
IV. FINDINGS OF FACT....................................................572
A. Overview of Uncontested Statistics....................................572
B. Expert Testimony...................................................576
1. The Liability Experts............................................576
2. The Fourth Amendment Claim....................................578
a. Overview of Key Issues.......................................578
b. Dr. Fagan’s Method of Classifying Stops.......................579
c. Unreliable Stop Factors......................................580
d. Quantifying the Magnitude of Apparently Unjustified Stops Based on UF-250 Stop Factors..............................582
3. The Fourteenth Amendment Claim................................583
a. Overview of Key Issues.......................................583
b. Competing Benchmarks................................. 583
c. Findings Based on Dr. Fagan’s Analyses.......................588
C. Institutional Evidence of Deliberate Indifference......................589
1. Early Notice: the 1999 AG Report.................................590
2. Pressure to Increase Stops........................................591
a. Compstat: Pressure on Commanders..........................592
b. Evidence of Pressure in Survey Data ..........................594
c. Further Evidence of Pressure on Officers......................596
i. Pressure Before the 2010 Quota Law ....................596
ii. Pressure After 2010 Quota Law .........................600
d. Conclusion..................................................602
3. Targeting “the Right People”.....................................602
4. Inadequate Monitoring and Supervision...........................607
a. Inadequate Documentation and Document Review..............607
b. Inadequate Supervision.......................................610
5. Partially Inadequate Training....................................613
6. Inadequate Discipline............................................617
7. Ongoing Notice of Constitutional Violations.......................620
D. Individual Stops.....................................................624
1. Unconstitutional Stop and Frisk..................................625
a. Leroy Downs................................................625
i. Findings of Fact.......................................625
ii. Mixed Findings of Fact and Law........................627
b. Devin Almonor ..............................................628
i. Findings of Fact.......................................628
ii. Mixed Findings of Fact and Law........................630
c. Cornelio McDonald..........................................630
i. Findings of Fact.......................................630
ii. Mixed Findings of Fact and Law........................632
d. Nicholas Peart — August 5, 2006 ...............................633
i. Findings of Fact.......................................633
ii. Mixed Findings of Fact and Law........................635
e. Nicholas Peart — April 13, 2011 Stop............................635
i. Findings of Fact.......................................636
*555ii. Mixed Findings of Fact and Law........................636
f. Ian Provost..................................................637
i. Findings of Fact.......................................637
ii. Mixed Findings of Fact and Law........................639
g. David Ourlicht — January 30, 2008 Stop.........................640
i. Findings of Fact.......................................640
ii. Mixed Findings of Fact and Law........................641
h. Clive Lino — February 5,2008 Stop.............................642
i. Findings of Fact.......................................642
ii. Mixed Findings of Fact and Law........................644
i. Lalit Clarkson...............................................645
i. Findings of Fact.......................................645
ii. Mixed Findings of Fact and Law........................646
2. Unconstitutional Frisk Only......................................646
a. Dominique Sindayiganza.....................................646
i. Findings of Fact.......................................646
ii. Mixed Findings of Law and Fact........................648
b. David Floyd — April 20, 2007 Stop..............................649
i. Findings of Fact.......................................649
ii. Mixed Finding of Fact and Law.........................650
c. David Floyd — February 27, 2008 Stop..........................650
i. Findings of Fact.......................................650
ii. Mixed Findings of Fact and Law............ 652
d. Clive Lino — February 24, 2011 Stop............................652
i. Findings of Fact.......................................652
ii. Mixed Findings of Fact and Law........................654
e. Deon Dennis.................................................655
i. Findings of Fact.......................................655
ii. Mixed Findings of Fact and Law........................656
3. Failure of Proof.................................................656
a. John Doe Stops of Nicholas Peart in Spring 2008 and February 2010 and David Ourlicht in February and June 2008.......................................................656
b. Kristianna Acevedo Stop......................................657
c. Clive Lino — August 3, 2008....................................658
V. CONCLUSIONS OF LAW................................................658
A. The City Is Liable for Violations of Plaintiffs’ Fourth Amendment Rights............................................................658
1. Deliberate Indifference...........................................658
2. Widespread Practice.............................................659
B. The City Is Liable for Violations of Plaintiffs’ Fourteenth Amendment Rights ......................................................660
1. Policy of Indirect Racial Profiling................................660
a. Intentionally Discriminatory Application of a Facially Neutral Policy.............................................661
b. Express Classification........................................663
c. Conclusion..................................................664
2. Deliberate Indifference...........................................665
VI. CONCLUSION..........................................................667
Courts can take no better measure to assure that laws will be just than to require that laws be equal in operation.
— Railway Express Agency v. People of State of New York, 336 U.S. 106, 112-13, 69 S.Ct. 463, 93 L.Ed. 533 (1949) (Jackson, J., concurring)
It is simply fantastic to urge that [a frisk] performed in public by a policeman while the citizen stands helpless, perhaps facing a wall with his hands raised>is a ^dignity.’
— Terry v. Ohio, 392 U.S. 1, 16-17, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968)
*556Whether you stand still or move, drive above, below, or at the speed limit, you will be described by the police as acting suspiciously should they wish to stop or arrest you. Such subjective, promiscuous appeals to an ineffable intuition should not be credited.
— United States v. Broomfield, 417 F.3d 654, 655 (7th Cir.2005) (Posner, J.)
I. INTRODUCTION
New Yorkers are rightly proud of their city and seek to make it as safe as the largest city in America can be. New Yorkers also treasure their liberty. Countless individuals have come to New York in pursuit of that liberty. The goals of liberty and safety may be in tension, but they can coexist — indeed the Constitution mandates it.
This case is about the tension between liberty and public safety in the use of a proactive policing tool called “stop and frisk.” The New York City Police Department (“NYPD”) made 4.4 million stops between January 2004 and June 2012. Over 80% of these 4.4 million stops were of blacks or Hispanics. In each of these stops a person’s life was interrupted. The person was detained and questioned, often on a public street. More than half of the time the police subjected the person to a frisk.
Plaintiffs — blacks and Hispanics who were stopped — argue that the NYPD’s use of stop and frisk violated their constitutional rights in two ways: (1) they were stopped without a legal basis in violation of the Fourth Amendment, and (2) they were targeted for stops because of their race in violation of the Fourteenth Amendment. Plaintiffs do not seek to end the use of stop and frisk. Rather, they argue that it must be reformed to comply with constitutional limits. Two such limits are paramount here: first, that all stops be based on “reasonable suspicion” as defined by the Supreme Court of the United States;1 and second, that stops be conducted in a racially neutral manner.2
I emphasize at the outset, as I have throughout the litigation, that this ease is not about the effectiveness of stop and frisk in deterring or combating crime. This Court’s mandate is solely to judge the constitutionality of police behavior, not its effectiveness as a law enforcement tool. Many police practices may be useful for fighting crime — -preventive detention or coerced confessions, for example — but because they are unconstitutional they cannot be used, no matter how effective. “The enshrinement of constitutional rights necessarily takes certain policy choices off the table.”3
This case is also not primarily about the nineteen individual stops that were the subject of testimony at trial.4 Rather, this case is about whether the City has a policy or custom of violating the Constitution by making unlawful stops and conducting unlawful frisks.5
The Supreme Court has recognized that “the degree of community resentment *557aroused by particular practices is clearly relevant to an assessment of the quality of the intrusion upon reasonable expectations of personal security.”6 In light of the very active and public debate on the issues addressed in this Opinion — and the passionate positions taken by both sides — it is important to recognize the human toll of unconstitutional stops. While it is true that any one stop is a limited intrusion in duration and deprivation of liberty, each stop is also a demeaning and humiliating experience. No one should live in fear of being stopped whenever he leaves his home to go about the activities of daily life. Those who are routinely subjected to stops are overwhelmingly people of color, and they are justifiably troubled to be singled out when many of them have done nothing to attract the unwanted attention. Some plaintiffs testified that stops make them feel unwelcome in some parts of the City, and distrustful of the police. This alienation cannot be good for the police, the community, or its leaders. Fostering trust and confidence between the police and the community would be an improvement for everyone.
Plaintiffs requested that this case be tried to the Court without a jury. Because plaintiffs seek only injunctive relief, not damages, the City had no right to demand a jury. As a result, I must both find the facts and articulate the governing law. I have endeavored to exercise my judgment faithfully and impartially in making my findings of fact and conclusions of law based on the nine-week trial held from March through May of this year.
I begin with an Executive Summary of the most important points in the Opinion. Next, I address the legal standards governing the ability of police to conduct stops and frisks. I provide a statistical overview of the 4.4 million stops made between January 2004 and June 2012, followed by a discussion of the expert analyses of those stops. I then address the question of whether the City had notice of allegations of racial profiling in the conduct of stops and frisks, and the institutional response to that notice in terms of monitoring, supervision, training, and discipline. After addressing these big picture issues, I make findings of fact with respect to each of the nineteen stops of the twelve class members who provided testimony at trial.
Finally, I present my conclusions of law based on my findings of fact. I will address the question of remedies in a separate opinion, because the remedies overlap with a different case involving stop and frisk in which I have already found that preliminary injunctive relief is warranted.7
It is important that this Opinion be read synergistically. Each section of the Opinion is only a piece of the overall picture. Some will quarrel with the findings in one section or another. But, when read as a whole, with an understanding of the interplay between each section, I hope that this Opinion will bring more clarity and less disagreement to this complex and sensitive issue.
II. EXECUTIVE SUMMARY
Plaintiffs assert that the City, and its agent the NYPD, violated both the Fourth Amendment and the Equal Protection Clause of the Fourteenth Amendment of the United States Constitution. In order to hold a municipality liable for the violation of a constitutional right, plaintiffs “must prove that ‘action pursuant to official municipal policy’ caused the alleged constitutional injury.”8 “Official municipal *558policy includes the decisions of a government’s lawmakers, the acts of its policy-making officials, and practices so persistent and widespread as to practically have the force of law.”9
The Fourth Amendment protects all individuals against unreasonable searches or seizures.10 The Supreme Court has held that the Fourth Amendment permits the police to “stop and briefly detain a person for investigative purposes if the officer has a reasonable suspicion supported by articulable facts that criminal activity ‘may be afoot,’ even if the officer lacks probable cause.”11 “Reasonable suspicion is an objective standard; hence, the subjective intentions or motives of the officer making the stop are irrelevant.” 12 The test for whether a stop has taken place in the context of a police encounter is whether a reasonable person would have felt free to terminate the encounter.13 “ ‘[T]o proceed from a stop to a frisk, the police officer must reasonably suspect that the person stopped is armed and dangerous.’ ”14
The Equal Protection Clause of the Fourteenth Amendment guarantees to every person the equal protection of the laws. It prohibits intentional discrimination based on race. Intentional discrimination can be proved in several ways, two of which are relevant here. A plaintiff can show: (1) that a facially neutral law or policy has been applied in an intentionally discriminatory manner; or (2) that a law or policy expressly classifies persons on the basis of race, and that the classification does not survive strict scrutiny. Because there is rarely direct proof of discriminatory intent, circumstantial evidence of such intent is permitted. “The impact of the official action — whether it bears more heavily on one race than another — may provide an important starting point.” 15
The following facts, discussed in greater detail below, are uncontested:16
• Between January 2004 and June 2012, the NYPD conducted over 4.4 million Terry stops.
• The number of stops per year rose sharply from 314,000 in 2004 to a high of 686,000 in 2011.
• 52% of all stops were followed by a protective frisk for weapons. A weapon was found after 1.5% of these frisks. In other words, in 98.5% of the 2.3 million frisks, no weapon was found.
• 8% of all stops led to a search into the stopped person’s clothing, ostensibly based on the officer feeling an object during the frisk that he suspected to be a weapon, or immediately perceived to be contraband other than a weapon. In 9% of these searches, the felt object was in fact a weapon. 91% of the time, it was not. In 14% of these searches, the felt object was in fact contraband. 86% of the time it was not.
• 6% of all stops resulted in an arrest, and 6% resulted in a summons. The *559remaining 88% of the 4.4 million stops resulted in no further law enforcement action.
• In 52% of the 4.4 million stops, the person stopped was black, in 31% the person was Hispanic, and in 10% the person was white.
• In 2010, New York City’s resident population was roughly 23% black, 29% Hispanic, and 33% white.
• In 23% of the stops of blacks, and 24% of the stops of Hispanics, the officer recorded using force. The number for whites was 17%.
• Weapons were seized in 1.0% of the stops of blacks, 1.1% of the stops of Hispanics, and 1.4% of the stops of whites.
• Contraband other than weapons was seized in 1.8% of the stops of blacks, 1.7% of the stops of Hispanics, and 2.3% of the stops of whites.
• Between 2004 and 2009, the percentage of stops where the officer failed to state a specific suspected crime rose from 1% to 36%.
Both parties provided extensive expert submissions and testimony that is also discussed in detail below.17 Based on that testimony and the uncontested facts, I have made the following findings with respect to the expert testimony.
With respect to plaintiffs’ Fourth Amendment claim,18 I begin by noting the inherent difficulty in making findings and conclusions regarding 4.4 million stops. Because it is impossible to individually analyze each of those stops, plaintiffs’ case was based on the imperfect information contained in the NYPD’s database of forms (“UF-250s”) that officers are required to prepare after each stop. The central flaws in this database all skew toward underestimating the number of unconstitutional stops that occur: the database is incomplete, in that officers do not prepare a UF-250 for every stop they make; it is one-sided, in that the UF-250 only records the officer’s version of the story; the UF-250 permits the officer to merely check a series of boxes, rather than requiring the officer to explain the basis for her suspicion; and many of the boxes on the form are inherently subjective and vague (such as “furtive movements”). Nonetheless, the analysis of the UF-250 database reveals that at least 200,000 stops were made without reasonable suspicion.
The actual number of stops lacking reasonable suspicion was likely far higher, based on the reasons stated above, and the following points: (1) Dr. Fagan was unnecessarily conservative in classifying stops as “apparently unjustified.” For example, a UF-250 on which the officer checked only Furtive Movements (used on roughly 42% of forms) and High Crime Area (used on roughly 55% of forms) is not classified as “apparently unjustified.” The same is true when only Furtive Movements and Suspicious Bulge (used on roughly 10% of forms) are checked. Finally, if an officer checked only the box marked “other” on either side of the form (used on roughly 26% of forms), Dr. Fagan categorized this as “ungeneralizable” rather than “apparently unjustified.” (2) Many UF-250s did not identify any suspected crime (36% of all UF-250s in 2009). (3) The rate of arrests arising from stops is low (roughly 6%), and the yield of seizures of guns or other contraband is even lower (roughly 0.1% and 1.8% respectively). (4) “Furtive Movements,” “High Crime Area,” and “Suspicious Bulge” are vague and subjective terms. Without an accompanying narrative explanation for the stop, these *560checkmarks cannot reliably demonstrate individualized reasonable suspicion.
With respect to plaintiffs’ Fourteenth Amendment claim,191 reject the testimony of the City’s experts that the race of crime suspects is the appropriate benchmark for measuring racial bias in stops. The City and its highest officials believe that blacks and Hispanics should be stopped at the same rate as their proportion of the local criminal suspect population. But this reasoning is flawed because the stopped population is overwhelmingly innocent — not criminal. There is no basis for assuming that an innocent population shares the same characteristics as the criminal suspect population in the same area. Instead, I conclude that the benchmark used by plaintiffs’ expert — a combination of local population demographics and local crime rates (to account for police deployment) is the most sensible.
Based on the expert testimony I find the following: (1) The NYPD carries out more stops where there are more black and Hispanic residents, even when other relevant variables are held constant. The racial composition of a precinct or census tract predicts the stop rate above and beyond the crime rate. (2) Blacks and Hispanics are more likely than whites to be stopped within precincts and census tracts, even after controlling for other relevant variables. This is so even in areas with low crime rates, racially heterogenous populations, or predominately white populations. (3) For the period 2004 through 2009, when any law enforcement action was taken following a stop, blacks were 30% more likely to be arrested (as opposed to receiving a summons) than whites, for the same suspected crime. (4) For the period 2004 through 2009, after controlling for suspected crime and precinct charaeteristics, blacks who were stopped were about 14% more likely — and Hispanics 9% more likely — than whites to be subjected to the use of force. (5) For the period 2004 through 2009, all else being equal, the odds of a stop resulting in any further enforcement action were 8% lower if the person stopped was black than if the person stopped was white. In addition, the greater the black population in a precinct, the less likely that a stop would result in a sanction. Together, these results show that blacks are likely targeted for stops based on a lesser degree of objectively founded suspicion than whites.
With respect to both the Fourth and Fourteenth Amendment claims, one way to prove that the City has a custom of conducting unconstitutional stops and frisks is to show that it acted with deliberate indifference to constitutional deprivations caused by its employees — here, the NYPD. The evidence at trial revealed significant evidence that the NYPD acted with deliberate indifference.20
As early as 1999, a report from New York’s Attorney General placed the City on notice that stops and frisks were being conducted in a racially skewed manner. Nothing was done in response. In the years following this report, pressure was placed on supervisors to increase the number of stops. Evidence at trial revealed that officers have been pressured to make a certain number of stops and risk negative consequences if they fail to achieve the goal.21 Without a system to ensure that stops are justified, such pressure is a predictable formula for producing unconstitutional stops. As one high ranking police official noted in 2010, this pressure, without a comparable emphasis on ensuring that the activities are legally justified, *561“could result in an officer taking enforcement action for the purpose of meeting a quota rather than because a violation of the law has occurred.”22
In addition, the evidence at trial revealed that the NYPD has an unwritten policy of targeting “the right people” for stops. In practice, the policy encourages the targeting of young black and Hispanic men based on their prevalence in local crime complaints.23 This is a form of racial profiling. While a person’s race may be important if it fits the description of a particular crime suspect, it is impermissible to subject all members of a racially defined group to heightened police enforcement because some members of that group are criminals. The Equal Protection Clause does not permit race-based suspicion.
Much evidence was introduced regarding inadequate monitoring and supervision of unconstitutional stops. Supervisors routinely review the productivity of officers, but do not review the facts of a stop to determine whether it was legally warranted. Nor do supervisors ensure that an officer has made a proper record of a stop so that it can be reviewed for constitutionality. Deficiencies were also shown in the training, of officers with respect to stop and frisk and in the disciplining of officers when they were found to have made a bad stop or frisk. Despite the mounting evidence that many bad stops were made, that officers failed to make adequate records of stops, and that discipline was spotty or non-existent, little has been done to improve the situation.
One example of poor training is particularly telling. Two officers testified to their understanding of the term “furtive movements.” One explained that “furtive movement is a very broad concept,” and could include a person “changing direction,” “walking in a certain way,” “[ajcting a little suspicious,” “making a movement that is not regular,” being “very fidgety,” “going in and out of his pocket,” “going in and out of a location,” “looking back and forth constantly,” “looking over their. shoulder,” “adjusting their hip or their belt,” “moving in and out of a car too quickly,” “[tjurning a part of their body away from you,” “[gjrabbing at a certain pocket or something at their waist,” “getting a little nervous, maybe shaking,” and “stutter[ing].”24 Another officer explained that “usually” a furtive movement is someone “hanging out in front of [a] building, sitting on the benches or something like that” and then making a “quick movement,” such as “bending down and quickly standing back up,” “going inside the lobby ... and then quickly coming back out,” or “all of a sudden becom[ing] very nervous, very aware.”25 If officers believe that the- behavior described above constitutes furtive movement that justifies a stop, then it is no surprise that stops so rarely produce evidence of criminal activity-
I now summarize my findings with respect to the individual stops that were the subject of testimony at trial.26 Twelve plaintiffs testified regarding nineteen stops. In twelve of those stops, both the plaintiffs and the officers testified. In seven stops no officer testified, either because the officers could not be identified or because the officers dispute that the stop ever occurred. I find that nine of the *562stops and frisks were unconstitutional— that is, they were not based on reasonable suspicion. I also find that while five other stops were constitutional, the frisks following those stops were unconstitutional. Finally, I find that plaintiffs have failed to prove an unconstitutional stop (or frisk) in five of the nineteen stops. The individual stop testimony corroborated much of the evidence about the NYPD’s policies and practices with respect to carrying out and monitoring stops and frisks.
In making these decisions I note that evaluating a stop in hindsight is an imperfect procedure. Because there is no contemporaneous recording of the stop (such as could be achieved through the use of a body-worn camera), I am relegated to finding facts based on the often conflicting testimony of eyewitnesses. This task is not easy, as every witness has an interest in the outcome of the case, which may consciously or unconsciously affect the veracity of his or her testimony. Nonetheless, a judge is tasked with making decisions and I judged the evidence of each stop to the best of my ability. I am also aware that a judge deciding whether a stop is constitutional, with the time to reflect and consider all of the evidence, is in a far different position than officers on the street who must make split-second decisions in situations that may pose a danger to themselves or others. I respect that police officers have chosen a profession of public service involving dangers and challenges with few parallels in civilian life.27
In conclusion, I find that the City is liable for violating plaintiffs’ Fourth and Fourteenth Amendment rights. The City acted with deliberate indifference toward the NYPD’s practice of making unconstitutional stops and conducting unconstitutional frisks. Even if the City had not been deliberately indifferent, the NYPD’s unconstitutional practices were sufficiently widespread as to have the force of law. In addition, the City adopted a policy of indirect racial profiling by targeting racially defined groups for stops based on local crime suspect data. This has resulted in the disproportionate and discriminatory stopping of blacks and Hispanics in violation of the Equal Protection Clause. Both statistical and anecdotal evidence showed that minorities are indeed treated differently than whites. For example, once a stop is made, blacks and Hispanics are more likely to be subjected to the use of force than whites, despite the fact that whites are more likely to be found with weapons or contraband. I also conclude that the City’s highest officials have turned a blind eye to the evidence that officers are conducting stops in a racially discriminatory manner. In their zeal to defend a policy that they believe to be effective, they have willfully ignored overwhelming proof that the policy of targeting “the right people” is racially discriminatory and therefore violates the United States Constitution. One NYPD official has even suggested that it is permissible to stop racially defined groups just to instill fear in them that they are subject to being stopped at any time for any reason — in the hope that this fear will deter them from carrying guns in the streets. The goal of deterring crime is laudable, but this method of doing so is unconstitutional.
I recognize that the police will deploy their limited resources to high crime areas. This benefits the communities where the *563need for policing is greatest. But the police are not permitted to target people for stops based on their race. Some may worry about the implications of this decision. They may wonder: if the police believe that a particular group of people is disproportionately responsible for crime in one area, why should the police not target that group with increased stops? Why should it matter if the group is defined in part by race?28 Indeed, there are contexts in which the Constitution permits considerations of race in law enforcement operations.29 What is clear, however, is that the Equal Protection Clause prohibits the practices described in this case. A police department may not target a racially defined group for stops in general — that is, for stops based on suspicions of general criminal wrongdoing — simply because members of that group appear frequently in the police department’s suspect data.30 The Equal Protection Clause does not permit the police to target a racially defined group as a whole because of the misdeeds of some of its members.
To address the violations that I have found, I shall order various remedies including, but not limited to, an immediate change to certain policies and activities of the NYPD, a trial program requiring the use of body-worn cameras in one precinct per borough, a community-based joint remedial process to be conducted by a court-appointed facilitator, and the appointment of an independent monitor to ensure that the NYPD’s conduct of stops and frisks is carried out in accordance with the Constitution and the principles enunciated in this Opinion, and to monitor the NYPD’s compliance with the ordered remedies.
III. APPLICABLE LAW
A. Monell Liability
Section 1983 of Title 42 of the United States Code (“section 1983”) creates “ ‘a species of tort liability’ ” for, among other things, certain violations of constitutional rights.31 As the Supreme Court established in Monell v. New York City Department of Social Services,32 in order to have recourse against a municipality or other local government under section 1983, plaintiffs “must prove that ‘action pursuant to official municipal policy’ caused the alleged constitutional inju*564ry.”33 “In other words, municipalities are ‘responsible only for their own illegal acts,’ and cannot be held ‘vicariously liable under § 1983 for their employees’ actions.’ ”34 In general, “[official municipal policy includes the decisions of a government’s lawmakers, the acts of its policy-making officials, and practices so persistent and widespread as to practically have the force of law.”35 Such policies “may be pronounced or tacit and reflected in either action or inaction.”36
One way to establish the existence of a municipal policy or custom is through a showing of “deliberate indifference” by high-level officials. “ ‘[W]here a policymaking official exhibits deliberate indifference to constitutional deprivations caused by subordinates, such that the official’s inaction constitutes a deliberate choice, that acquiescence may be properly thought of as a city policy or custom that is actionable under § 1983.’ ”37 Deliberate indifference requires “ ‘proof that a municipal actor disregarded a known or obvious consequence of his action.’ ”38 Recognizing that deliberate indifference is “a stringent standard of fault,” the Second Circuit requires “that the policymaker’s inaction was the result of ‘conscious choice’ and not ‘mere negligence.’ ”39
A municipality may incur Monell liability based on deliberate indifference through its training and supervision practices. “[Deliberate indifference may be inferred where ‘the need for more or better supervision to protect against constitutional violations was obvious,’ but the policymaker ‘fail[ed] to make meaningful efforts to address the risk of harm to plaintiffs[.]’ ”40 Although “[a] municipality’s culpability for a deprivation of rights is at its most tenuous where a claim turns on a failure to train,”41 the Supreme Court has held that “[w]hen city policymakers are on actual or constructive notice that a particular omission in their training program causes city employees to violate citizens’ constitutional rights, the city may be deemed deliberately indifferent if the policymakers choose to retain that program.” 42
In Walker v. City of New York, the Second Circuit framed the deliberate indifference inquiry in three parts:
(1) [the] policymaker knows “to a moral certainty” that its employees will con*565front a given situation; (2) either [the] situation presents employees with [a] difficult choice that will be made less so by training or supervision, or there is a record of employees mishandling [the] situation; and (3) [a] wrong choice by employees will frequently cause [the] deprivation of constitutional rights.43
“Where the plaintiff establishes all three elements, then ... the policymaker should have known that inadequate training or supervision was ‘so likely to result in the violation of constitutional rights, that the policymakers of the city can reasonably be said to have been deliberately indifferent to the need.’ ” 44 “A pattern of similar constitutional violations by untrained employees is ‘ordinarily necessary’ to demonstrate deliberate indifference for purposes of failure to train.”45
B. Stops, Frisks, and Searches Under the Fourth Amendment
The Fourth Amendment, made applicable to the States by the Fourteenth Amendment,46 states: “The right of the people to be secure in their persons, houses, papers, and effects, against unreasonable searches and seizures, shall not be violated____”47 The Supreme Court of the United States has repeatedly affirmed that “the ultimate touchstone of the Fourth Amendment is ‘reasonableness.’ ”48 The Supreme Court has held that under the Fourth Amendment, it is constitutionally reasonable for the police to “stop and briefly detain a person for investigative purposes if the officer has a reasonable suspicion supported by articulable facts that criminal activity ‘may be afoot,’ even if the officer lacks probable cause.”49 This form of investigative detention is now known as a Terry stop.50
1. The Definition of a Stop
As the Supreme Court reaffirmed in Florida v. Bostick, the test for determining whether a Terry stop is taking place “is whether a reasonable person would feel free to decline the officers’ requests or otherwise terminate the encounter.”51 Whether a stop has taken place depends on “whether, taking into account *566all of the circumstances surrounding the encounter, the police conduct would ‘have communicated to a reasonable person that he was not at liberty to ignore the police presence and go about his business.’ ”52
While the Supreme Court explicitly refrained from determining whether a seizure occurred in Bostick,53 it noted several types of police encounters that were not necessarily stops.54 However, the Court confirmed that even in these cases, the “free to terminate the encounter” standard applies: “[Ejven when officers have no basis for suspecting a particular individual, they may generally ask questions of that individual; ask to examine the individual’s identification; and request consent to search his or her luggage — as long as the police do not convey a message that compliance with'their requests is required.”55 The Bostick majority emphasized that police officers may not “demand of passengers their ‘voluntary’ cooperation” through “ ‘an intimidating show of authority.’ ”56
The Second Circuit has held that the following factors are indicative of a “seizure,” a term that encompasses both Terry stops and arrests:
the threatening presence of several officers; the display of a weapon; the physical touching of the person by the officer; language or tone indicating that compliance with the officer was compulsory; prolonged retention of a person’s personal effects, such as airplane tickets or identification; and a request by the officer to accompany him to the police station or a police room.57
*567The following summarizes two examples of police encounters that the Second Circuit held to be Terry stops, despite their arguably low level of coercion:
The Second Circuit has held ... that a stop took place where an officer twice ordered a person to “hold on a second,” and after the second order the person stopped. The Second Circuit also held that a stop occurred where an officer pointing a spotlight at a person said, “What, are you stupid? Come here. I want to talk to you,” and then told the person to show his hands.58
By contrast, the Second Circuit held that no Terry stop took place “where a person encountered two officers in his dorm lobby, and the officers asked him to show them his hands.”59
In sum, the test for whether a Terry stop has taken place in the context of a police encounter is whether a reasonable person would have felt free to terminate the encounter. The Second Circuit has further held: “[a] seizure occurs when (1) a person obeys a police officer’s order to stop or (2) a person that does not submit to an officer’s show of authority is physically restrained.”60
2. Stops Must Be Based on Reasonable Suspicion
In order for a Terry stop to comply with the Fourth Amendment, it must be based on a reasonable suspicion that criminal activity “may be afoot.”61 That is, the police may make a Terry stop “when the officer has reasonable, articulable suspicion that the person has been, is, or is about to be engaged in criminal activity.” 62 At minimum, “ ‘[t]he officer [making a Terry stop] ... must be able to articulate something more than an inchoate and unparticularized suspicion or hunch.’ ”63 That is, “[p]olice ‘must be able to point to specific and articulable facts which, taken together with rational inferences from those facts, reasonably warrant [the] intrusion [on a citizen’s liberty interest].’ ”64 “Reasonable suspicion is an objective standard; hence, the subjective intentions or motives of the officer making the stop are irrelevant.”65
*568In general, reasonable suspicion requires an individualized suspicion of wrongdoing.66 While the Supreme Court has recognized certain narrow exceptions to this requirement, there is no exception for stops of pedestrians for the general purpose of controlling crime.67
Courts reviewing stops for reasonable suspicion “must look at ‘the totality of the circumstances’ of each case to see whether the detaining officer has a ‘particularized and objective basis’ for suspecting legal wrongdoing.”68 “[T]he proper inquiry is not whether each fact considered in isolation denotes unlawful behavior, but whether all the facts taken together support a reasonable suspicion of wrongdoing.” 69 “While ‘reasonable suspicion’ is a less demanding standard than probable cause and requires a showing considerably less than preponderance of the evidence, the Fourth Amendment requires at least a minimal level of objective justification for making the stop.”70
3. Protective Frisks for Weapons
The Supreme Court has recognized that a police officer making an investigatory stop “should not be denied the opportunity to protect himself from attack by a hostile suspect.”71 As a result, “a law enforcement officer, for his own protection and safety, may conduct a patdown to find weapons that he reasonably believes or suspects are then in the possession of the person he has accosted.”72 “ ‘[T]o proceed from a stop to a frisk, the police officer must reasonably suspect that the person stopped is armed and dangerous.’ ”73 “The test is an objective rather than a subjective one, ... and thus it is not essential that the officer actually have been in fear.”74
*569“The purpose of [a frisk for weapons] is not to discover evidence of crime, but to allow the officer to pursue his investigation without fear of violence.”75 Thus, the frisk must be “limited in scope to this protective purpose,”76 and “strictly ‘limited to that which is necessary for the discovery of weapons which might be used to harm the officer or others nearby.’ ”77 However, when an officer “lawfully pats down a suspect’s outer clothing and feels an object whose contour or mass makes its identity [as contraband] immediately apparent,” the officer may seize the contraband without a warrant.78 In sum, “[n]othing in Terry can be understood to allow a generalized ‘cursory search for weapons’ or indeed, any search whatever for anything but weapons.”79
4. Searching into Clothing for Weapons
Just as reasonableness is the touchstone for the Fourth Amendment generally, reasonable suspicion provides the standard at each stage of a Terry stop. Once an officer has lawfully stopped someone based on reasonable suspicion of criminal activity, the officer may lawfully frisk the stopped person based on reasonable suspicion that the person is armed and dangerous. If the frisk gives rise to reasonable suspicion that an object in the clothing of the stopped person is a weapon that could be used to harm the officer, then the officer may take whatever action is necessary to examine the object and protect himself — including removing the object from the clothing of the stopped person.80
5. De Bour and the Fourth Amendment
The NYPD’s training materials place great importance on the New York state common law of stops, as articulated in People v. De Bour and its progeny.81 Because De Bour and the Fourth Amendment draw the line between permissible *570and impermissible police encounters in different ways, De Bour is in some respects more protective of liberty from governmental intrusion than the Fourth Amendment, and in other respects less.82 The Supreme Court has held that although states may impose greater restrictions on police conduct than those established by the Fourth Amendment, a state “may not authorize police conduct which trenches upon Fourth Amendment rights, regardless of the labels which it attaches to such conduct.”83 Thus, even where a police encounter would be permissible under De Bour, it remains unlawful if it violates the Fourth Amendment.
C. Equal Protection Under the Fourteenth Amendment
The Fourteenth Amendment’s Equal Protection Clause declares that “[n]o State shall ... deny to any person within its jurisdiction the equal protection of the laws.”84 The Clause “is essentially a direction that all persons similarly situated should be treated alike.”85 It prohibits intentional discrimination on the basis of race, but not government action that merely has a disproportionate racial impact.86
The Second Circuit has outlined “several ways for a plaintiff to plead intentional discrimination that violates the Equal Protection Clause.”87 First, “[a] plaintiff could point to a law or policy that ‘expressly classifies persons on the basis of race.’ ”88 Second, “a plaintiff could identify *571a facially neutral law or policy that has been applied in an intentionally discriminatory manner.”89 Third, “[a] plaintiff could also allege that a facially neutral statute or policy has an adverse effect and that it was motivated by discriminatory animus.”90 In none of these three cases is a plaintiff “obligated to show a better treated, similarly situated group of individuals of a different race in order to establish a claim of denial of equal protection.”91
In order to show intentional discrimination under the second and third models of pleading above, plaintiffs need not prove that the “ ‘challenged action rested solely on racially discriminatory purposes,’ ”92 or even that a discriminatory purpose “was the ‘dominant’ or ‘primary’ one.”93 Rather, plaintiffs must prove that “a discriminatory purpose has been a motivating factor” in the challenged action.94 That is, plaintiffs must show that those who carried out the challenged action “selected or reaffirmed a particular course of action at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon an identifiable group.”95 As the Supreme Court and the Second Circuit have explained:
Because discriminatory intent is rarely susceptible to direct proof, litigants may make “a sensitive inquiry into such circumstantial and direct evidence of intent as may be available. The impact of the official action — whether it bears more heavily on one race than another — may provide an important starting point.”96
The consequences of government action are sometimes evidence of the government’s intent: “proof of discriminatory intent must necessarily usually rely on objective factors.... The inquiry is practical. What a legislature or any official entity is ‘up to’ may be plain from the results its *572actions achieve, or the results they avoid.”97 “ ‘Once it is shown that a decision was motivated at least in part by a racially discriminatory purpose, the burden shifts to the defendant to show that the same result would have been reached even without consideration of race.’ ”98 “ ‘If the defendant comes forward with no such proof or if the trier of fact is unpersuaded that race did not contribute to the outcome of the decision, the equal protection claim is established.’ ”99
IV. FINDINGS OF FACT
A non-jury trial on liability and remedies was held between March 18 and May 20, 2013.100 Based on the preponderance of the credible evidence,101 as well as the parties’ post-trial submissions, the following are my findings of fact pursuant to Federal Rule of Civil Procedure 52(a).102
A. Overview of Uncontested Statistics
Officers are required to complete a UF-250 form, also known as a “Stop, Question and Frisk Report Worksheet,” after each Terry stop. Each side of the form contains checkboxes and fields in which officers are required to indicate the nature of the stop and the circumstances that led to and justified the stop (the “stop factors”). A copy of a blank UF-250 appears as Appendix A to this Opinion.103
Plaintiffs’ liability expert, Dr. Jeffrey Fagan, conducted various statistical analyses of UF-250s based on an electronic database containing the information on the forms.104 The more complicated and con*573tested statistical analyses will be discussed below.105 In this section, I summarize the most relevant uncontested statistics culled from the UF-250 database:106
• Between January 2004 and June 2012,107 the NYPD conducted over 4.4 million Terry stops.108
• The number of stops per year rose sharply from 314,000 in 2004 to a high of 686,000 in 2011.109
• 52% of all stops were followed by a protective frisk for weapons. A weapon was found after 1.5% of these frisks. In other words, in 98.5% of the 2.3 million frisks, no weapon was found.110
• 8% of all stops led to a search into the stopped person’s clothing, ostensibly based on the officer feeling an object during the frisk that he suspected to be a weapon, or immediately perceived to be contraband other than a weapon. In 9% of these searches, the felt object was in fact a weapon. 91% of the time, it was not. In 14% of these searches, the felt object was in fact contraband. 86% of the time it was not.111
• 6% of all stops resulted in an arrest, and 6% resulted in a summons. The remaining 88% of the 4.4 million stops resulted in no further law enforcement action.112
*574• In 52% of the 4.4 million stops, the person stopped was black.113
• In 31% of the stops, the person stopped was Hispanic.114
• In 10% of the stops, the person stopped was white.115
• In 2010, New York City’s resident population was roughly 23% black, 29% Hispanic, and 33% white.116
• In 23% of the stops of blacks, and 24% of the stops of Hispanics, the officer recorded using force. The number for whites was 17%.117
• Weapons were seized in 1.0% of the stops of blacks, 1.1% of the stops of Hispanics, and 1.4% of the stops of whites.118
• Contraband other than weapons was seized in 1.8% of the stops of blacks, 1.7% of the stops of Hispanics, and 2.3% of the stops of whites.119
• For the years 2004 to 2009,120 the two most commonly checked boxes indicating the reasons for a stop were “Furtive Movements” and “Area Has Incidence Of Reported Offense Of Type Under Investigation” (“High Crime Area”). Setting aside stops based on radio runs, officers marked “Furtive Movements” as a basis for the stop on 42% of the forms, and “High Crime Area” on 55% of the forms. In 2009, officers indicated “Furtive Movements” as a basis for the stop nearly 60% of the time.121
*575• Both “Furtive Movements” and “High Crime Area” are weak indicators of criminal activity. For the years 2004 to 2009, stops were 22% more likely to result in arrest if “High Crime Area” was not checked, and 18% more likely to result in arrest if “Furtive Movements” was not checked.122
• Between 2004 and 2009, as the number of stops per year soared from 314,000 to 576,000, the percentage of UF-250s on which the officer failed to state a specific suspected crime rose from 1% to 36%.123
Finally, I note that the City’s attempt to account for the low rate of arrests and summonses following stops was not persuasive. The City states that “[v]arious witnesses testified, including former Chief of Department Joseph Esposito, that many stops interrupt a crime from occurring, for example an individual casing a location or stalking an individual late at night.”124 No evidence was offered at trial, however, of a single stop that was: (1) based on reasonable suspicion, and (2) prevented the commission of a crime, but (3) did not result in probable cause for an arrest. While I have no doubt that such a stop has taken place at some time, it is highly implausible that successful “preventive” stops take place frequently enough to affect the conclusion that in at least 88% of the NYPD’s 4.4 million stops between January 2004 and June 2012, the suspicion giving rise to the stop turned out to be misplaced.
Indeed, for several reasons, the 12% “hit rate” likely overstates the percentage of stops in which an officer’s suspicions turn out to be well-founded. First, officers are trained to prepare UF-250s only for stops based on suspicion of a misdemeanor or felony. The UF-250 itself states: “Specify Which Felony/P.L. Misdemeanor Suspected.” 125 By contrast, a summons may be issued for offenses less serious than a misdemeanor, such as violations.126 Although the parties did not offer evidence on the types of summonses recorded in the UF-250 database, it is likely that many of these summonses were for violations rather than misdemeanors or felonies.127 In these cases, the issuance of the summons provides no evidence that the suspicion *576giving rise to the stop was well-founded, because if the officer was following NYPD procedures, the stop cannot have been initiated based on suspicion of the summonsed offense. Similarly, when a stopped person provides identification and is then arrested for an unrelated open warrant, the arrest does not prove that the suspicion leading to the stop was well-founded.128
Second, the fact that many post-stop summonses are dismissed further undermines the reliability of the 6% post-stop summons rate as a true “hit rate,” that is, a measure of validated suspicions.129 The same argument applies to post-stop arrests that were not charged.130
Third, both summonses and arrests may be unrelated to the suspected crime for which a person was stopped. For example, it has been reported that the most common arrest after a stop is for marijuana possession.131 The NYPD has recognized concerns that some marijuana arrests are based, improperly, on “occasions when the officers recover marihuana pursuant to a search of the subject’s person or upon direction of the subject to surrender the contents of his/her pockets.”132 If it is true that officers sometimes carry out arrests for marijuana possession following stops that were based on suspicion of another crime, then these arrests do not provide evidence that the officers’ initial suspicions were well-founded.133
B. Expert Testimony
Both parties offered expert testimony about whether the NYPD’s stop and frisk practices violate the Constitution. After describing the qualifications of the competing experts and discussing their differing views on the central issues in dispute here, I first determine which expert I find more reliable and the basis for that decision. I then make certain findings based on the credible expert testimony with respect to both the Fourth and Fourteenth Amendment claims.
1. The Liability Experts
Dr. Fagan is a Professor of Law at Columbia Law School and Professor of Epidemiology at the Mailman School of Public Health at Columbia University. He has been studying the policies at issue in this case for over a decade.134 Dr. Fagan’s *577honors, academic and professional appointments, and publications, make him an expert in criminology, with special expertise in the statistical study of racial disparities in police enforcement activities.135
The City’s liability experts are Dr. Dennis Smith, an Associate Professor of Public Administration at the Robert F. Wagner Graduate School of Public Service at New York University; and Dr. Robert Purtell, an Assistant Professor of Finance at the University of Albany’s Nelson A. Rockefeller College of Public Affairs and Policy.136 Dr. Smith has a Ph.D. in political science and is an expert at evaluating the effectiveness of police organizations. However, Dr. Smith is not a statistician. For this reason, Dr. Smith collaborated with Dr. Purtell, a statistical expert.137 Dr. Purtell has a BS in Mathematics, an MBA with an emphasis on finance and economics, and a Ph.D. in Public Administration; began his career as a research mathematician writing code used to run regression analyses; spent over thirty years working in finance and management; and now teaches finance.138 Dr. Purtell is not an expert in the study of policing, criminology, or racial discrimination.139
I find Dr. Fagan a more reliable expert than Drs. Smith and Purtell. While Dr. Smith’s research makes him specially qualified to opine on the effectiveness of the NYPD’s practices in controlling crime, the effectiveness of stop and frisk is not at issue in this case, as I have repeatedly explained. Unlike Dr. Fagan, Dr. Smith had never worked on a statistical study of racial disparities in any context until he became the City’s expert.
In addition, while both parties’ experts made errors in the course of their analyses that were later corrected,140 one error by Dr. Purtell called into question the general reliability of his interpretations of Dr. Fagan’s statistical analyses. Dr. Purtell conflated Tables 5 and 7 in Fagan’s Reports, which are at the center of Dr. Fagan’s conclusions regarding racial disparities in the NYPD’s stop practices. Table 5 deals with the effect of the racial composition of a geographic area on the number of stops that take place there — without reference to the race of the individuals being stopped. Table 7 deals with the races of individuals who are stopped.141 Yet, in his testimony, Dr. Purtell described how the numbers in Dr. Fagan’s Table 5 “are comparing the probability of a black person being stopped to the chances of a white person being stopped,”142 and persisted in
*578defending that analysis even after plaintiffs’ counsel explicitly pointed out the error.143
Finally, while the “battle of the experts” between Dr. Purtell and Dr. Fagan showed that Dr. Purtell has a sophisticated understanding of the purely mathematical aspects of statistics, Dr. Fagan has a deeper understanding of the practical, real-world meaning and implications of the statistical analyses in this case. Given a choice between relying on highly sophisticated mathematical analysis but limited practical understanding, or deep practical understanding informed by established statistical expertise, I favor the latter.
2. The Fourth Amendment Claim
a. Overview of Key Issues
Dr. Fagan performed an analysis of the NYPD’s UF-250 database in order to evaluate how often the NYPD’s stops lack reasonable suspicion. Before delving into Dr. Fagan’s Fourth Amendment analysis and my findings, I highlight several general points.
First, it is impossible to assess individually whether each of the 4.4 million stops at issue in this case was based on an officer’s reasonable articulable suspicion that criminal activity was afoot. It took weeks of testimony to try nineteen stops. It would take multiple lifetimes of many judges to try each of the 4.4 million stops.144 The best available information for assessing those stops comes from the UF-250s prepared by officers shortly after the stops.
Second, while the UF-250 database is the best available source of information, it is highly flawed for the following reasons: (1) Officers do not always prepare a UF-250, either because the officer does not believe she made a Terry stop or because the officer failed to prepare the form. (2) A UF-250 is one-sided, in that the UF-250 only records the officer’s version of the story. (3) Even NYPD commanders and supervisors have acknowledged that UF-250s do not provide enough information to determine whether reasonable suspicion existed for a stop.145 (4) Many of the checkboxes on the UF-250 that officers use to indicate the basis for a stop are problematic. “Furtive Movements” is vague and subjective. In fact, an officer’s impression of whether a movement was “furtive” may be affected by unconscious racial biases. “Fits Description” is a troubling basis for a stop if the description is so general that it fits a large portion of the population in the area, such as black males between the ages of 18 and 24. “High Crime Area” is also of questionable value when it encompasses a large area or an entire borough, such as Queens or Staten Island.
Third, Dr. Fagan was extremely conservative in characterizing stops as lacking reasonable suspicion. He categorized each stop as “apparently justified,” “apparently unjustified,” or “ungeneraliza*579ble.” The City argued that because Dr. Fagan characterized only 6% of the stops as “apparently unjustified,” that is, lacking reasonable suspicion, the plaintiffs have failed to demonstrate that the City has a policy or custom of carrying out stops without reasonable suspicion. However, in light of Dr. Fagan’s very generous assumptions in categorizing the stops, his analysis can best be understood as providing a very rough minimim number of unjustified stops. The actual number of unjustified stops was likely far higher. Moreover, even if I were to accept that Dr. Fagan’s 6% figure accurately reflects the number of stops lacking reasonable suspicion — which I do not for the reasons stated here and below — that relatively small percentage still represents 200,000 individuals who were stopped without reasonable suspicion. Even this number of wrongful stops produces a significant human toll.
b. Dr. Fagan’s Method of Classifying Stops
Dr. Fagan estimated the number of stops apparently lacking reasonable suspicion by analyzing the UF-250 database. He began with the assumption that all the forms had been filled out accurately and completely, then distinguished the stop factors on Side 1 of the UF-250 from the stop factors on Side 2.146 Dr. Fagan identified the following Side 1 boxes as providing a sufficient basis for a Terry stop, standing alone: (1) “Actions Indicative Of ‘Casing’ Victims Or Location” (“Casing”), (2) “Actions Indicative Of Engaging In Drug Transaction” (“Drug Transaction”), and (3) “Actions Indicative Of Engaging In Violent Crimes” (“Violent Crime”). Dr. Fagan defined the remaining Side 1 stop factors — except the “Other” box — as “conditionally justified,” that is, contributing to reasonable suspicion, but not generally providing an independently sufficient basis for a Terry stop: (4) “Carrying Objects In Plain View Used In Commission Of Crime e.g., Slim Jim/Pry Bar, etc.,” (5) “Suspicious Bulge/Object (Describe),” (6) “Actions Indicative Of Acting As A Lookout,” (7) “Fits Description,” (8) “Furtive Movements,” and (9) “Wearing Clothes/Disguises Commonly Used In Commission Of Crime.”147
Based on these classifications, Dr. Fagan categorized the stops recorded in the UF-250 database as (a) “apparently justified,” that is, based on reasonable suspicion; (b) “apparently unjustified,” that is, lacking reasonable suspicion; or (c) “ungeneralizable,” meaning that the UF-250 contains insufficient information to make a determination without further analysis.148 All stops in which officers checked the “Other” box are categorized as either ungeneralizable or apparently justified.149 A stop is “apparently unjustified” in Dr. Fagan’s analysis if: (a) no Side 1 stop circumstances are indicated, and only one Side 2 additional circumstance is indicated — unless the Side 2 additional circumstance is “Other (Describe),” in which case the stop is ungeneralizable; or (b) only one Side 1 stop circumstance is indicated, that stop circumstance is only “conditionally justified,” and no Side 2 additional circumstances are indicated — unless the Side 1 *580stop circumstance is the “Other” box, in which case the stop is ungeneralizable.150
c. Unreliable Stop Fáctors
Dr. Fagan thoroughly undermined the assumption that the two most frequently checked stop factors provide a reliable basis for suspecting criminality: Furtive Movements on Side 1, and High Crime Area on Side 2.151 Part of Dr. Fagan’s argument against the reliability of these factors rested on the uncontested statistics cited above,152 including that stops were more likely to result in arrest when Furtive Movements and High Crime Area were not checked than when they were. Courts have also recognized that furtive movements, standing alone, are a vague and unreliable indicator of criminality.153 As Judge Richard Posner has stated in a related context: “Whether you stand still or move, drive above, below, or at the speed limit, you will be described by the police as acting suspiciously should they wish to stop or arrest you. Such subjective, promiscuous appeals to an ineffable intuition should not be credited.”154 Recent psychological research has also provided evidence that officers may be more likely to perceive a movement as indicative of criminality if the officer has been primed to look for signs that “crime is afoot.”155 As I stated in Ligón, “[g]iven the nature of their work on patrol, officers may have a systematic tendency to see and report furtive movements where none objectively exist.” 156
Other recent psychological research has shown that unconscious racial bias continues to play an objectively measurable role in many people’s decision processes.157 It would not be surprising if many police officers share the latent biases that pervade our society.158 If so, such biases *581could provide a further source of unreliability in officers’ rapid, intuitive impressions of whether an individual’s movements are furtive and indicate criminality: Unconscious bias could help explain the otherwise puzzling fact that NYPD officers check “Furtive Movements” in 48% of the stops of blacks and 45% of the stops of Hispanics, but only 40% of the stops of whites.159 There is no evidence that black people’s movements are objectively more furtive than the movements of white people.
The High Crime Area stop factor is likewise problematic. Presence in an area with high rates of crime is not a sufficient basis for a stop, although it may contribute to reasonable suspicion.160 Plaintiffs offered evidence that the High Crime Area checkbox has been interpreted so broadly by at least some officers that it would contribute very little to the justification for a stop.161 In addition, Dr. Fagan has shown that the rate at which officers check High Crime Area in a precinct or census tract is roughly 55%, regardless of the amount of crime in the precinct or census tract as measured by crime complaints.162
Dr. Fagan also showed that over time, officers increasingly developed “scripts” for checking off stop factors.163 Not only did the average number of stop factors checked on UF-250s increase, but this increase reflected a growing use of several of the more subjective stop factors, such as Furtive Movements, Evasive Response, High Crime Area, and “Actions Indicative Of Engaging In Violent Crimes.”164 There was also credible evidence of script*582ing in the UF-250s of officers who testified at trial. During a sample quarter in 2009, Officer Edgar Gonzalez — who carried out a notably high number of stops — checked the same four boxes on 99% of his UF-250s: Fits Description, Casing, High Crime Area, and Time of Day.165 Officer Kha Dang — another aberrantly high stopper — checked an average of 4.2 boxes on the UF-250s he prepared during a sample quarter, and checked both High Crime Area and Time of Day on 75% of the forms, despite the stops being widely geographically and temporally dispersed.166
d. Quantifying the Magnitude of Apparently Unjustified Stops Based on UF-250 Stop Factors
Dr. Fagan’s extremely conservative definition of “apparently unjustified” almost guarantees that the roughly 200,000 stops he placed in that category underestimate the true number of stops lacking legal justification.167 For example, a UF-250 on which the officer checked only Furtive Movements on Side 1, and only High Crime Area on Side 2, is not classified as “apparently unjustified” according to Dr. Fagan’s definition. Similarly, a UF-250 on which the officer checked only Furtive Movements and Suspicious Bulge on Side 1, and no boxes on Side 2, is not classified as “apparently unjustified.” While some stops in which an officer checked only these factors might be based on reasonable suspicion, there is little doubt that many others would not be. In addition, any UF-250 on which an officer checked only the “Other” box on Side 1, or only the “other” additional circumstances box on Side 2, is not classified by Dr. Fagan as “apparently unjustified.”
The finding that Dr. Fagan’s stop factor analysis likely significantly undercounts the number of unjustified stops is corroborated by evidence in this case, as well as in Ligón, showing that officers sometimes fail to fill out UF-250s for stops thát lack reasonable suspicion, or fail to fill out UF-250s because they misunderstand when an encounter evolves into a Terry stop, or fill out UF-250s inaccurately and in a way that increases the apparent justification for a stop.168 In addition, several of the uncontested statistics suggest that far more than 6% of stops are “apparently unjustified,” including: the number of UF-250s that do not identify a suspected crime (36% of all forms in 2009), the fact that the two most commonly checked stop factors (Furtive Movements and High Crime *583Area) are negatively correlated with a summons or arrest, and the fact that only 12% of all stops result in an arrest or summons.169
The problems with Dr. Fagan’s Fourth Amendment analysis of the UF-250s result not from analytical failures but from the inadequacy of the NYPD’s systems for identifying unjustified stops when they occur. As a result, the magnitude of Fourth Amendment violations that have taken place in this case — beyond the rough minimum indicated by Dr. Fagan’s statistics— will almost certainly never be known.
3. The Fourteenth Amendment Claim
a. Overview of Key Issues
The crux of plaintiffs’ Fourteenth Amendment claim is that blacks and Hispanics are stopped more frequently than they would be if police officers did not discriminate based on race when deciding whom to stop. Assessing this claim required comparing statistics about rates of stops of blacks and Hispanics to “[a] standard, or point of reference, against which [those statistics] can be compared, assessed, measured or judged” — what is known in statistics as a “benchmark.”170 In this case, the benchmark was meant to capture “what the racial distribution of the stopped pedestrians would have been if officers’ stop decisions had been racially unbiased.”171
Conclusions regarding racial bias drawn from statistics “may vary drastically based on which benchmark is used.”172 As such, a central dispute between the experts regarding the Fourteenth Amendment claim was the appropriate benchmark for measuring racial bias in stops.
b. Competing Benchmarks
Each expert submitted voluminous reports and testified at trial in support of his choice of benchmark. Of necessity, I must simplify their very detailed and complex submissions and testimony to focus on the question at the heart of the parties’ dispute: is there statistical evidence of racial discrimination in the NYPD’s stop practices? With that caveat, I endeavor to summarize their differing benchmarks.
Dr. Fagan explained his choice of benchmark as follows:
[A] valid benchmark requires estimates of the supply of individuals of each racial or ethnic group who are engaged in the targeted behaviors and who are available to the police as potential targets for the exercise of their stop authority. Since police often target resources to the places where crime rates and risks are highest, and where populations are highest, some measure of population that is conditioned on crime rates is an optimal candidate for inclusion as a benchmark.173
Accordingly, Dr. Fagan’s “analyses use both population and reported crime as benchmarks for understanding the racial distribution of police-citizen contacts.”174 *584While there is scholarly disagreement regarding the best benchmark to use in such measurements, none of the sources Drs. Smith and Purtell cited criticized the benchmark used by Dr. Fagan. In addition, at least one other study of a police department’s stop patterns — -a study of stop patterns in Los Angeles by Dr. Ian Ayres, the William K. Townsend Professor of Law at Yale Law School — used an “[a]l-most identical” benchmark to Dr. Fagan’s.175
The City’s experts, by contrast, used a benchmark consisting of the rates at which various races appear in suspect descriptions from crime victims — in other words, “suspect race description data.”176 The City’s experts assumed that if officers’ stop decisions were racially unbiased, then the racial distribution of stopped pedestrians would be the same as the racial distribution of the criminal suspects in the area.177
I conclude that Dr. Fagan’s benchmark is the better choice. The reason is simple and reveals a serious flaw in the logic applied by the City’s experts: there is no basis for assuming that the racial distribution of stopped pedestrians will resemble the racial distribution of the local criminal population if the people stopped are not criminals. The City defends the fact that blacks and Hispanics represent 87% of the persons stopped in 2011 and 2012 by noting that “approximately 83% of all known crime suspects and approximately 90% of all violent crime suspects were Black and Hispanic.” 178 This might be a valid comparison if the people stopped were criminals, or if they were stopped based on fitting a specific suspect description. But there was insufficient evidence to support either conclusion. To the contrary, nearly 90% of the people stopped are released without the officer finding any basis for a summons or arrest,179 and only 13% of stops are based on fitting a specific suspect description.180 There is no reason to believe that the nearly 90% of people who are stopped and then subject to no further *585enforcement action are criminals. As a result, there is no reason to believe that their racial distribution should resemble that of the local criminal population, as opposed to that of the local population in general. If the police are stopping people in a race-neutral way, then the racial composition of innocent people stopped should more or less mirror the racial composition of the areas where they are stopped, all other things being equal. Dr. Fagan’s benchmark captures what the NYPD’s stops would look like in the absence of racial discrimination: his use of local population data reflects who is available to be stopped in an area (assuming, as the evidence shows, that the overwhelming majority of stops are not of criminals), and his use of local crime rates reflects the fact that stops are more likely to take place in areas with higher crime rates.
By contrast, Dr. Smith rejected the assumption that 88% of those stopped were innocent. “[H]ow do we know ... [i]f they were utterly innocent[?]” Dr. Smith asked at trial. He then proposed a “hypothetical” in which “the stop prevents a crime.”181 If one assumes that those stopped with no further enforcement action are nevertheless criminals, then it is natural to conclude, as Dr. Smith did, that a valid benchmark for measuring racial disparities in stops must “enable us to know who is committing the crime in [an] area.”182 Thus, he concludes that the best benchmark for the population of people who will be stopped in the absence of racial discrimination is the local criminal population. As Dr. Smith testified, “the best proxy for the share of the population by race engaged in the targeted behaviors that lead officers to make Terry stops” is the percentage of each racial category that appears in crime suspect data, or more precisely a combination of crime suspect data and arrestee data, because “[t]hat’s what we know about who is committing crime.”183
Based on this analysis, Dr. Smith concludes that the disproportionate stopping of black people can be explained by the disproportionately black composition of the pool of criminals.184 But even if all stops by the NYPD were based on reasonable suspicion — which is highly unlikely for reasons already stated — the low hit rate would undermine the assumption that the stopped people were in fact engaged in criminal activity, and thus members of the criminal population. The City failed to establish that a significant number of the approximately 3.9 million stops that resulted in no further enforcement action were stops of people who were about to commit, *586but were prevented from committing, a crime.185 Dr. Smith’s theory that a significant number of these stops resulted in the prevention of the suspected crime is pure speculation and not reliable.
Crime suspect data may serve as a reliable proxy for the pool of criminals exhibiting suspicious behavior. But there is no reason to believe that crime suspect data provides a reliable proxy for the pool of non-criminals exhibiting suspicious behavior. Because the overwhelming majority of people stopped fell into the latter category, there is no support for the City’s position that crime suspect data provides a reliable proxy for the pool of people exhibiting suspicious behavior. Moreover, given my finding that a significant number of stops were not based on reasonable suspicion — and thus were stops drawn from the pool of non criminals not exhibiting suspicious behavior — the use of crime suspect data as a benchmark for the pool of people that would have been stopped in the absence of racial bias is even less appropriate.186
When confronted by plaintiffs’ counsel with similar reasoning, Dr. Smith ultimately appeared willing to entertain the possibility that black people, even when they are law-abiding, might simply be more likely to engage in suspicious behavior than white people:
Q. So is it your testimony that law-abiding black people in New York City are more likely to engage in suspicious behavior than law-abiding white people?
A. I’m only saying that that’s the evidence from the stop patterns, which we have said, according to Professor Fagan, are ninety percent apparently justified.187
Dr. Smith’s position, while surprising, is not illogical once his premises are accepted. Dr. Smith apparently does not find it plausible that officers’ decisions regarding whether to stop a person may be swayed by conscious or unconscious racial bias.188 *587If a researcher begins with this premise, he will attempt to find a credible, race-neutral explanation for the NYPD’s stopping of blacks and Hispanics out of proportion to their share of the population. For example, the researcher may seek to explain the disproportionate stopping of minorities as the result of the characteristics of the criminal population. However, as already explained, there is no evidence that 88% of the people stopped are, in fact, members of the criminal population. Next, the researcher may analyze the deployment of police to high crime areas or “hot spots.” If these areas happen to be disproportionately minority, then heavy deployment to these areas will provide a race-neutral basis for the disproportionate stopping of minorities. But Dr. Fagan’s “Table 5” analysis showed that blacks and Hispanics are overstopped even after controlling for police deployment to high crime areas.189 In the end, if the researcher cannot think of any relevant race-neutral factors for which Dr. Fagan did not control, the only remaining race-neutral explanation for the NYPD’s stop patterns may be that members of the overstopped racial groups have a greater tendency to appear suspicious than members of other racial groups, even when they are not breaking the law.
Rather than being a defense against the charge of racial profiling, however, this reasoning is a defense of racial profiling. To say that black people in general are somehow more suspicious-looking, or criminal in appearance, than white people is not a race-neutral explanation for racial disparities in NYPD stops: it is itself a racially biased explanation. This explanation is especially troubling because it echoes the stereotype that black men are more likely to engage in criminal conduct than others.190 In a recent speech responding to the public controversy surrounding the shooting of a black teenager, President Obama noted his personal experience with this stereotype:
There are very few African-American men in this country who haven’t had the experience of being followed when they were shopping in a department store. That includes me. There are very few African-American men who haven’t had the experience of walking across the street and hearing the locks click on the doors of cars. That happens to me, at least before I was a senator. There are very few African-Americans who haven’t had the experience of getting on an elevator and a woman clutching her purse nervously and holding her breath until she had a chance to get off. That happens often.191
Another commentator observed in even starker terms:
What is reasonable to do, especially in the dark of night, is defined by preconceived social roles that paint young black men as potential criminals and *588predators. Black men, the narrative dictates, are dangerous, to be watched and put down at the first false move. This pain is one all black men know; putting away the tie you wear to the office means peeling off the assumption that you are owed equal respect. Mr. Martin’s hoodie struck the deepest chord because we know that daring to wear jeans and a hooded sweatshirt too often means that the police or other citizens are judged to be reasonable in fearing you.192
No doubt many people have heard similar fears and stereotypes expressed, whether intentionally or unintentionally. But race alone is not an objective basis for suspicion by the police. Because there is no evidence that law-abiding blacks or Hispanics are more likely to behave objectively more suspiciously than law-abiding whites, Dr. Smith’s — and the City’s — refuge in this unsupported notion is no refuge at all. It is effectively an admission that there is no explanation for the NYPD’s disproportionate stopping of blacks and Hispanics other than the NYPD’s stop practices having become infected, somewhere along the chain of command, by racial bias.
Why would the people stopped by the NYPD, both criminal and law-abiding, so closely resemble the criminal population- — ■ or, more precisely, the NYPD’s understanding of the criminal population, based on its limited suspect data? 193 A simple explanation exists: the racial composition of the people stopped by the NYPD resembles what the NYPD perceives to be the racial composition of the criminal population because that is why they were stopped. Evidence discussed later in this Opinion shows that the NYPD has an unwritten policy of targeting racially defined groups for stops, based on the appearance of members of those groups in crime suspect data.194 A strong correlation between the races of people stopped and the known races of criminal suspects is the natural result.
In short, the correlation highlighted by the City and its experts in their attempt to refute the allegation of racial profiling in fact provides evidence of racial profiling. Rather than revealing a valid race-neutral variable that explains the NYPD’s disproportionate stopping of blacks and Hispanics, the correlation highlighted by the City’s experts suggests how the racial disparities identified by Dr. Fagan might have come about — namely, through a widespread practice of racial profiling based on local criminal suspect data.
c. Findings Based on Dr. Fagan’s Analyses
Because I accept Dr. Fagan’s benchmark for measuring racial disparity and *589find his statistical analyses generally reliable, I make the following findings.
First, as reflected in Dr. Fagan’s Table 5, the NYPD carries out more stops in areas with more black and Hispanic residents, even when other relevant variables are held constant. The best predictor for the rate of stops in a geographic unit— be it precinct or census tract — is the racial composition of that unit rather than the known crime rate.195 These findings are “robust,” in the sense that the results persist even when the units of analysis are changed from precincts to census tracts, or from calendar quarters to months.
Second, as reflected in Dr. Fagan’s Table 7, within any area, regardless of its racial composition, blacks and Hispanics are more likely to be stopped than whites. This is different from the first finding — that the best predictor for the stop rate in a geographic area is the racial composition of that area. Table 7, by contrast, shows that blacks and Hispanics are more likely to be stopped than whites within precincts and census tracts, even after controlling for the racial composition, crime rate, patrol strength, and various socioeconomic characteristics of the precincts or census tracts where the stops take place. These findings are also robust. They apply not only when the spatial and temporal units of the analysis are changed, but also when the analysis is limited to areas with low crime rates, racially heterogenous populations, or predominately white populations.196
Third, for the period 2004 through 2009, blacks who were subject to law enforcement action following their stop were about 30% more likely than whites to be arrested (as opposed to receiving a summons) after a stop for the same suspected crime, even after controlling for other relevant variables.197
Fourth, for the period 2004 through 2009, after controlling for suspected crime and precinct characteristics, blacks who were stopped were about 14% more likely — and Hispanics 9% more likely — than whites to be subjected to the use of force.198
Fifth, for the period 2004 through 2009, all else being equal, the odds of a stop resulting in any further enforcement action were 8% lower if the person stopped was black than if the person stopped was white. In addition, the greater the black population in a precinct, the less likely that a stop would result in a sanction. These results show that blacks are likely targeted for stops based on a lesser degree of objectively founded suspicion than whites.199
C. Institutional Evidence of Deliberate Indifference
The previous two sections addressed the statistical evidence of unconstitutional stops. This section addresses the evidence *590regarding the NYPD’s awareness of and response to those unconstitutional stops. In short, I find that the “institutional evidence” — evidence regarding the actions or inactions of the NYPD — shows that the City has been deliberately indifferent to violations of the plaintiff class’s Fourth and Fourteenth Amendment rights.
The NYPD has known for more than a decade that its officers were conducting unjustified stops and frisks and were disproportionately stopping blacks and Hispanics. Despite this notice, the NYPD expanded its use of stop and frisk by seven-fold between 2002 and 2011. This increase was achieved by pressuring commanders at Compstat meetings to increase the numbers of stops. The commanders, in turn, pressured mid-level managers and line officers to increase stop activity by rewarding high stoppers and denigrating or punishing those with lower numbers of stops.
This pressure to increase the quantity of stops was not accompanied by attention to the constitutionality of the stops. No policies were implemented to ensure that officers were recording each stop with sufficient detail to permit an assessment of the constitutionality of the stops. Similarly, no study was done to ensure that officers were not reflexively creating a “script” of checkmarks — especially Furtive Movements and High Crime Area — by searching the OF-250 database to identify such patterns. No effort was made to identify outliers — meaning those officers with the highest stop numbers, officers who stopped only or almost only blacks or Hispanics, or officers who routinely checked the same boxes on the OF-250. No rewards or punishments turned on the quality of stops conducted. Indeed, when officers were found to have made “bad” stops, little or no discipline was imposed. The evidence showed that the NYPD turned a blind eye to its duty to monitor and supervise the constitutionality of the stops and frisks conducted by its officers.
In addition, I find that the NYPD instituted a policy of indirect racial profiling by directing its commanders and officers to focus their stop activity on “the right people” — the demographic groups that appear most often in a precinct’s crime complaints. This policy led inevitably to impermissibly targeting blacks and Hispanics for stops and frisks at a higher rate than similarly situated whites.
1. Early Notice: the 1999 AG Report
In 1999, New York’s Attorney General investigated the constitutionality of the NYPD’s stop and frisk practices under the Fourth and Fourteenth Amendments. The investigation was prompted in part by the Attorney General’s finding that despite a decade of falling crime rates, “the climate in many of New York’s minority neighborhoods ... was one of resentment and distrust of the NYPD.” Many of the complaints involved “lower-level police involvement in the everyday lives of minority residents,” such as stop and frisk encounters.200
The Attorney General sought the assistance of a team of researchers from Columbia University’s Center for Violence Research and Prevention, led by Dr. Fagan. The researchers performed statisti*591cal analyses of 175,000 UF-250s from January 1,1998 through March 31, 1999. The resulting Report was apparently the first-ever quantitative analysis of pedestrian stop and frisk practices in the United States.201
For their Fourth Amendment investigation, the researchers analyzed and grouped 15,000 UF-250s using a methodology that resolved “every ambiguity of factual or legal interpretation ... in favor of a determination that ‘reasonable suspicion’ existed.” Nevertheless, the researchers found that 15% of the UF-250s contained facts that did not meet the legal test for reasonable suspicion.202
For their Fourteenth Amendment investigation, the researchers tested thenNYPD Commissioner Howard Safir’s theory — which largely remains the City’s theory in this case — that the apparently disproportionate stopping of blacks and Hispanics can be explained on race-neutral grounds by police deployment to high crime areas, and by racial differences in crime rates.203 The researchers found, however, that “blacks and Hispanics were significantly more likely than whites to be ‘stopped’ [even] after controlling for race-specific precinct crime rates and precinct population composition by race.” 204 In addition, the Report found that different crime rates among precincts did not explain the higher overall stop rate in majority-minority precincts as opposed to majority-white precincts.205
The Report called for a broad, public dialogue among the Office of the Attorney General, the NYPD, and members of the community.206 This dialogue never occurred. Instead, senior officials at the NYPD either found pretexts for rejecting the Report’s findings,207 or ignored the Report entirely — like Chief Esposito, who testified that he had never read it.208
2. Pressure to Increase Stops
Between 2002 and 2011, the number of stops increased from roughly 97,000 to *592roughly 686,000 per year.209 How did the NYPD increase its stop activity by roughly 700%, despite the fact that crime continued to fall during this period?210
Based on numerous, mutually reinforcing sources of evidence at trial including live testimony, depositions, roll call recordings, internal NYPD documents, and survey results, the most plausible explanation is that NYPD officers prior to and during the class period experienced significant pressure to increase their stop activity.211
a. Compstat: Pressure on Commanders
Introduced in 1994, Compstat is the NYPD’s statistics-based performance management system.212 The system collects and analyzes statistical and other data about local crime and enforcement activities, conducts weekly meetings during which senior officials question local commanders about the data, and holds commanders accountable for addressing crime conditions and improving the quantitative measures of their performance.
Chief Esposito, who chaired Compstat’s meetings until his retirement earlier this year, and Assistant Chief Raymond Diaz testified that the amount of UF-250s that a unit has completed is a factor in evaluating the unit’s performance.213 Excerpts from heavily redacted minutes of Comps-tat meetings show Chief Esposito questioning commanders about their low stop numbers.214 Assistant Chief Diaz also *593stated that one way of measuring the effectiveness of impact response teams is to look at the number of UF-250s they generate 215 and that “an increase in the 250s is usually a good sign ... that the unit that is being reviewed is engaging in more activity as opposed to less.”216
Chief Esposito and other NYPD officials testified that the quality of UF-250s is also reviewed at Compstat meetings. Indeed, there was evidence that attention is paid at Compstat meetings to the quality of enforcement activity in the sense of its effectiveness.217 For example, Chief Esposito often questions commanders at Compstat about whether enforcement activity was responding to crime conditions in specific places and times.218 There was no evidence, however, that the quality of stops in the sense of their constitutionality receives meaningful review or plays a role in the evaluation of commanders’ performance during Compstat meetings.
Several NYPD officials conceded in testimony that Compstat focuses on effectiveness, not constitutionality. For example, Chief Esposito was asked to explain an excerpt from the Compstat meeting notes in which he is recorded as stating:
Quality on 250s[,] forget the number. 5% enforcement rate off 250s, 102 [Precinct] is the worst with enforcement] off 250s. A lot of it is probably training. But quality of 250 in [Queens] South has a lot to be desired.219
When asked by plaintiffs’ lawyers whether “quality” in this passage could refer to whether stops were based on reasonable suspicion, Chief Esposito stated: “No. I think we talk more about where and when. Does it match up with the crime picture? That’s what is paramount.” 220 None of the excerpts from the Compstat meeting notes regarding UF-250s include a discussion of racial profiling or use the term reasonable suspicion.221
*594Similarly, to the extent that Chief of Patrol James Hall and his staff “raise issues or concerns about the UF-250s with COs at the meetings,” 222 these relate to the effectiveness of stops and officers’ basic compliance with paperwork requirements.223 There was no credible evidence that Chief Hall or his staff perform regular or meaningful reviews of the constitutionality of stops before Compstat meetings.224
In sum, Compstat exists to measure the effectiveness of police enforcement activities, not their constitutionality.
b. Evidence of Pressure in Survey Data
The evidence discussed in the previous section shows that senior NYPD officials at Compstat meetings routinely place pressure on commanders to increase their enforcement activity, including their stop numbers. The survey evidence in this section shows that subordinate managers in the NYPD have communicated this pressure to the rank and file.
Dr. Eli Silverman and Dr. John Eterno, a retired NYPD captain, conducted two surveys of retired members of the NYPD, one in 2008 and the other in 2012.225 The 2008 survey was sent to 1,197 retired NYPD personnel with the rank of captain or above. 41% responded. The 2012 survey was sent to 4,069 retired NYPD personnel of all ranks who had listed themselves as “active retirees,” that is, available if needed to serve in an emergency. 48% responded.226
The 2008 survey asked: “With respect to the following criteria and based on your personal experience, on a scale of 1 to 10 (with 1 being the least and 10 the most), how much pressure was there from man*595agement/supervisors to ... ” — followed by a list including, among other -items, “Increase summonses,” “Increase arrests,” and “Increase Stop and Frisk Reports.” The final page of the survey asked: “Did you serve on [the] NYPD after 1994?”227 As noted above, 1994 was the year in which Compstat was introduced.
The responses to the 2008 survey show that mid-level managers in the NYPD who served during the Compstat era perceived significantly greater pressure to increase stops, arrests, and summons than those who retired prior to 1994. Specifically, when asked to rate the amount of pressure they perceived to increase stops, the mean response of Compstat era personnel was a statistically significant 2.2 points higher on the scale from 1 to 10 than pre-Compstat era personnel.228 In addition, between the pre-Compstat and Compstat eras, the percentage of NYPD personnel who reported low pressure (1 to 3 on the scale) fell from 55% to 23%, while the percentage who reported high pressure (8 to 10 on the scale) rose from 5% to 28%.229
The 2012 survey began by stating: “For all questions, if you retired before 1994, base your answers on your overall impressions over your entire career; if you retired in 1994 or after, base your answers on experiences that occurred only in 1994 and after.” The second set of questions presented a refinement of the question on the first page of the 2008 survey: “With respect to the following criteria and based on your personal experienee/knowledge, on a scale of 1 to 10 (with 1 being the least and 10 the most), how much pressure did precinct (patrol) personnel receive from management/supervisors to ... ”■ — -followed by a list including “Increase summonses,” “Increase stop and frisk,” and “Increase arrests,” as well as a new item: “Obey Legal/Constitutional Rules.” The survey later asked respondents to state the year in which they retired.230
Dr. Silverman divided the respondents to the 2012 survey into three categories: those who retired in 1994 or earlier (preCompstat), those who retired between 1995 and 2001, and those who retired after 2002 (the year of Mayor Bloomberg’s arrival and his appointment of Commissioner Kelly). The responses showed that the percentage of NYPD personnel reporting high pressure to increase stops increased from 9% in the pre-Compstat era, to 19% in the Compstat era before Mayor Bloom-berg, to 35% of all respondents who retired after 2002. The percentage of personnel reporting low pressure fell from 58% for pre-Compstat retirees to 37% for pre-Bloomberg retirees to 24% for postBloomberg retirees. The increases in. reported pressure to raise summons and arrest numbers was similarly stark.231
Finally, while the 2012 survey showed an increase from 35% to 42% in those reporting medium pressure to obey legal and constitutional rules, it also showed a significant post-2002 decrease in those reporting high pressure to do so. 45% of pre*596Compstat retirees and 47% of early Compstat era retirees reported high pressure to obey legal and constitutional rules, while only 36% of post-Bloomberg retirees reported high pressure. Dr. Silverman testified that this represented a modest but statistically significant decline.232
Although the City attempted to undercut the reliability of the 2008 and 2012 findings,233 I find that the City’s criticisms do not undermine the surveys’ central finding for the purposes of this case: NYPD personnel experienced or were aware of pressure to increase the number of stops after the introduction of Comps-tat, and especially after the arrival of May- or Bloomberg and Commissioner Kelly. In addition, this rising pressure for stop numbers was not accompanied by equivalent pressure to obey constitutional restrictions.
c. Further Evidence of Pressure on Officers
Additional anecdotal evidence supports plaintiffs’ argument that the NYPD pressured officers to increase stops without due regard to the constitutionality of those stops. For convenience, I divide the evidence into two periods: before the enactment of New York’s Quota Law in 2010, and after.
i. Pressure Before the 2010 Quota Law
Before 2010, the NYPD had no written policy prohibiting quotas for stops, arrests, or other enforcement activities.234 There is abundant evidence during this period of supervisors directing officers to meet numerical enforcement goals, as well as threatening the officers with negative consequences if they did not achieve those goals. In particular, three NYPD officers from three precincts made secret recordings revealing institutional pressure to increase enforcement numbers: Officers Adrian Schoolcraft, Adhyl Polanco, and Pedro Serrano.235 The three officers’ recordings provide a rare window into how the NYPD’s policies are actually carried out. I give great weight to the contents of these recordings.
Officer Schoolcraft’s recordings take place at the 81st Precinct in the Bedford Stuyvesant area of Brooklyn. Many of the recordings are of roll calls, the period before a tour when officers assemble to receive assignments and training.236 Requests or commands to issue UF-250s are common on the recordings, though sometimes innocuous.237 The recordings also show repeated calls for increased “activity,” including summonses, stops, and arrests. Sometimes supervisors use numerical goals backed by the threat of negative *597consequences if the goals are not met. For example, at a June 12, 2008 roll call, Lieutenant Jean Delafuente stated:
[W]e had the CO’s meeting today.... First and foremost, we need more activities, all right? The CO wants more activity. The XO wants more activity. The borough is monitoring the activity sheets. So, if your activity falls below par, they’re going to have either you or I or the Sergeant or the CO have to explain what’s going on, all right? So, let’s not let it get that far, all right?238
Later, Lieutenant Delafuente states: “The XO was in the other day.... He actually laid down a number, all right?” Lieutenant Delafuente says, perhaps jokingly, that he is not going to quote the number, then proceeds to say that the Executive Officer “wants at least three seatbelts, one cellphone, and 11 others.” He also suggests that he has criticized officers whose numbers were not high enough: “The CO gave me some names. I spoke to you. I’m not going to embarrass you in front of everyone.” 239
The most striking aspect of the School-craft recordings is the contempt and hostility of supervisors toward the local population. For example, at a roll call on November 8, 2008, Lieutenant Delafuente states:
All right, I went out there [to Howard and Chauncey] yesterday and ... we’ve got the old man out there with the grey hairs. A loud mouth. He thinks since he’s 55 years old he’s not going to get locked up. Well, guess what? I don’t tolerate shit out there. He went in and two of his pals went in. All right? So we’ve got to keep the corner clear.... Because if you get too big of a crowd there, you know, ... they’re going to think that they own the block. We own the block. They don’t own the block, all right? They might live there but we own the block. All right? We own the streets here. You tell them what to do.240
Similarly, Lieutenant Delafuente reminds the officers at a roll call on November 1, 2008 that they are “not working in Midtown Manhattan where people are walking around smiling and happy. You’re working in Bed-Stuy where everyone’s probably got a warrant.”241 Because Bedford Stuyvesant is a historically black neighborhood and continues to have a majority black population,242 Lieutenant Delafuente’s comment carries troubling racial overtones.
As further evidence of a culture of hostility in the 81st Precinct, Sergeant Raymond Stakes said the following at a roll call on March 13, 2009:
If you see guys walking down the street, move ‘em along. Two or three guys you can move, you can’t move 15, all right? If you want to be a[n] asshole or whatever you want to call it, make a move. If they won’t move, call me over and lock them up [for disorderly conduct]. No big deal. We could leave them there all night.... The less people on the street, the easier our job will be.... If you stop them[] 250, how hard is a 250. I’m not saying make it up but you can always articulate robbery, burglary, *598whatever the case may be. That’s paperwork ... It’s still a number. It keeps the hounds off, I’ve been saying that for months.243
Similarly, Sergeant Stukes states at a roll call on November 23, 2008: “If they’re on a corner, make them move. They don’t want to move, you lock them up. Done deal. You can always articulate later.”244
In a speech at roll call on Halloween in 2008, Deputy Inspector Mauriello states:
Tonight is zero tolerance. It’s New Years Eve all over again. Everybody goes. I don’t care.... They’re throwing dice? They all go, promote gambling. I don’t care. Let the DA discuss what they’re going to do tomorrow.... They got [bandanas] on and they’re running like nuts down the block, chasing people? Grab them. Fuck it. You’re preventing a robbery.... You know that and I know that.245
When asked to explain what he meant in these remarks, Deputy Inspector Mauriello testified that throwing dice is a quality of life infraction.246
In addition to revealing a virulent precinct culture that the NYPD failed to address until forced to do so by the publication of the recordings, some of the recorded statements are directly relevant to plaintiffs’ claims. Sergeant Stukes’ statements on more than one occasion that “you can always articulate” some basis for a stop after the fact encourages officers to stop first and develop a justification later.247 The same Sergeant Stukes later directs his officers to “[s]hake everybody up. Anybody moving, anybody coming out that building, 250, verticals, and give me a couple of community visits. C-summons as well.” 248 Because exiting a building— even in a high crime area — is not a sufficient basis for reasonable suspicion, these words are an instruction to carry out stops and other enforcement activity without legal justification.
Likewise, the following words by Deputy Inspector Mauriello at a roll call on November 8, 2008 provide evidence that the *599rapid escalation in stops between 2002 and 2011 may have been accomplished in part by encouraging stops without reasonable suspicion of any crime:
I’m tired of bandanas on their waist and I’m tired of these beads. Red and black beads mean Bloods. Their bandanas — if they’re walking down the street and they’ve got a bandana sticking out their ass, coming out there — they’ve got to be stopped. A 250 at least. At least.249
Most significantly, Sergeant Stakes repeatedly instructs the officers that their careers depend on carrying out high levels of activity, and shows utter disregard for the requirement that a stop only be made based on a reasonable suspicion that crime is afoot. At a December 8, 2008 roll call, Sergeant Stakes explained:
This job is so easy. Just keep the hounds off. A parker, a 250, you could book somebody walking down the street. You know what? I stopped and asked— so what? I did a 250. What’s the big deal? Let him go. He doesn’t want to give you no information, who cares? It’s still a 250.250
Officer Polanco’s recordings take place in 2009 at the 41st Precinct in the Bronx, and reveal similar pressure on officers to achieve enforcement activity numbers regardless of whether there is a reasonable suspicion of criminal activity. In one recording, Officer Polanco’s union delegate, Officer Angel Herran, refers repeatedly to a requirement that officers complete “20 and 1,” which Officer Polanco testified meant twenty summonses and one arrest per month.251 Officer Herran encourages the other officers to “[cjrush the fucking city” and make the required numbers, because someone who does not is a “zero,” and he will not fight for a zero.252 Officer Herran confirmed in his testimony that “20 and 1” referred to the goal of 20 summonses and 1 arrest, which patrol officers in the 41st Precinct were expected to achieve (at least prior to the Quota Law) in approximately 20 to 22 days on patrol.253 Officer Polanco’s recordings also appear to show Lieutenant Andrew Valenzano instructing officers to stop anyone on a bike who is carrying a bag near an area where there have been car break-ins.254 “[T]hose are good stops,” Lieutenant Valenzano states.255
Officer Serrano’s recordings take place at the 40th Precinct in the Bronx, and also show the pressure for enforcement activity. During a roll call on June 30, 2010, *600Lieutenant Stacy Barrett directs each officer to get five summonses or UF-250s,256 and says it should not be difficult. She tells the officers to “go crazy” in St. Mary’s Park: “If we get every single summons in St. Mary’s, I don’t care.” Her primary concern is to “get those numbers.” 257
Beyond the recordings, many other officers offered credible, consistent testimony regarding the pressure to increase enforcement activity and the threat of adverse consequences for failing to achieve high enough numbers.258
ii. Pressure After 2010 Quota Law
In 2010, after the Village Voice publicized Officer Schoolcraft’s recordings,259 the State of New York enacted the Quota Law, which prohibits retaliation against officers for failing to meet quotas for tickets, stops, summonses, and arrests.260
Subsequently, Chief Hall sent a memo to the commanders of every patrol borough purporting to clarify the NYPD’s position on performance quotas. The memo first clarifies that “a requirement that a specific number of summonses be issued or arrests be made over a specific period of time has always been prohibited.” The memo then states that “[o]fficers who avoid engaging in enforcement activities ... can be subjected to adverse consequences.”261 Furthermore, “[a]n obvious way of gauging an officer’s activity level is to count the number of enforcement encounters that an officer has over time,” and to compare an officer’s activity level to that of similarly situated officers. The memo only explicitly prohibits “discussing specific numerical objectives ” or linking “the failure to reach a specific numerical goal with an adverse employment consequence.”262
In 2011, the NYPD introduced the “Quest for Excellence” program, a set of new policies for evaluating the performance of officers and encouraging the use of performance goals.263 One of the central documents in the Quest program is Operations Order 52 (“OO 52”), which was issued October 17, 2011 and describes officers’ performance objectives. OO 52 made clear that supervisors must evaluate officers based on their activity numbers, with particular emphasis on summonses, stops, and arrests, and that officers whose numbers are too low should be subjected to increasingly serious discipline if their low numbers persist. Specifically, NYPD managers “can and must” set “performance goals” for “proactive enforcement activities,” with “particular attention” to “self-initiated arrests, issuing summonses, [and] conducting stops.” 264 Deputy Commissioner Beirne testified that “performance goals” could include “setting a goal of *601a certain number of stops.” 265 Officers who “fail to engage in proactive activities,” and thus continue to fail in “addressing sector/post conditions,” will ultimately be referred to the “Performance Monitoring Unit” for potential “transfer, reassignment or other appropriate disciplinary action.” 266
In addition, the form used to track officer performance reflects the NYPD’s emphasis on enforcement activity numbers and effectiveness without attention to the constitutional justifications for enforcement. An officer is required to tally her activities each day on the Police Officer’s Monthly Conditions Impact Measurement Report.267 The form contains columns for a number of activities, including vertical patrols, radio runs, arrests and summonses, and the preparation of various reports, including UF-250s.268 At the end of the month, the officer tallies her total activities, and the supervisor provides a brief written evaluation of whether the officer’s “impact on declared conditions” was “effecfive.” Each quarter, the supervisor reviews the officer’s activity over the prior three months and evaluates the officer’s effectiveness.269
In contrast to this detailed review system for effectiveness, there is no process for evaluating whether enforcement activities are legally justified.270 For the purposes of performance review, an unconstitutional stop is no less valuable to an officer’s career than a constitutional one — because the two are indistinguishable. In fact, a review of several monthly activity reports suggests that in practice, many officers are evaluated almost exclusively based on the number of stops, arrests, and summonses that they carry out. Based on these reports, as well as corroborating testimony, an “effective” “impact on declared conditions,” in the context of performance reviews, is sometimes nothing more than a euphemism for an acceptable number of stops, arrests, and summonses in targeted locations.271
*602Officers may be subject to warnings and more severe adverse consequences if they fail to achieve what their superiors perceive as appropriate enforcement activity numbers. Deputy Commissioner Beirne acknowledged that an officer’s failure to engage in enough proactive enforcement activities could result in a negative performance evaluation and reassignment to a different command, and that both of these steps represent “an adverse employment action.” 272 As noted earlier, the only proactive enforcement activities mentioned in 00 52 are stops, summonses, and arrests.273
It is difficult to see any difference between a performance goal and a quota if “performance goals” operate as Deputy Commissioner Beirne testified.274 It is not surprising, then, that since 2010, there have been at least nine grievances filed by police officers against the NYPD alleging adverse employment action as a result of quotas.275
d. Conclusion
The foregoing evidence shows that officers are routinely subjected to significant pressure to increase their stop numbers, without corresponding pressure to ensure that stops are constitutionally justified. Together with evidence described in the next section, this is a predictable formula for producing unjustified stops. To paraphrase a statement by Chief Hall from his 2010 memo, imposing numerical performance goals for enforcement activities, without providing effective safeguards to ensure the activities are legally justified, “could result in an officer taking enforcement action for the purpose of meeting a [performance goal] rather than because a violation of the law has occurred.”276
3. Targeting “the Right People”
The role of race in stop and frisk has been a source of contention since the Supreme Court first sanctioned the practice in 1968. In Terry, the Supreme Court recognized that “ ‘[i]n many communities, field interrogations are a major source of friction between the police and minority groups,’ ” and that friction “ ‘increases as more police departments [encourage] officers ... routinely to stop and question persons on the street.’ ” 277 In 1996, the Ninth Circuit noted that these stops “are humiliating, damaging to the detainees’ self-esteem, and reinforce the reality that racism and intolerance are for many Afri*603ean-Americans a regular part of their daily lives.”278
The NYPD maintains two different policies related to racial profiling in the practice of stop and frisk: a written policy that prohibits racial profiling and requires reasonable suspicion for a stop 279 — and another, unwritten policy that encourages officers to focus their reasonable-suspicion-based stops on “the right people, the right time, the right location.”280
Based on the evidence summarized below, I find that the NYPD’s policy of targeting “the right people” encourages the disproportionate stopping of the members of any racial group that is heavily represented in the NYPD’s crime suspect data. This is an indirect form of racial profiling. In practice, it leads NYPD officers to stop blacks and Hispanies who would not have been stopped if they were white. There is no question that a person’s race, like a person’s height or weight, is a permissible consideration where a stop is based on a specific description of a suspect.281 But it is equally clear that it is impermissible to subject all members of a racially defined group to heightened police enforcement because some members of that group appear more frequently in criminal complaints. The Equal Protection Clause does not permit race-based suspicion.
Chief Esposito, the highest ranking uniformed member of the NYPD throughout the class period and the chair at Compstat meetings, was especially frank about the NYPD’s policy of targeting racially defined groups for stops, provided that reasonable suspicion is also present:
Q. Quality stops are stops that are in the right place at the right time, correct?
A. Yes.
Q. And targeting ... the right people, correct?
A. Among other things.
Q. And the right people would be young black and Hispanic youths 14 to 20, correct?
A. At times, [pause] You failed to mention reasonable suspicion.282
Chief Esposito conceded that not all stops are based on a specific suspect description from a crime complaint. In fact, officers check “Fits Description” on only 13% of UF-250s. Nevertheless, Esposito testified, the NYPD uses criminal suspect data to target certain individuals for stops even when there is no suspect description:
Q: Do you believe the disparity in stop, question and frisk among black and Latino men is evidence of racial profiling?
A: No. I don’t believe that.... Because the stops are based on complaints that we get from the public.
THE COURT: But there are many street stops that have nothing to do with complaints, right?
*604THE WITNESS: Correct.
THE COURT: It’s observed conduct____It’s not based on a complaint of a victim.
THE WITNESS: It’s based on the totality of, okay, who is committing the— who is getting shot in a certain area? ... Well who is doing those shootings? Well, it’s young men of color in their late teens, early 20s.283
Thus, within the pool of people that an officer could reasonably suspect of criminal activity, the people who match the general demographics of the local criminal suspect data are “the right people” to be stopped.284
Other evidence corroborates this interpretation of Chief Esposito’s testimony. On one of the Serrano recordings, Deputy Inspector Christopher McCormack explained to Officer Serrano that stopping “the right people, [at] the right time, [in] the right location” meant not stopping “a 48-year-old lady [who] was walking through St. Mary’s Park when it was closed.” 285 He continued as follows:
INSPECTOR: This is about stopping the right people, the right place, the right location.
SERRANO: Okay.
INSPECTOR: Again, take Mott Haven where we had the most problems. And the most problems we had, they was robberies and grand larcenies.
SERRANO: And who are those people robbing?
INSPECTOR: The problem was, what, male blacks. And I told you at roll call, and I have no problem telling you this, male blacks 14 to 20, 21. I said this at roll call.286
Deputy Inspector McCormack testified that his statements in the recording were based on suspect descriptions from victims. But he also acknowledged that the descriptions of the suspects consisted only of the information stated here: males, black, between 14 and 21.287
Earlier in the recording, when challenged by Officer Serrano, Deputy Inspector McCormack clarified that he does not believe “every black and Hispanic” is subject to being stopped based on the crime suspect data.288 Deputy Inspector McCor*605mack, like Chief Esposito, recognized that reasonable suspicion is required for every stop. But both believe that, within the pool of people displaying reasonably suspicious behavior, those who fit the general race, gender, and age profile of the criminal suspects in the area should be particularly targeted for stops.289
The stop of Cornelio McDonald illustrates the NYPD’s policy of indirect racial profiling based on crime suspect data. Officer Edward French (now a detective) was aware of crime reports that a black male had been burglarizing residences in Queens, as well as reports of a black male committing armed robberies of commercial establishments in the borough. The only information known about the suspects in these robbery patterns was that they were male and black- — -a very general description that would not justify a stop based on “fits description.” 290 Nevertheless, Officer French stopped McDonald in part because he fit the suspect description. McDonald was a black man crossing the street late on a winter night with his hands in his pockets, and as a black man he was treated as more suspicious than an identically situated white man would have been.291 In other words, because two black males committed crimes in Queens, all black males in that borough were subjected to heightened police attention.
The UF-250s prepared by Officer Gonzalez, one of the most aggressive stoppers in 2009, provide a different illustration of an officer responding to the NYPD’s policy of indirect racial profiling based on crime suspect data. Officer Gonzalez checked “Fits Description” on 132 of his 134 UF-250s, although he also indicated that not a single one of those stops was based on an ongoing investigation, a report from a victim, or a radio run.292 Nonetheless, Gonzalez’s supervisor, then-Sergeant Charlton Telford, testified that he was not concerned by this discrepancy. Telford insisted that Officer Gonzalez’s stops were based on “the race, the height, [and] the age” of criminal suspects.293 The following were the suspect descriptions that formed the basis for Officer Gonzalez’s 134 stops:
The burglaries, the description we had was a male Hispanic, between 5'8", 5'9", in his 30s. The robberies were male blacks, anywhere from four to five [in number], between the ages of 14 to 19. And the shooting was a male black in his 20s.294
Perhaps as a result of Officer Gonzalez’s reliance on this general suspect data, 128 of the 134 people he stopped were black or Hispanic.295 This is roughly in line with *606the percentage of criminal suspects in his precinct who are either black or Hispanic (93%), but far exceeds the percentage of blacks and Hispanics in the local population (60%).296 Thus, Officer Gonzalez’s UF-250s provide a perfect example of how racial profiling leads to a correlation between the racial composition of crime suspects and the racial composition of those who are stopped by the police. By checking “Fits Description” as a basis for nearly every stop, Officer Gonzalez documented what appears to be a common practice among NYPD officers — treating generic crime complaint data specifying little more than race and gender as a basis for heightened suspicion.
New York State Senator Eric Adams’ testimony provided further evidence of official acquiescence in racial profiling by NYPD leadership. Senator Adams, a former NYPD captain, testified about a small meeting he attended at the Governor’s office in Manhattan in July 2010. Former New York Governor David Paterson, Senator Adams, another state senator, a state assemblyman, and Commissioner Kelly were all present to discuss a bill related to stop and frisk. Senator Adams raised his concern that a disproportionate number of blacks and Hispanics were being targeted for stops. Commissioner Kelly responded that he focused on young blacks and Hispanics “because he wanted to instill fear in them, every time they leave their home, they could be stopped by the police.”297 Senator Adams testified that he was “amazed” that Commissioner Kelly was “comfortable enough to say that in the setting.”298
I find Senator Adams’ testimony credible, especially in light of the Senator’s former affiliation with the NYPD, Commissioner Kelly’s decision not to appear at trial to rebut the testimony, the City’s failure to offer any rebuttal evidence regarding Commissioner Kelly’s statement at this meeting, and the other evidence of tolerance toward racial profiling at the NYPD. In fact, the substance of Commissioner Kelly’s statement is not so distant from the City’s publicly announced positions. Mayor Bloomberg stated in April that the NYPD’s use of stop and frisk is necessary “to deter people from carrying guns.... [I]f you end stops looking for guns, ... there will be more guns in the hands of young people and more people will be getting killed.” 299 At the same time, the City emphasized in its opening arguments that “blacks and Hispanics account for a disproportionate share of ... crime perpetrators,” 300 and that “90 percent of all violent crime suspects are black and Hispanic.”301 When these premises are combined — that the purpose of stop and frisk is to deter people from carrying guns and that blacks and Hispanics are a disproportionate source of violent crime— it is only a short leap to the conclusion that blacks and Hispanics should be targeted for stops in order to deter gun violence, regardless of whether they appear objectively suspicious. Commissioner Kelly simply made explicit what is readily inferrable from the City’s public positions.
*6074. Inadequate Monitoring and Supervision
Previous sections have described the institutional pressure on NYPD officers to make more stops and the reward for high stop activity. The City argues that the NYPD has taken effective measures to counter the risk that this pressure will lead to legally unjustified stops. However, after nine weeks of trial, the City failed to establish that the NYPD has any effective mechanism for identifying or tracking unconstitutional stops.
a. Inadequate Documentation and Document Review
The City concedes that a UF-250, standing alone, often provides inadequate information to indicate whether a stop was based on reasonable suspicion.302 An earlier version of the UF-250 included a large blank space in which officers were required to state the “factors which caused [the] officer to reasonably suspect [the] person stopped.” 303 The current UF-250, which has been in use since November 2002, does not require the officer to articulate in writing the facts justifying the stop.304 Instead, officers are directed to record the details of each stop in their memo books, also known as “activity logs.”305
In practice, however, officers do not in fact record the factors justifying a stop in their memo books, and supervisors do not address this deficiency. Officer Dennis testified that “while it was suggested” that officers record the “time, date, location of the stop, name of the person stopped, crime suspected, and that a 250 was prepared,” it was not “always” necessary and he did not always do so.306 Officer Dennis’s memo book entry regarding his stop of Devin Almonor included only the time of the stop and the following two statements: “2 males stopped,” “one male refused [illegible mark], UF-250.” With regard to Almonor’s arrest and transfer to the 30th Precinct, Dennis wrote only “one under to 30.” 307 No one ever discussed these memo book entries with him.308 In fact, throughout his many years as a police officer, no superior ever told him that his activity log entries concerning stops were insufficient.309
*608Some supervisors are not even aware that officers are required to record the factors justifying a stop in their memo books. Sergeant Michael Loria could not remember ever being told in training that officers should record more information in their memo books than appears on the UF-250.310 In addition, Chief Esposito openly expressed resistance to the policy of requiring officers to state the justifications for a stop in their memo books, suggesting that the practice was “redundant many times” and interfered with an officer’s “ability to do police work.”311
In fact, the NYPD has been aware for the last decade of a systematic failure by officers to record the justifications for stops in their memo books. Each month, every enforcement command is required to perform a self-inspection of a sample of twenty-five of its UF-250s and corresponding activity log entries. As part of the self-inspection, the reviewer is required to indicate whether, for each stop, the officer made an activity log entry detailing the circumstances of the stop.312
According to QAD, which is responsible for ensuring that the NYPD as an organization complies with its written policies and procedures,313 every patrol borough has failed every annual audit of activity log entries corresponding to stops for the last decade.314 Numerous commanders acknowledged that they received QAD audits showing their officers’ failure to prepare activity log entries for stops, but did not correct the failure.315 Assistant Chief Thomas Dale, the Commanding Officer of Patrol Borough Queens South, acknowledged that failure to complete activity log entries for stops is “a serious problem,” but testified that he took no corrective actions even after every precinct in Queens South failed the QAD audit of activity log entries for three consecutive years.316 Failure on this scale indicates, in the words of plaintiffs’ police practices expert Lou Reiter, that “the operational way memo books are used in the field is contrary to ... all of the written training and all of the policy,” and the supervisors “are not holding their officers accountable for it.”317
On March 5, 2013, five years after the commencement of this litigation and two weeks before the beginning of trial, Chief Hall circulated a memo requiring all patrol borough officers to include nine categories *609of information in every activity log entry for a stop; to elaborate the basis for a stop in the “Additional Circumstances/Factors” section of the UF-250; and to photocopy every activity log entry for a stop and attach the photocopy to the UF-250 before submitting it to a supervisor.318 Just as I gave little weight to the equally ambitious memo that Chief Hall circulated shortly before the beginning of the Ligón preliminary injunction hearing,319 I give little weight to the March 5 memo or to the City’s uncorroborated anecdotal evidence of compliance with it.320
An untested, last-minute adjustment— even if undertaken in good faith — cannot undo ten uninterrupted years of willful disregard. I take the March 5 memo neither as evidence that the NYPD has solved the problem of documenting stops, nor as evidence of the NYPD’s commitment to finding a solution, but rather as a belated recognition of the obvious inadequacies of the existing system of documentation.321
Finally, I note that the NYPD has no meaningful procedures for auditing stop paperwork to monitor the constitutionality of stops. The City agreed to conduct regular audits of whether stops recorded on UF-250s were based on reasonable suspicion when it executed the Stipulation of Settlement in Daniels v. City of New York on September 24, 2003.322 After signing the settlement, however, QAD simply continued to use audit protocols that it had introduced in 2002. These protocols, which remain in effect today, require every command to conduct the monthly “self-inspection” described above, using two forms, Worksheets 802 and 802-A, which are then reviewed by QAD.323 The only arguably substantive element of these forms is a column on Worksheet 802 in which the reviewer is asked to note whether at least one box on the UF-250 was checked under “What Were Circumstances Which Led To Stop?” 324 In other words, the Worksheets lead to a superficial review of whether paperwork was completed, not *610a substantive review of whether a stop was constitutional. QAD’s annual audit of the self-inspections is similarly ineffective.325
Despite the obvious inadequacy of the QAD audits as a means to identify either unjustified stops or racial profiling,326 officials in the NYPD continue to defend the QAD audit process.327 Inspector Mary Cronin, the former Executive Officer and current Commanding Officer of QAD, testified that the QAD audit worksheets allow reviewers to evaluate whether a stop was based on reasonable suspicion, because any form on which at least one stop circumstance has been checked — for example “furtive movements” as the sole basis for a criminal trespass stop — is sufficient “for purposes of [QAD’s] review.” 328 This testimony demonstrates the patent inadequacy of QAD audits. A review of whether a single stop factor has been checked on a OF-250 is not an effective review of the constitutionality of a stop.
b. Inadequate Supervision
As summarized above, the NYPD’s documentation provides an inadequate basis for monitoring whether officers are conducting unconstitutional stops. Nevertheless, the City argues that in practice, the NYPD identifies unconstitutional stops through “a chain of command of supervisors, namely, sergeants, lieutenants, and precinct commanders.” 329 At trial, the City introduced a large volume of testimony and other evidence concerning the details of this chain of command, and the responsibilities of NYPD personnel at various ranks.330 Very little of this evidence was relevant, however, because very little of it concerned the supervision of the constitutionality of stops. In fact, the City notes only two concrete mechanisms for identifying unconstitutional stops: first, sergeants “routinely witness stops made by officers”; and second, sergeants “review their officers’ OF-250s and frequently discuss the underlying facts of stops with officers to determine whether an officer is able to articulate a proper basis for the stop.”331
The evidence showed that neither mechanism provides an effective means for *611monitoring the constitutionality of stops. With regard to sergeants witnessing stops, there was no quantitative evidence concerning how many of the 4.4 million stops during the class period were witnessed by sergeants or how many of the observed stops were self-initiated versus based on radio runs.332 More importantly, the evidence showed that sergeants do not effectively monitor the constitutionality of stops even when they are present.333
With regard to supervisors’ review of stops after the fact, there are no written policies requiring supervisors to evaluate the constitutionality of stops. The section of the Patrol Guide containing the procedures for stops and frisks directs uniformed members of the service to “[i]n-form [the] desk officer, [in the] precinct of occurrence, of [the] facts [of the stop].”334 Chief Giannelli explained that this requirement is satisfied when the officer submits a UF-250 to her supervisor.335 Despite the fact that NYPD officials like Chief Esposito present supervisory review as the central mechanism for monitoring the constitutionality of stops, the section of the Patrol Guide describing the desk officer’s duty to review UF-250s contains no mention of a review for constitutionality, or of any substantive review for that matter.336 In addition, while the lesson plan for training newly promoted sergeants on stop and frisk mentions the need for reasonable suspicion and the prohibition on racial proffling, the plan does not instruct sergeants to perform any review of stops for constitutionality.337
*612The evidence also shows that in practice, supervisors do not review the constitutionality of stops after the fact. Rather, supervisors review whether a UF-250 was fully completed.338 Supervisors are not required to review activity logs alongside UF-250s,339 nor do they routinely discuss the circumstances of a stop with the stopping officer in order to determine whether the stop was justified.340
Alternately, the City suggests that Integrity Control Officers (“ICOs”) are available to monitor the constitutionality of stops.341 ICOs are “the eyes and ears of the Precinct Commander,” and they are tasked with performing various inspections and reviews to identify misconduct by officers.342 While ICOs might in theory be well-positioned to review the constitutionality of officers’ stops,343 there is no evidence that ICOs do so in practice, or are instructed to do so through training or written policies.344 The detailed thirty-one-point list of an ICO’s duties and responsibilities contains no mention of monitoring the constitutionality of officer behavior in general or of stops specifically, nor does it require ICOs to review officers’ UF-250s or activity log entries for stops.345 *613Former Chief Giannelli testified that he expects ICOs to review UF-250s to determine whether officers’ stops are based on reasonable suspicion.346 In practice, however, ICOs review UF-250s and officers’ activity logs solely for completeness, if at all,347 rather than for the constitutionality of the underlying stop.348
In sum, neither the NYPD’s review of stop documentation nor its supervision of officers provides an adequate mechanism for identifying unconstitutional stops. Consequently, the NYPD is unable to hold officers accountable for those stops or prevent them from happening in the future. I also note that the failure of supervisors and ICOs to effectively supervise the constitutionality of stops is not the result of oversight by subordinates, but rather stems from the failure of senior NYPD managers and officials to direct supervisors and ICOs to perform this task.349
5. Partially Inadequate Training
The core constitutional standards governing stop and frisk are well established.350 Training officers to comply with these standards, however, is no simple task, because there are no mechanical rules for their application to the varied circumstances of an officer’s work. Viewed in light of this difficulty, the NYPD’s efforts to train its recruits have been largely adequate.351 The gravest problems in the NYPD’s stop and frisk pfactices stem not from inadequate training but from a divergence between the NYPD’s written training materials and the “operational policy”' carried out in the streets.352
Nevertheless, plaintiffs accurately note several problematic and obvious omissions or errors in the NYPD’s training programs.353 Some of these flaws effectively encourage officers to commit constitutional violations, and the evidence in this case shows that predictably, such violation have in fact occurred on a widespread basis. The following are some of the problems I find in the written training materials.
First, the Police Student’s Guide’s training on reasonable suspicion explains “Furtive Behavior” as follows: “If an officer observes strange, suspicious, or evasive behavior, he or she may have reasonable suspicion. The officer’s experience and/or expertise are often taken into account in *614these situations.” 354 The vagueness and overbreadth of this description invites officers to make stops based on “hunches,” in violation of Terry. Given the frequency with which Furtive Movements is checked (roughly 42% of forms), and the obvious risk that stops based merely on “strange, suspicious, or evasive behavior” may lack reasonable suspicion, the Guide’s description of furtive movements is inadequate.
The danger of this inadequate training is illustrated by the testimony of Officer Christopher Moran, who stopped David Ourlicht for walking in a suspicious way with an ostensible bulge under his winter clothing.355 Officer Moran testified that “people acting nervous” could “[o]f course” provide reasonable suspicion for a stop.356 Officer Moran also explained that “furtive movement is a very broad concept,” and could include “changing direction,” “walking a certain way,” “acting a little suspicious,” “making a movement that is not regular,” being “very fidgety,” “going in and out of his pocket,” “going in and out of a location,” “looking back and forth constantly,” “looking over their shoulder,” “adjusting their hip or their belt,” “moving in and out of a car too quickly,” “[tjurning a part of their body away from you,” “[gjrabbing at a certain pocket or something at their waist,” “[gjetting a little nervous, maybe shaking,” and “stutter[ing].” 357 To the extent that Officer Moran views nervousness or fidgeting, standing alone, as an adequate basis for seizing, questioning, and potentially frisking a person under the Fourth Amendment, he is incorrect. But his view is also a natural response to the vague and overly broad description of furtive movement in the Police Student’s Guide. Misconceptions like Officer Moran’s are a predictable consequence of the training reflected in the Guide, and likely lead to unconstitutional stops.358
Second, the NYPD’s training on the identification of weapons invites unjustified stops based on “suspicious bulges” that are not in fact suspicious, and constitutionally unjustified frisks and searches based on objects that officers cannot reasonably suspect to be weapons. In particular, the training draws attention to several “unusual firearms,” such as a gun shaped like a pen, a gun shaped like an old-model cell phone, and a folding gun that fits into a wallet.359 It is no doubt valuable for officers’ safety to know that such weapons exist.360 However, the outline of a commonly carried object such as a wallet or cell phone does not justify a stop or frisk, nor does feeling such an object during a frisk justify a search. The training materials are misleading and unclear on this point.361 The materials encourage officers *615to perform stops and frisks without reasonable suspicion based on the now-ubiquitous bulge created by a cell phone or other common objects — as was the case in the stops of Leroy Downs, Devin Almonor, Cornelio McDonald and Nicholas Peart,362 and was likely the case in the vast majority of stops involving suspicion that the suspect was carrying a weapon, based on the extremely low seizure rate.363
Third, more generally, the NYPD’s training materials fail to make clear the legal standard for when a frisk may be undertaken. Rather than simply stating that a frisk must be based on reasonable suspicion that a person is armed and dangerous,364 the training materials present a four-part rule that includes an invitation for officers to conduct frisks whenever “they are in fear of their safety,” without clarifying that the fear must be both reasonable or related to a weapon.365 Based on this four-part rule, which officers have apparently internalized,366 it is not surprising that several of the individual stops in this case involved unconstitutional frisks,367 or that the 2.3 million frisks during the class period resulted in the recovery of weapons only 1.5% of the time.368
Fourth, while the NYPD deserves praise for its use of scenarios and role playing in training on the legal justification for stops, plaintiffs correctly observe that these exercises are exclusively based on radio runs rather than self-initiated stops.369 As noted above, 78% of stops during the class period were self-initiated,370 and self-initiated stops create a different set of constitutional risks than stops based on suspect descriptions from a radio run. Just as ICOs observing stops based on radio runs cannot provide adequate monitoring of self-initiated stops, role playing that focuses exclusively on radio runs is inadequate to train officers about the application of *616constitutional standards to self-initiated stops.
Fifth, the NYPD’s training regarding racial profiling does not clearly define the difference between the constitutionally permissible use of race in a stop based on a specific, reliable suspect description, and the constitutionally impermissible targeting of racially defined groups for stops in general.371 Because the NYPD has an unwritten policy of conducting the latter type of stops, as described above, this omission is not surprising.
Sixth, I have already noted in another opinion the constitutional infirmities in the NYPD’s stop and frisk refresher course at the Rodman’s Neck training center in the Bronx, as well as other recently introduced materials.372 In that opinion, I expressed concern that the recent training materials misstate what constitutes a stop, which likely leads some officers not to prepare OF-250s in cases when they wrongly conclude that the encounter did not rise to the level of a stop.373 Although Chief Shea testified that the Rodman’s Neck training was developed in part because the NYPD wants officers to fill out OF-250s more selectively, a more significant concern is the failure of officers to fill out OF-250s after encounters that clearly were stops.374 In addition, testimony in the present case confirmed that officers misunderstand what constitutes a stop based on training materials used at Rodman’s Neck.375
*6176. Inadequate Discipline
As described in earlier sections, the NYPD has chosen to document, review, and supervise its officers’ stops in such a way that unconstitutional stops are unlikely to be identified, and as a result the responsible officer cannot be held accountable. Moreover, when confronted with evidence of unconstitutional stops, the NYPD routinely denies the accuracy of the evidence, refuses to impose meaningful discipline, and fails to effectively monitor the responsible officers for future misconduct.
Civilian complaints are one source of notice to the NYPD that an unconstitutional stop may have taken place. Anyone can make a complaint against an NYPD officer for misconduct through a variety of channels, including the CCRB, an all-civilian municipal agency tasked with investigating allegations against the NYPD of excessive force, abuse of authority, discourtesy, or offensive language (collectively referred to as “FADO”); 376 the Bloomberg administration’s “311” information and government services hotline; or by contacting the NYPD directly.377 Of course, a complaint made to the NYPD may not even be recorded if the officer receiving that complaint decides to ignore it — as Officer Anthony Moon did in response to Downs’s complaint.378 It is impossible to know how often the NYPD simply disregards complaints.
The correct practice when the NYPD receives a complaint regarding a stop is to either forward the complaint to the CCRB, or, if the complaint falls outside of the FADO categories, to the Investigative Review Section of the Office of the Chief of Department (“OCD”).379 The NYPD’s Internal Affairs Bureau (“IAB”) can, in theory, initiate its own investigations into alleged misconduct based on media reports,380 although no evidence was offered that IAB has in fact done this in response to the media reports over the last decade concerning racially biased and/or constitutionally unjustified stops and frisks.
Once the CCRB receives a complaint of an unconstitutional stop, CCRB investigators will generally “interview complainants, witnesses, and officers and determine whether allegations are substantiated, unsubstantiated, exonerated, or unfounded.”381 A complaint is “substantiated” if, based on the preponderance of the evidence, “[tjhere is sufficient credible evidence to believe that the subject officer committed the act charged in the allegation and thereby engaged in misconduct.” 382 A complaint is “unsubstantiated” if there is insufficient evidence to determine whether the officer committed misconduct.383 The in*618vestigator’s decision will be reviewed by a supervisor and then passed along to a three-member panel, which makes the final decision after reviewing the evidence. If the panel finds a complaint substantiated, the case is forwarded to the NYPD’s Department Advocate’s Office (“DAO”), which serves as an internal prosecutor for officer misconduct at the NYPD.384
It is at this stage that the CCRB complaint process founders. Rather than accepting the CCRB’s findings, the DAO conducts its own review of the materials that the CCRB’s three-member panel has just reviewed. Instead of applying the well-established “preponderance of the evidence” standard like the CCRB, the DAO applies its own evidentiary standards. Deputy Commissioner Julie Schwartz, an attorney who has lead the DAO since 2005, testified that if the CCRB bases its findings on a credibility determination in favor of a witness and against a police officer, the DAO will as a rule reject the CCRB’s findings. In Deputy Commissioner Schwartz’s words, it “doesn’t matter” that the CCRB has seen and heard the witnesses, while the DAO has not, “[bjecause if one witness says A happened and the other witness says B happened,” the complaining witness cannot satisfy the preponderance of the evidence standard: “there is nothing that brings [the allegation] to 51 percent.” 385 In this “he said ... she said” situation, the DAO will only accept the CCRB’s findings if there is “a little corroboration,” like visible marks on a complainant alleging overly tight handcuffs.386
Deputy Commissioner Schwartz stated further that “in any stop-and-frisk case in which you only have the complainant’s version and the officer’s version, the department cannot pursue discipline.” 387 It appears that in the eight years that she has led the DAO, Deputy Commissioner Schwartz has consistently applied the unique evidentiary standard that a complainant’s testimony can never be sufficient standing alone.
Deputy Commissioner Schwartz also testified that the CCRB and the OCD apply the same evidentiary standard,388 although she was unable to cite any source articulating that standard. When pressed, Deputy Commissioner Schwartz changed her testimony and stated that in some instances a complainant’s word is enough to sustain a complaint.389 I do not find this later testimony credible.390 Instead, I find that the DAO follows a policy of rejecting CCRB determinations when they are based only on the uncorroborated testimony of a civilian witness. Deputy Commissioner Schwartz testified that she was not concerned about allegations that the DAO is biased in favor of police officers.391
Because many of the complaints related to stop and frisk involve precisely the sce*619nario described by Deputy Commissioner Schwartz, the DAO’s evidentiary theory seriously undermines the NYPD’s ability to hold officers accountable for unconstitutional stops or frisks.392 In light of this evidentiary standard, as well as other indications of the DAO’s resistance to evidence of unconstitutional stops provided by the CCRB,393 it is not surprising that the DAO frequently declines to pursue any discipline against officers who have been the subject of substantiated CCRB complaints. Between 2007 and 2011, the DAO declined to pursue discipline in between 16% and 36% of the substantiated complaints forwarded by the CCRB, and in 2012 the percentage of substantiated cases resulting in no discipline rose again.394 In addition, the NYPD consistently downgrades the discipline recommended by the CCRB, imposing only instructions — the least serious form of discipline — in the majority of cases in most years.395 The DAO’s frequent rejection of the CCRB’s disciplinary recommendations has likely undermined public confidence in the CCRB and discouraged the filing of complaints — many of which may have been meritorious.396
The relatively few stop complaints that pass through the OCD are no more likely to result in officer discipline than those processed by the CCRB and the DAO.397 *620Despite the fact that the OCD is apparently the only entity responsible for addressing civilian complaints of racial profiling, its system for categorizing and tracking complaints contains no tracking code for either stop complaints or complaints of racial profiling.398 Not surprisingly, the evidence showed that the OCD has been ineffective in monitoring and imposing discipline in response to allegations of racial profiling. For example, the OCD received an allegation that Officer Jonathan Rothenberg stopped and arrested someone based on racial profiling a year before his stop of Ian Provost,399 yet he was never questioned about the allegation,400 and the NYPD is unable to determine whether an investigation ever occurred401 The OCD’s inadequate response to civilian complaints of wrongful stops is all the more troubling because the OCD receives most of the racial profiling complaints that are addressed either to the CCRB or the NYPD.402
7. Ongoing Notice of Constitutional Violations
The 1999 AG Report put the NYPD on notice that its stop and frisk practices were resulting in constitutional violations. Despite that notice, senior NYPD officials significantly increased the risk of constitutional violations by applying pressure throughout the chain of command to raise the number of stops without imposing a countervailing pressure to ensure their constitutionality, and without instituting adequate supervisory, monitoring, or disciplinary procedures. This section describes the various ways the NYPD has continued to receive notice since 1999 of widespread constitutional violations in its practice of stop and frisk.
The NYPD has received thorough and consistent notice of constitutional problems in its stop practices from multiple sources, including the media, community members, community and legal organizations, individual police officers, and the class members and attorneys in this case. As Reiter testified, the prominent media coverage of complaints about baseless and racially motivated stops would have given “any reasonable police department” notice of the need “to take a look at” these issues.403 In fact, Chief Esposito is aware of media reports that the NYPD racially profiles young black and Hispanic males in its stop activities, and has personally heard complaints from community organizations, civil liberties groups, and elected officials about racial profiling in stops.404 Deputy Chief Marino testified that the NYPD’s stop and *621frisk practices are “always a concern and a complaint in any precinct I’ve worked in,” and that community leaders and community members had complained to him about suspicionless stops and frisks.405 Similarly, Inspector Lehr has heard complaints from black residents about being stopped for no reason, and has witnessed demonstrations in overwhelmingly black precincts that raised concerns about stop and frisk and racial profiling.406 Chief Morris testified that he, too, had heard complaints from individuals who felt they were stopped for no reason.407 Chief Shea testified that when members of the community' participate in the multicultural immersion course for newly graduated officers, community leaders have repeatedly shared complaints involving stops based on racial profiling.408 Similarly, Chief Shea testified that when he was a precinct commander, “there were occasions when community members might come to a community meeting and claim a family member or themselves was the subject of police action and they fear it was influenced by their race.” 409 Despite this extensive evidence, Chief Esposito and other NYPD officials testified that they had never heard an individual complain about being stopped based on his race.4101 do not find this testimony credible given the other testimony in this case and the widespread public concern regarding racially biased policing by the NYPD.411 In any event, whether or not high officials heard individual complaints is immaterial, given the many other sources of notice discussed in this section.
In addition to receiving notice from the public and ■ community organizations, the NYPD has been apprised of unconstitutional practices by some of its own officers. In 2009, Officer Polanco delivered an anonymous letter 412 to his ICO, reporting that officers were engaging in racial profiling and other misconduct toward minority communities:
[W]e were handcuffing kids for no reason. They would just tell us handcuff them. And boss, why are we handcuffing them? Just handcuff them. We’ll make up the charge later.
Some of those kids were not doing anything. Some of those kids were just walking home. Some of those kids were just walking from school413
In his letter, which he believed would be forwarded to IAB, Officer Polanco described the following incident, from 2009:
I remember one incident where one kid — and I reported this — they stopped his brother. He was 13. And he was waiting for him from school at the corner to bring him home. When he came to us, the officer — Officer, what’s wrong with my little brother? Was he acting *622out? He wind[s] up with handcuffs too. For simply asking what was going on with his brother.414
Officer Polanco also reported in the letter that on more than one occasion, he was required to drive on patrol with supervisors who directed him to stop individuals without what he believed to be reasonable suspicion. A supervisor would point to a “group of black kids or Hispanic kids on the corner, in the park, or anywhere,” and direct Officer Polanco to “just go grab, go 250 them, go summons them. Sometimes they will ask me to summons them. We will ask the supervisor why. And they will say unlawful assembly or something like that ... [bjecause there’s more than three of them on the corner.”415
After his initial letter in 2009, Officer Polanco anonymously called IAB to express his concerns regarding the treatment of minority youths and the manipulation of crime statistics.416 Finally, Officer Polanco gave IAB the recordings played at trial, which corroborated his testimony about institutional pressure to meet target numbers of enforcement activity by making stops and arrests without an adequate legal basis.417
CCRB complaints regarding stop and frisk, discussed above, provided a further source of ongoing notice to the NYPD.418 In addition, the NYPD has received mounting evidence regarding unconstitutional stops through the proceedings in Daniels v. City of New York, filed in 1999; 419 this case, filed in 2008; 420 and *623the two related stop and frisk cases, Dams v. City of New York and Ligón v. City of New York, filed in 2010 and 2012, respectively.421
The City attempts to rebut evidence of notice by pointing to a 2007 report by the RAND Corporation that found little evidence of pervasive racial profiling in the NYPD’s pedestrian stop and frisk activity.422 But the RAND study used violent crime suspect data as its benchmark, which is problematic for reasons discussed at length in the section on expert testimony.423 As a result, the Report concluded that the NYPD’s stop patterns are racially neutral because the people stopped by the NYPD resemble the suspects in reported violent crimes — even though less than a quarter of stops are based on suspicion of violent crimes,424 and even though blacks are more highly represented in the violent crime suspect data than in the suspect data for other types of crime.425 Nonetheless, the Report cautioned that the use of benchmarks “can either detect or hide racial bias due to unobserved or unmeasured factors that affect both the racial distribution that the benchmark establishes and the racial distribution of the stops.”426
Moreover, even if the NYPD did not question the Report’s choice of benchmark, the Report concluded that “for some particular subsets of stops, there are racial disparities, and, in some boroughs for some outcomes, the disparities are fairly large” — such as in Staten Island for frisks *624of blacks, and Brooklyn South for the use of force against blacks.427 The Report recommended a “closer review” of stop outcomes in these boroughs, which was never carried out.428 The NYPD also failed to effectively implement other recommendations in the Report.429 In light of these warnings and recommendations, the NYPD’s reliance on the RAND Report as proof that its stops are racially neutral was and is unreasonable. In sum, the numerous sources discussed above were more than adequate to put the NYPD on notice that its officers were engaging in racially motivated stops, and nothing in the RAND report justified ignoring those clear signs.
D. Individual Stops
Plaintiffs offered evidence from twelve individuals regarding nineteen stops. In twelve of those stops the plaintiff testified as did some or all of the police officers involved in the stop. Evaluating whether these stops — and often frisks — complied with the Fourth Amendment turned, in large part, on the credibility of the witnesses. In some cases the testimony of the plaintiff and the officers was irreconcilable, and I was required to accept one version and disregard the other.
In the remaining seven stops, no officer testified — either because the officers involved in the stop could not be identified or, in one case, because the officers identified dispute that the stop ever occurred. In the majority of these cases the plaintiffs testimony was either not sufficiently credible or lacked sufficient detail to establish that the stop was unconstitutional. In others, I found the plaintiffs testimony sufficiently credible and sufficiently detailed to make a determination about the constitutionality of the stop. I recognize the dilemma posed by making findings based on only one side of the story, particularly given the NYPD’s apparent good faith efforts to locate the officers involved. However, finding a failure of proof in all stops where no officer was identified would create a perverse incentive for officers not to record stops where the basis for those stops or the police conduct during the stop was clearly problematic, which would immunize the most egregious stops from scrutiny. This appears to be what happened in the stop of Leroy Downs, discussed below, and it was only through his extraordinary persistence that the officers involved were identified.
Evaluating the police conduct in each of these stops is an imperfect science. There is no objective contemporaneous recording of the stop — either audio or visual. I am relegated to finding facts based on the often conflicting testimony of eyewit*625nesses. The task is particularly challenging where everyone who testified had an interest in the outcome, which may have, consciously or otherwise, affected the veracity of his or her testimony. I understand that a judge reviewing the facts in hindsight is in an entirely different position from officers on the beat making split-second decisions in situations which may pose a danger to themselves and others. With this in mind, I have endeavored to exercise my judgment faithfully and impartially in making, as I am required to do, the following findings of fact with respect to each of the nineteen stops at issue.430 I have placed each stop in one of three categories: (1) unconstitutional stop and frisk (if frisk occurred); (2) unconstitutional frisk only; (3) insufficient evidence to find that the stop or frisk was unconstitutional.
1. Unconstitutional Stop and Frisk
a. Leroy Downs
i. Findings of Fact
Leroy Downs is a black male resident of Staten Island in his mid-thirties.431 On the evening of August 20, 2008, Downs arrived home from work and, before entering his house, called a friend on his cell phone while standing in front of a chain link fence in front of his house. Downs used an earpiece connected to the phone by a cord, and held the cell phone in one hand and the black mouthpiece on the cord in the other.432
Downs saw a black Crown Victoria drive past and recognized it as an unmarked police car. The car stopped, reversed, and double-parked in front of Downs’s house, at which point Downs told his friend he would call back.433 Two white plainclothes officers, later identified as Officers Scott Giacona and James Mahoney, left the car and approached Downs.434 One officer said in an aggressive tone that it looked like Downs was smoking weed. They told him to “get the [fuck] against the fence,” then pushed him backwards until his back was against the fence. Downs did not feel free to leave.435
Downs explained that he was talking on his cell phone, not smoking marijuana, that he is a drug counselor, and that he knows the captain of the 120th Precinct. Without asking permission, the officers patted down the outside of his clothing around his legs and torso, reached into his front and back pants pockets and removed their contents: a wallet, keys, and a bag of cookies from a vending machine. The officers also searched his wallet.436
After the officers failed to find any contraband, they started walking back to the car. Downs asked for their badge numbers. The officers “laughed [him] off” and said he was lucky they did not lock him up. Downs said he was going to file a com*626plaint, and one of them responded by saying, “I’m just doing my [fucking] job.” Charles Joseph, a friend of Downs who lives on the same block, witnessed the end of the stop. After the officers drove away, Downs walked to the 120th Precinct to file a complaint.437
Downs told Officer Anthony Moon at the front desk that he wanted to make a complaint and described whát had happened. Officer Moon said that he could not take the complaint because Downs did not have the officers’ badge numbers, and that Downs should file a complaint with the CCRB. As Downs left the station he saw the two officers who stopped him driving out of the precinct in their Crown Victoria, and he wrote down its license plate number on his hand.438
Downs then returned to the station. He tried to give Officer Moon the license plate information, but Officer Moon said that he should give the information to the CCRB instead. Downs waited at the station until he saw the two officers come through the back door with two young black male suspects.439
Downs pointed out the two officers to Officer Moon and asked him, “Can you get their badge numbers?” Officer Moon talked to the officers and then told Downs “maybe you can ask them.” At that point, Downs went outside again and took a picture of the license plate on the Crown Victoria, which was the same number he had written on his hand.440
Eventually, Downs spoke with a supervisor, who said he would try to get the officers’ badge numbers and then call Downs. The call never came. Having spent a few hours at the station, Downs went home.441
The next day, Downs submitted a complaint to the CCRB. Five months later, Officers Mahoney and Giacona both testified under oath to the CCRB that they had no memory of stopping and frisking Downs — an assertion that was “not entirely credited” by the CCRB, because it is “unlikely that PO Giacona and PO Mahoney would not recall their actions immediately prior to effecting two arrests.” The CCRB substantiated Downs’s complaint that Officers Mahoney and Giacona failed to provide their badge numbers. The CCRB found the complaints that the officers stopped Downs without reasonable suspicion, and used profanity unsubtantiated. The CCRB found Downs’s allegation of a search into his pants pockets “unfounded,” based in part on Joseph’s testimony that he did not witness a search. The CCRB substantiated the complaint against Officer Moon for failing to process Downs’s complaint.442
Neither Officer Mahoney nor Officer Giacona received any discipline as a result of the CCRB’s recommendations. Instead, each lost five vacation days for failing to *627make a memo book entry for the Downs stop. They also failed to prepare a UF-250 for the stop, but received no discipline for this. Officer Mahoney has since been promoted to Sergeant.443
Officers Mahoney and Giacona testified that they have no recollection of the Downs stop. Like the CCRB, I do not find their denials of recollection credible.444
Downs testified that he has been stopped “[m]any times” other than the stop on August 20, 2008.445
ii. Mixed Findings of Fact and Law
Downs was stopped when the officers told him to “get the [fuck] against the fence.” The officers lacked reasonable suspicion to stop Downs. The officers seized Downs based on a glimpse of a small object in Downs’s hand from the window of their passing car. The officers’ hunch, unaided by any effort to confirm that what they glimpsed was contraband, was too unreliable, standing alone, to serve as a basis for a Terry stop.
Moreover, whatever legal justification the officers might have had for the stop dissipated shortly after they approached Downs. The absence of any physical evidence, smoke or marijuana smell, and Downs’s explanation that he was talking on his mouthpiece, negated any ground for reasonable suspicion. Just as an officer may not reach into the pocket of a suspect after a frisk has negated the possibility that the pocket contains a dangerous weapon or immediately perceptible contraband,446 so an officer may not persist in stopping a person after the suspicion giving rise to the stop has been negated.447 Officers Mahoney and Giacona violated Downs’s rights under the Fourth Amendment by stopping him based on a hunch, and continuing to detain him after it became clear that he had not been smoking marijuana.
The officers further violated the Fourth Amendment by frisking Downs without any objective basis for suspecting that he was armed and dangerous. Nothing about the suspected infraction — marijuana use — in combination with the facts summarized above provides reasonable suspicion that Downs was armed and dangerous.
The officers further violated Downs’s Fourth Amendment rights by searching his pockets and wallet after the frisk. Such a search would only have been justified if the officers’ frisk of the outer surfaces of Downs’s pockets gave rise to reasonable suspicion that his pockets contained a dangerous weapon, or if the frisk made it immediately apparent that an object in his pockets was a form of contra*628band. Nothing in Downs’s pockets could have provided reasonable suspicion that he was armed; nor could it have been immediately apparent from the patdown that Downs’s pockets contained contraband.448
b. Devin Almonor
i. Findings of Fact
Devin Almonor is a sixteen-year-old black male high school student living in Manhattan. In 2010, Almonor was thirteen years old.449 He was approximately five foot ten and weighed approximately 150 pounds.450
On March 20, 2010, a Saturday, around 8:45 p.m., Almonor left his house to walk his friend Levon Loggins to the bus stop at 145th Street and Amsterdam.451 After Loggins boarded the bus, Almonor began to walk home along Hamilton Place toward a bodega where he planned to meet his brother Malik.452 A group of males was standing outside the bodega and, after talking to friends outside, Almonor continued home with another individual.453
Around 10:00 p.m., Officer Brian Dennis and Sergeant Jonathan Korabel 454 were driving an unmarked vehicle in the vicinity of Hamilton Place in response to nine 911 calls describing a group of about forty youths fighting, throwing garbage cans, and setting off car alarms. A few calls indicated the possibility that weapons were involved.455 The calls suggested that the youths were dispersing when marked cars arrived and then returning. When the officers arrived at Hamilton Place there were garbage cans in the middle of the street and car alarms still going off.456 The only description they had of the individuals was that they were young black males.457
The officers briefly observed Almonor and another individual walking on Hamilton Place in the direction from which the calls originated.458 The individuals crossed 141st Street.459 The officers — two white males in plainclothes — pulled up alongside Almonor, at which point Almonor retreated onto the sidewalk.460 After the officers exited the car and approached Almonor, Officer Dennis grabbed Almonor’s arm and said: “Police.” 461 Almonor pulled away and within moments, Officer Dennis pushed Almonor down on the hood of the police car because he was not “satisfied *629[that Almonor] did not have something in his waist.”462
Together the officers handcuffed Almonor.463 Without explanation, Officer Dennis patted Almonor down from his feet to his torso, during which Almonor was saying, “What are you doing? I’m going home. I’m a kid.” 464 The officers did not recover anything — Almonor only had a cell phone in his right front pocket and a few dollars.
The officers did not ask Almonor his name until after he was handcuffed.465 Almonor did not have ID but identified himself as “Devin Al.” 466 Almonor told the officers that he was thirteen years old and was going home, which was a few blocks away.467 At some point, though not initially, Almonor gave the officers his full address.468 The officers did not ask for Almonor’s phone number or whether his parents were home — instead the officers put Almonor in the back of the patrol car, took him to the precinct, and placed him in the juvenile room because of the possibility that he was thirteen.469
After Almonor was released, Officer Dennis completed a handwritten UF 250 form and a juvenile report.470 The suspected crime was criminal possession of a weapon, and the circumstances of the stop indicated on the form were “fits description” and “furtive movements.” The “suspicious bulge” box was not checked and Officer Dennis testified that he did not see a suspicious bulge that night.471 No contemporaneous document noted that Almonor was touching his waistband.472 The juvenile report form indicated that Almonor was “resisting arrest,” although Almonor was never arrested.473 The next morning, Officer Dennis filled out a computerized *630UF-250 and another juvenile report worksheet, both of which noted a suspicious bulge.474
ii. Mixed Findings of Fact and Law
Almonor was stopped when the officers approached him on the sidewalk, and Officer Dennis grabbed Almonor’s arm and said: “Police.” Even if credited, Almonor’s alleged furtive movements — looking over his shoulder and jaywalking — in combination with the generic description of young black male does not establish the requisite individualized suspicion that Almonor was engaged in criminal activity.475 The officers could have approached Almonor and asked him some questions, but instead chose to physically restrain and handcuff him first, and ask questions later. The circumstances did not justify any restraint of Almonor’s liberty, much less immediate physical restraint and the use of handcuffs.476
Even if the officers had possessed the requisite basis to stop Almonor — which they did not — they had no basis to frisk him. While some of the 911 calls suggested that some youths involved in the fight-mg may have had weapons, that alone does not establish individualized suspicion that Almonor was armed and dangerous. No contemporaneous document indicates a suspicious bulge, and Almonor was not in possession of anything that would have created a suspicious bulge. Almonor’s actions did not indicate that he was armed.477
Finally, not only were Almonor’s Fourth Amendment rights violated at the inception of both the stop and the frisk, but the officers made no effort to minimize the intrusion on his liberty. Instead, they used the most intrusive methods at their disposal, thereby exacerbating the violation of his rights.478
c. Cornelio McDonald
i. Findings of Fact
Cornelio McDonald is a middle-aged black male who resides on Parsons Boulevard, in a private co-op apartment building in Queens.479 McDonald often cares for his mother who lives across the street from him in a New York City Housing Authority (“NYCHA”) complex called Pomonok Houses.480 Parsons Boulevard is a wide *631street with a concrete island dividing two lanes and cars parked on both sides- of each lane.481 The majority of residents on McDonald’s side of the block are white, while the majority of residents on his mother’s side of the block are black.482
On December 18, 2009, a Friday, McDonald spent approximately ten hours at his mother’s house and left her building around 1:00 a.m.483 Because the weather that night was below freezing, McDonald was wearing a zipped-up jacket, with his hands in his pockets the entire time he was crossing the street. He had his cell phone in his left jacket pocket, his keys in his right pants pocket, and his wallet in his back pocket. McDonald turned his body sideways to pass through the cars parked along the divider.484
McDonald had crossed the first lane of Parsons Boulevard and was standing between two parked cars on the far side of the island, getting ready to cross the second lane, when he saw an unmarked red van with plainclothes individuals inside make a u-turn and pull up in front of him, trapping him between two parked cars.485
The driver rolled down the window and, without identifying himself as police, asked McDonald where he was coming from, to which McDonald responded, “Why you stopping me for?” 486 At that point both officers in the van, one of whom was Officer Edward French, and both of whom were white,487 stepped out of the car, identified themselves as police, and began to search McDonald without explanation.488 Officer French told McDonald to remove his hands from his pockets, patted down the outside of McDonald’s pockets, asked McDonald to take out his keys — -which McDonald did — placed his hand inside McDonald’s pocket, and removed a cell phone.489 When McDonald asked why he was being frisked, the officer said he wanted to be sure McDonald did not have a weapon.490
After conducting the frisk, which failed to produce any contraband, Officer French asked McDonald for ID. McDonald obliged and then asked for the officers’ identification. Only Officer French identified himself and gave his shield number.491 McDonald was then permitted to leave. No summons was issued and the entire incident took seven to ten minutes.492
On the UF-250, Officer French listed McDonald’s suspected crime as criminal *632possession of a weapon.493 Officer French did not believe he needed to include the reason he suspected a person of a crime in his memo book- — he believed that the reason for stopping a person was enough.494 At trial, however, Officer French testified that his suspicion was based on crime patterns and a suspicious bulge.
Officer French testified about three crime patterns. First, he made an arrest for armed robbery a month earlier in the general vicinity where he stopped McDonald. Second, he was aware of a robbery pattern somewhere in Queens on the night he stopped McDonald — specifically, a black male holding up commercial establishments. Third, he was aware that a black male had been burglarizing residential establishments in Queens, but could not be more specific about the location of the burglaries.495 The other explanation Officer French gave for his suspicion was his observation that McDonald had his hands in his pockets and was leaning to one side, and had a “suspicious bulge” in his left front pocket.496 I do not believe that Officer French saw a suspicious bulge.497
McDonald believes that he was stopped based on his race because other, non-black individuals — whites or Asians — were coming out of a bowling alley about twenty-five feet from where he was stopped, and none of them were stopped.498
ii. Mixed Findings of Fact and Law
McDonald was stopped when two officers pulled up to him in a police van, trapped him between two cars, and proceeded to question him. The officers made a u-turn and specifically targeted him in a manner that would not have made a reasonable person feel free to simply walk away.
The only articulated bases for the stop were the existence of highly generalized crime patterns involving black males — a month-old armed robbery, a robbery pattern somewhere in Queens and a burglary pattern somewhere in Queens 499 —the fact that McDonald was walking with his hands in his pockets in December and a supposedly suspicious bulge, which turned out to be a cell phone. “[Presence in an area of expected criminal activity, standing alone, is not enough to support a reasonable, particularized suspicion that the person is committing a crime.” 500 Moreover, a crime area defined as the entire borough of Queens is far too broad to contribute to a totality of the circumstances establishing reasonable suspicion, let alone to form the sole basis.501 This, combined with the vague description of “black males” and the entirely unsuspicious act of putting one’s hands in one’s pockets in the wintertime, is a far cry from *633the individualized suspicion of wrongdoing that constitutes reasonable suspicion. Absent any other justification, there was no basis for a Terry stop, and there was certainly no basis to believe that McDonald was armed and dangerous.
I also find that McDonald was stopped because of his race. The only suspect description was “black male,” the street was racially stratified, and other non-black individuals were present and presumably behaving no differently than McDonald — yet only McDonald was stopped. In sum, McDonald was stopped, in violation of the Fourth and Fourteenth Amendments, because he was a black man crossing the street late at night in Queens.
d. Nicholas Peart — August 5, 2006
i. Findings of Fact
Nicholas Peart is a twenty-four-year-old black resident of Harlem. He is the legal guardian of three younger siblings and works for a non-profit organization as an after-school program facilitator.502 On August 5, 2006, around 5:00 a.m., Peart was with his cousin and a friend, both of whom are also black, after celebrating his eighteenth birthday at his sister’s house.503 It was dark out and Peart was wearing light blue basketball shorts and a white — or black and white — tank top. At least one of the men was wearing a hat.504
Peart and his companions were standing in the median at 96th and Broadway when three marked police cars pulled up on 96th Street.505 Approximately five uniformed officers exited the cars with their guns drawn.506 Officer White stated: “Police. Don’t move. Let me see your hands” and, when the men did not immediately show their hands, ordered them to get down on the ground.507 Peart and his companions protested but eventually obeyed the order.508 Once Peart was on the ground, Officer White patted him down over his shorts and in the groin area and buttocks, without consent.509 Officer White also frisked the other two men.510
After frisking the men, Officer White told them to stand up and asked for ID.511 In response to Peart’s inquiries about why they were stopped, Officer White played the radio call three times to show that the men fit the description in the radio call.512 *634The dispatcher’s report described a call coming from a payphone at 96th and Broadway about three black males — one carrying a firearm and wearing blue pants and a black shirt, and two others wearing blue and white tank tops, shorts and red hats — -walking uptown on Broadway toward 98th Street.513 After hearing the radio run, Peart and his companions were free to leave.514 However, they continued to ask why they had been stopped, and stated their belief that the only reason they were stopped was because they were black.515
Following the stop, Peart filed a complaint with the CCRB.516 He declined to state his race in filing the complaint.517 Over a year later, Peart received a letter informing him that the officers had been exonerated.518
The officers who stopped Peart and his friends were responding to a radio run based on a report from an anonymous caller.519 The officers stopped the three men because they resembled the description relayed by the dispatcher and were in the same location from which the call had been made only minutes before.520 Based on the descriptions from the radio run, Officer White believed he had stopped the right people, although no weapons were found.521
Officer White completed three UF-250s after the stop. He checked the boxes for “fits description” and “suspicious bulge.” 522 As reasons for the frisk, he listed violent crime suspected, suspicious bulge, and refusal to comply with directions (relating to the command that the men show their hands and get on the ground).523 On each form Officer White noted that the “bulge in pocket” had in fact been a cell phone. As additional circumstances, Officer White noted “report from victim witness” and “proximity to crime location.” However, nothing about *635the radio run suggested that the caller was a victim rather than a mere observer.524
ii. Mixed Findings of Fact and Law
Peart and his companions were stopped when the officers exited their car with guns drawn. Although it is a close call, I am constrained by controlling Supreme Court law to find that the officers lacked reasonable suspicion to forcibly stop the men. In Florida v. J.L., the Supreme Court held that “an anonymous caller[’s report] that a young black male standing at a particular bus stop and wearing a plaid shirt was carrying a gun” did not establish reasonable suspicion for a stop and frisk.525 The Court held that “reasonable suspicion ... requires that a tip be reliable in its assertion of illegality, not just in its tendency to identify a determinate person.” 526 The Court rejected an exception for reports of firearms, explaining that “[s]ueh an exception would enable any person seeking to harass another to set in motion an intrusive, embarrassing police search of the targeted person simply by placing an anonymous call falsely reporting the target’s unlawful carriage of a gun.” 527 Florida v. J.L. makes clear that a suspect description from an anonymous caller cannot by itself justify the forcible stop and frisk of Peart and his companions. Although the Second Circuit held that a 911 call reporting that an assault (or any crime for that matter) was in progress would require less corroboration, the calls here did not indicate that an assault was in progress.528
The only additional factor cited as the basis for the stop was the bulges in the men’s waistband area.529 The only items recovered from the frisk were cell phones. Because the men were wearing minimal clothing, it would be difficult to mistake a cell phone for a firearm. Moreover, to conclude that the bulge created by the now ubiquitous cell phone can provide the additional corroboration necessary to justify an invasive stop and frisk would eliminate the corroboration requirement entirely.530
I do not find credible the assertion that the officers saw any suspicious bulges that would corroborate the anonymous caller’s statement that the men stopped were armed. Rather, I find that they stopped the men solely on the basis of the description in the radio run. No other factor establishes reasonable suspicion for the highly intrusive stop. Because there was no legally sufficient justification for the stop, the frisk was also unconstitutional,
e. Nicholas Peart — April 13, 2011 Stop531
*636i. Findings of Fact
On April 13, 2011, around 11:00 p.m., Peart was walking on 144th Street between Lenox and Seventh Avenue — the block on which he resides — on his way to the corner store. Peart was wearing sneakers, jeans and a red hooded sweatshirt. He was sending a text message while walking when two uniformed officers appeared directly in front of him. One officer was white, shorter than Peart, and wore glasses (“Officer A”). The other officer was roughly Peart’s height and had salt-and-pepper hair (“Officer B”).532
One of the officers took Peart’s cell phone and instructed Peart to put his hands up against the wall of a church. Officer A patted Peart down outside his clothing over his entire body and put his hands in his pockets. Officer B also put his hands in Peart’s pockets, removed Peart’s keys and wallet, and searched the wallet for ID. Officer B did not ask permission to search the wallet and Peart did not consent.533 During the search Peart asked, “why is this happening?” In response to questions about what building he was coming from, he explained that he was coming from his apartment in the Frederick Samuel House, which is a NYCHA building.534
Officer A then grabbed Peart’s sweatshirt with his fist near Peart’s chest area and handcuffed him.535 Officer B, who had Peart’s keys, asked which key opened Peart’s door, and Peart identified the key in order to prove that he lived where he said he did. He did not give Officer B permission to enter the apartment, but Officer B entered the building and remained for about five minutes.536
While Officer B was in the building, Officer A, who was still holding Peart’s sweatshirt, placed Peart, still handcuffed, in the back of an unmarked police vehicle parked in front of the church. Officer A removed Peart’s sneakers, patted down his socks and asked Peart if he had weed on him. Peart said he did not.537 Eventually, Officer B came out of Peart’s building. The officers opened the car, let Peart out, removed the handcuffs, and returned his keys, phone and wallet.538 The officers explained that Peart fit the description of someone who had been ringing a doorbell at the Frederick Samuel House.539 Peart was then free to go.
ii. Mixed Findings of Fact and Law
Peart was stopped when the officers blocked his path and told him to put his arms against the wall. The stated reason for the stop was that he fit a suspect description of someone who was ringing doorbells in NYCHA housing. While I cannot know from Peart’s testimony whether the description he allegedly fit *637was sufficiently detailed to form the basis for reasonable suspicion,540 the stop the officers conducted was not justified by the circumstances.541 The officers had every right to ask Peart whether he lived in the Frederick Samuel House — a question he could have answered easily. Instead, they forced him up against a wall and handcuffed him.
The officers violated Peart’s Fourth Amendment rights by frisking him, going into his pockets and searching his wallet.542 The officers further abused their authority — and Peart’s rights — when Officer B took Peart’s keys and entered his apartment without permission while Officer A continued to search Peart and question him about drugs.
f. Ian Provost
i. Findings of Fact
Ian Provost is a forty-two-year-old black male who currently resides in North Carolina but lived in Queens from 1978 to 2011.543 On November 24, 2009, Provost was at his girlfriend’s apartment in the Seth Low Houses at 365 Saekman Street in Brooklyn, which is a NYCHA housing project with a high crime rate.544 Provost had a key to access the apartment and the front door lock was broken.545 That day Provost had been doing odd jobs around his girlfriend’s apartment, which involved tools including a knife with a four-inch blade.546
Provost left his girlfriend’s apartment around 2:15 p.m. to get food at a restaurant across the street on Belmont Avenue. He was wearing jeans, a hooded sweatshirt and a down jacket, which did not cover his back pockets, and was carrying the knife in his right back pocket.547 As he reached the corner of Belmont and Saekman, Provost saw two uniformed police officers, since identified as Jonathan Rothenberg and David Furman.548 After he passed them, Officer Rothenberg said “excuse me,” and Provost stopped and turned around.549 The officers asked if Provost was from around there, where he was go*638ing, and where he was coming from. Provost responded that he was from Queens and that he was coming from a friend’s house.550 When Officer Rothenberg asked where he was going, Provost responded that it was not the officer’s business which led to an argument about whether the officer had the right to question Provost’s comings and goings.551 Provost said, “you have no reason to stop me. This is harassment.” 552 Officer Rothenberg told Provost he was being stopped for criminal trespass.553
Provost tried to use his cell phone and Officer Rothenberg commanded him not to. When Provost asked why, Officer Rothenberg said that he did not like people using their cell phones when he was talking to them.554 When Provost attempted to use his cell phone a second time, Officer Rothenberg grabbed his right hand, which was holding the cell phone, handcuffed him, and pushed him up against the fence.555 As he was being handcuffed, Provost repeated Officer Rothenberg’s name and badge number so he could remember it to file a complaint. Provost also yelled out to his girlfriend in hopes that she would hear him from her apartment.556
After Provost was handcuffed, Officer Rothenberg frisked him.557 Provost informed Officer Rothenberg that he had a knife and told him which pocket it was in.558 After retrieving the knife from Provost’s pocket, Officer Rothenberg called Sergeant Houlahan.559 Officer Rothenberg then searched Provost’s person and looked through Provost’s cell phone.560
When Sergeant Houlahan arrived, he felt it was unsafe to stay on the street where crowds sometimes gathered and became violent, so he instructed the officers to put Provost in his car and then drove him to the precinct.561 After the officers determined that the knife was not a gravity knife, Provost was given a summons for carrying a knife with a blade exceeding four inches and a summons for disorderly conduct for being loud and boisterous.562
Officer Rothenberg’s memo book entry for the stop reflected that he had stopped Provost for possible criminal trespass, but did not include any information about the circumstances leading to the stop.563 Roth*639enberg had observed Provost going in and out of the Seth Low Buildings and did not observe him use a key, but Rothenberg acknowledged that he did not have reasonable suspicion to believe that Provost was engaged in criminal trespass.564 Officer Rothenberg testified that he observed an inch or two of a knife handle sticking out of Provost’s pocket when Provost approached the corner where he and another officer were standing, and that this was the reason for his stop.565
Provost filed a CCRB complaint after the incident and was interviewed. The complaint was not substantiated.566 Provost does not believe that the officers saw the knife before he informed them that he was carrying it, which occurred after the stop but before the frisk.567
ii. Mixed Findings of Fact and Law
Provost was stopped when Officer Rothenberg informed him that he was being stopped on suspicion of criminal trespass, and possibly earlier. Whether the stop was unlawful turns on whether I credit Officer Rothenberg’s testimony that he stopped Provost because he saw the knife in his right back pants pocket. I do not. While it is plausible that an officer would delay handcuffing and frisking an individual suspected of possessing a gravity knife, the testimony suggests that Officer Rothenberg had not, in fact, seen the knife when he first stopped Provost. In particular, the fact that Officer Rothenberg wrote in his memo book — the only contemporaneous record of the stop — that he stopped Provost for possible criminal trespass rather than not possession of a weapon, suggests that the knife was not the original reason for the stop.568
In light of this conclusion, there was no basis to stop Provost on suspicion of criminal trespass. Going in and out of his girlfriend’s building is not in and of itself suspicious behavior and, without more, does not rise to the level of reasonable suspicion. Handcuffing and then frisking Provost was also unreasonable. Although Officer Rothenberg’s actions may have been influenced by the fact that he had been assaulted in the area on a prior occasion, nothing in Provost’s actions — arguing about his right not to disclose where he was coming from, reaching for his cell phone while stating that he was reaching for his cell phone — suggested that he presented a sufficient threat to warrant handcuffing him. The frisk was unreasonable for the same reason, particularly in light of the fact that Provost was already handcuffed and could not have harmed Officer Rothenberg or anyone else.569
*640g. David Ourlicht — January 30, 2008 Stop
i. Findings of Fact
David Ourlicht is a twenty-five-year-old male of mixed black and white heritage who grew up and currently lives in Manhattan. At the time of his testimony Ourlicht was applying for admission to law school.570 On January 30, 2008, Ourlicht was enrolled at St. John’s University in Queens. Around 2:00 p.m., he left school and walked his girlfriend to her job. Ourlicht then began walking north on 164th Street to a deli near his dorm. He was wearing a black down Marmot jacket, a sweatshirt, jeans and sneakers. The jacket had six pockets which held Ourlicht’s keys, cell phone, wallet, passport, ipod, pens, and a five-subject notebook. Most of the items were in the jacket’s interior pockets, but the notebook lay flat in one of the front external pockets, with about twenty-five percent sticking out of the pocket.571
As Ourlicht was walking up 164th Street, he saw a uniformed officer, Christopher Moran,572 on a police scooter drive past him from behind. They made brief eye contact but Ourlicht kept walking. When Ourlicht reached the intersection where Officer Moran’s scooter was stopped, Officer Moran asked what Ourlicht was doing in the area and where he was going.573 Ourlicht did not feel free to leave.574
Officer Moran left the scooter and asked Ourlicht for ID, to which Ourlicht responded, “why are you stopping me?” Officer Moran asked whether Ourlicht went to school around there, and Ourlicht said, “why are you asking me this?” Officer Moran again asked for ID and Ourlicht asked, “why do you need to see ID? What did I do?” in an irritated way. Officer Moran responded that it looked like Ourlicht had a gun on him, and proceeded to pat down his waist area.575 Officer Moran did not reach into Ourlicht’s pockets.576
As soon as Officer Moran told Ourlicht he suspected him of having a gun, Ourlicht asked if he could give him his ID, and then reached into the inside breast pocket of his jacket and handed Officer Moran his passport and his St. John’s student ID.577 Officer Moran recorded Ourlicht’s information at which point Ourlicht said, “now that you have my information do you mind if I take down yours?” Officer Moran said sure, and Ourlicht made clear that he was going to reach into his pocket to get a pen and paper and began to write down Officer *641Moran’s badge number, nameplate, and scooter number.578
Officer Moran had radioed for backup and a patrol car pulled up.579 Officer Moran then said, “okay now you’re going to get the full treatment, get against the wall.” 580 Two uniformed police officers got out of the car.581 Ourlicht faced the wall with his hands behind his head and the officers proceeded to pull everything out of his jacket pockets and reached into his jeans pockets, which were empty.582 The officers instructed Ourlicht to sit on the ground, which he did, while Officer Moran returned to the scooter with Ourlicht’s ID.583
Officer Moran returned and asked Ourlicht for his address, which Ourlicht provided, and Officer Moran accused him of lying. Ourlicht then provided his mailing address, which was different from his residence — a college dorm.584 Moran wrote Ourlicht a ticket for disorderly conduct. When Ourlicht learned what the ticket was for he said “that’s fucked up” and “I’m going to fight this.” Ourlicht was then free to go.585 The summons was dismissed. Ourlicht’s mother filed a CCRB complaint regarding the stop.586
Officer Moran observed Ourlicht for no more than two minutes — as he approached Ourlicht on the scooter and after he passed him, from his rearview mirror.587 Officer Moran believed Ourlicht was “blading” the right side of his body in order to protect something in his right waist area that was preventing him from taking normal steps.588 Officer Moran claimed he saw an object running from Ourlicht’s hip along his ribs. Based on that and the way Ourlicht was walking, Officer Moran decided to stop him.589
Officer Moran completed a UF-250 in connection with the stop of Ourlicht. He checked the “suspicious bulge” box but did not identify what the bulge ultimately was.590 Officer Moran noted in his memo book that he stopped a male around 2:15 p.m. based on a suspicious bulge, and issued a summons. No other details about the stop were recorded.591
ii. Mixed Findings of Fact and Law
Ourlicht was stopped when Officer Moran confronted him on the sidewalk and began questioning him — and certainly when the men began to argue about Officer Moran’s authority to demand information about Ourlicht’s comings and goings. The only articulated basis for the stop was the “blading.” Even if Officer Moran saw Ourlicht walking strangely because he had a five-subject notebook in his pocket, that is insufficient to form a “reasonable, particularized suspicion that the *642person is committing a crime.” 592 Nothing else about the circumstances provided added basis for suspicion. It was daytime and Ourlicht made eye contact with Officer Moran, and did not attempt to evade his presence.
I also do not find that Officer Moran reasonably believed that Ourlicht was armed and dangerous. The five-subject notebook, the only item that could have been the object that supposedly caused Officer Moran to believe Ourlicht was armed, could not be confused for a gun, and, in fact, would be easily identifiable as a notebook.593 Therefore, I also find that the first frisk was unreasonable.
Even if the first frisk were justified, the search when the other officers arrived was not justified because Moran had already frisked Ourlicht and found nothing. Nothing that Ourlicht did after the initial stop justified Moran’s suspicion that Ourlicht was hiding “something,” much less that he was hiding a weapon. Terry’s authorization of “strictly circumscribed” protective searches contemplated that evidence of criminal conduct might be missed, but concluded that the Constitution did not permit a more intrusive search absent probable cause.594
h. Clive Lino — February 5, 2008 Stop
i. Findings of Fact
Clive Lino is a thirty-two-year-old black resident of the Bronx. He works as a social worker at a non-profit faith-based organization. Lino is about five foot ten and weighs about 175 pounds.595
On February 5, Officer Brian Kovall, who is white, and Officer Edward Arias, who is black, learned during roll call of a robbery pattern involving two black males, one wearing a beige or yellow coat and the other wearing a blue or black coat, committing gunpoint robberies in the vicinity of a check-cashing location near 103rd Street and Lexington Avenue, a high crime area.596 The height range for the suspects was five foot six to six foot two, the weight range was 170 to 200 pounds, and the age range was mid-twenties.597 Officer Kovall watched a video of the two suspects running, which was taken around noon on January 30.598
*643Around 8:00 p.m. on February 5, Lino and his Mend James went to pick up takeout from a Chinese restaurant at 103rd Street and Lexington Avenue. At the time, Lino lived just two blocks from the restaurant. Lino and James ordered food and were waiting outside the restaurant facing the street.599 Lino was wearing a tan State Property-brand jacket and James was wearing the same jacket in a greenish color.600
As Lino and James waited outside for their food, Officers Kovall and Arias approached and ordered Lino and James to take their hands out of their pockets, which they did after several requests.601 The officers asked the men what they were doing on the corner, where they were going, where they were coming from, where they lived, and if they had ID. Lino said he did not have ID because he had just come from his apartment to get food and was going right back there.602 Lino and James were not free to leave.603
The officers informed Lino and James that they were stopped because they fit the description of armed robbery suspects. Officer Arias stated that they had orders to stop anyone on that particular corner whenever they felt like it.604 Officer Arias Msked Lino’s pockets and waist but did not reach inside his pockets.605 The officers obtained Lino and James’s names and addresses.
The officers called their supervising Lieutenant to come down and confirm whether these men were the robbery suspects, which took five to ten minutes.606 During that time, Officer Kovall permitted Lino to enter the restaurant and get the food because Officer Kovall was “satisfied that there were absolutely no weapons on either individual.”607
After Lino returned with the food, Lieutenant Gaglio and two other officers arrived in plainclothes. Lino knew they were officers because “they spoke to the officers who were already there, and they were able to have access to us.” 608 After asking the same questions that Officers Kovall and Arias had asked, Lieutenant Gaglio told Officer Kovall and Arias that these were not the men.609 Lieutenant Gaglio told Officers Kovall and Arias to run the men’s names. While they were doing so, one of the plainclothes officers Msked Lino again.610
*644Officers Kovall and Arias returned and reported a summons that Lino and James had received in January. The plainclothes officers left and Lino asked Kovall and Arias for their badge numbers. Officer Kovall provided his name and badge number, while Officer Arias just walked away.611 Lino was then free to go. He did not receive a summons and was not arrested.
Following the stop, Lino filed a complaint with the CCRB, which was unsubstantiated.612 Lino believed he and James were stopped because of their race.613
The officers stopped the men because their coats matched the description in the robbery pattern — one light coat and one dark coat. In addition, the officers observed Lino and James standing outside on a cold night in the vicinity of the robberies, and they were still there after the officers circled the block and returned.614 Officer Arias frisked Lino because the suspected crime was armed robbery.615
Officer Kovall filled out a UF-250 after stopping Lino. Under circumstances that led to the stop he checked “Fits Description”; “Area Has High Incidence Of Reported Offense Of Type Under Investigation”; and “Time Of Day, Day Of Week, Season Corresponding To Reports Of Criminal Activity” because of “Ongoing Investigations, e.g. Robbery Pattern.” As the reason for the frisk, Officer Kovall checked “Violent Crime Suspected” and “Knowledge Of Suspect’s Prior Criminal Violent Behavior/Use of Force/Use of Weapon,” to refer to the fact that he was looking for an armed robbery suspect. “Refusal To Comply With Officer’s Directions” was also checked, in reference to Officer Kovall’s initial instruction to Lino to remove his hands from his pockets.616
ii. Mixed Findings of Fact and Law
Lino and James were stopped when the officers approached and told them to remove their hands from then-pockets. Although it is a close question, I find that the officers lacked reasonable suspicion to forcibly stop Lino and James. The suspect description alone — two black men, of average height and weight, one wearing a light jacket and one wearing dark — is too generic to form the basis for reasonable suspicion, especially given the passage of time between the commission of the armed robberies and the stop of Lino and James.617
The act of standing outside in the cold near the check-cashing location for the amount of time it took the officers to circle the block does not raise the totality of the circumstances to reasonable suspicion justifying a forcible stop of Lino and James.618 *645No other circumstances provided additional cause for suspicion. Furthermore, Officer Arias told Lino that the officers had orders to stop anyone on that corner whenever they felt like it. Such an order ignores the requirement of individualized, articulable suspicion, and Officer Arias’ reference to it discredits the officers’ assertions that they had the requisite suspicion in this instance. Thus, the initial stop of Lino violated his Fourth Amendment rights.619
I note that the officers could easily have observed for a few minutes longer to determine whether there was an innocent explanation for this conduct — namely obtaining the food — at no cost to their safety or law enforcement objectives.620 Moreover, they would have been justified in approaching the men and asking them some questions — the equivalent of a De-Bour level one stop. But the intrusion here was considerably more severe and included a frisk of the men’s persons.621 The second frisk by the plainclothes officers further violated Lino’s rights, as there was clearly no threat at that point that the men were armed and dangerous.
i. Lalit Clarkson
i. Findings of Fact
Lalit Clarkson is a thirty-one-year-old black male who works as a union organizer and lives in New Jersey, but visits New York frequently.622 When he was stopped in January, 2006, Clarkson worked as a teacher’s assistant at Grand Concourse Academy, a school located at 169th Street in the Bronx. Clarkson was returning to the school after picking up lunch at a Subway around 1:00 p.m., wearing slacks, a tie, and a collared shirt.623 Clarkson entered a bodega on the corner of 169th and Walton, across from the school, holding a clear Subway bag containing a sandwich. He saw two plainclothes officers, one white and one Hispanic, standing in the back of the bodega, and assumed they were police because of the way they were standing and his experience in the neighborhood. Clarkson did not interact with the officers *646in the bodega. He purchased a food item, put it in his pocket and left the store.624
As he was about to cross the street, he heard a voice say, “hey” and he turned around. The white officer said, “come over here, can I talk to you,” and Clarkson went over. The officers showed their badges and identified themselves as police. The officers came closer to Clarkson so that eventually his back was against the bodega wall and the officers were standing between Clarkson and the street.625
The officers told Clarkson they had seen him walk past a building down the block that they knew to be a drug building. Clarkson had walked past the building because it was on the route he took back to school, but had not stopped or spoken to anyone. The officers asked twice if Clark-son had any contraband on him, and he said he did not. Then they asked: “if I go in your pockets you don’t have anything on you?” Again Clarkson said no. He did not consent to a search and the officers did not search him. The officers left after Clarkson said a third time that he had no contraband, and did not consent to be searched. The stop lasted a few minutes.626 At no point did Clarkson feel free to leave.627
ii. Mixed Findings of Fact and Law
Clarkson was stopped when the officers called him over and surrounded him with his back against the wall. Although merely blocking a means of egress may not constitute a seizure where the police do not actually prevent a person from leaving,628 the act of surrounding an individual on the street against a wall is an intentional “assertion of authority to restrain a person’s freedom of movement” sufficient to constitute a seizure.629 Clark-son was not required to attempt to push past the officers in order to test whether or not he was, in fact, free to leave. Because the officers lacked reasonable suspicion to stop Clarkson where all he did was walk past a building known to be associated with drugs, this stop violated the Fourth Amendment.630
2. Unconstitutional Frisk Only
a. Dominique Sindayiganza
i. Findings of Fact
Dominique Sindayiganza is a middle-aged, black male who resides in Queens with his wife and two daughters. In 2010, Sindayiganza worked at a non-profit organization in Manhattan.631 On February 12, 2010, Sindayiganza left his office at Lexington Avenue and 25th Street around 5:30 p.m. carrying a backpack and wearing dark blue rain pants and a green winter jacket. After running an errand in Union Square, he was walking along Broadway toward the F train at 14th Street and Sixth Avenue to go home. He entered Petco on the corner of Broadway and East 17th Street thinking it was a store that sold children’s clothes. He quickly real*647ized it was a pet store, and as he started to exit he heard someone say: “this is the guy.” Four young white male uniformed officers, including Luke White and Sean Gillespie, surrounded him and forcibly escorted him out of the store.632
Once outside, the officers surrounded Sindayiganza on the sidewalk and aggressively questioned him about what he was doing there, where he came from, and whether he was armed. The officers told him that a lady had identified him as the man who was following her and asking her for money.633 Sindayiganza replied that he was an educator, that he was buying supplies for a field trip and then going home to Queens, and that he had not followed any lady. He asked to see the woman so he could clear his name and was told he could not.634 At some point during this initial interaction, Officer White frisked Sindayiganza.635
Officer White asked Sindayiganza for ID and then went into Petco to see if the woman could identify Sindayiganza as the harasser. The other officers continued to aggressively question Sindayiganza.636 The woman identified Sindayiganza and stated that she just wanted him to leave.637 Officer White went back outside and told Sindayiganza that he could go but that he had to walk north up Broadway. Sindayiganza asked if he could go to the 4 train or the F train, both of which were south, and was told he could not. Sindayiganza asked indignantly why, if he was free to go, he could not go to the train that would take him directly home.638 He was speaking loudly at this point, but was not yelling.639
Officer White went back inside Petco and conferred with the woman who said she wanted Sindayiganza arrested. Officer White then told Sindayiganza that he was going to teach him a lesson and told him to put his hands against the wall, made him take off his backpack, handcuffed him, and forced him to sit down on the sidewalk.640 Another officer looked *648through Sindayiganza’s backpack and removed items, and opened his jacket and searched his pockets, all without consent. When Sindayiganza asked why he was being arrested, the officers told him it was for “excessive panhandling.” Sindayiganza was then taken in a police car to the precinct and given a summons for disorderly conduct, which was never prosecuted.641 Sindayiganza submitted a CCRB complaint online a day or two after the event and provided a sworn statement seven months later.642
Officer White believed he had reasonable suspicion to stop Sindayiganza because he matched the woman’s description — tall, light-skinned black male with dark hair, a big backpack, glasses, a green jacket, and green pants — and was in close proximity to the location of the incident.643 Officer White believed he had probable cause to arrest Sindayiganza for aggravated harassment based on the woman’s allegations and identification.644 He further believed that once he handcuffed Sindayiganza, he had the right to search him incident to arrest.645 No arrest report was filled out and only a summons for disorderly conduct was issued.646
ii. Mixed Findings of Law and Fact
Sindayiganza was stopped when the police surrounded him and escorted him outside. The police had reasonable suspicion to forcibly stop Sindayiganza for suspected harassment because he matched a specific description provided by an identified victim, and was in close proximity to the reported harassment just minutes after it allegedly occurred.647
There was, however, no basis to frisk Sindayiganza. His stop was based on a woman’s report that a man had been following her and asking her repeatedly for money, which caused her alarm.648 The woman never said she believed her harasser was armed, or that she had been physically threatened.
The frisk cannot be justified as a search incident to arrest that preceded the arrest. A,s a preliminary matter, the officers almost certainly lacked probable cause under New York law to arrest Sindayiganza. Based on the woman’s allegations, Officer White at most had reasonable suspicion that Sindayiganza had committed harassment in the second degree, a violation under New York law.649 New York law prohibits arrest based on see*649ond-degree harassment that does not occur in the arresting officer’s presence.650
Even if probable cause for an arrest had existed, when Officer White frisked Sindayiganza he had no intention of arresting him; rather he planned to let Sindayiganza leave, as long as he walked north. A frisk cannot be justified after the fact as a search incident to arrest, where there is no intent to arrest at the time the frisk is conducted.651
b. David Floyd — April 20, 2007 Stop652
i. Findings of Fact
David Floyd is a thirty-three-year-old black male who lived in the Bronx from 2001 until 2010, when he left to attend medical school.653 On or around April 20, 2007, a Friday, Floyd was coming from the subway walking towards his home at 1359 Beach Avenue in the Bronx, wearing jeans and sneakers, and carrying his wallet, keys and cell phone in his pocket.654 As he crossed East 172nd Street, Floyd saw two police officers about a block-and-a-half away interacting with another individual. The officers then got into a van and Floyd continued walking down Beach Avenue toward his home. Shortly thereafter, the van pulled up to Floyd and the officer in the driver’s seat said, “Excuse me, may I speak to you, sir?” Floyd immediately stopped walking.655
Three uniformed officers, one Latino male, one white male and one female, exited the vehicle and one asked Floyd for ID.656 Floyd asked whether he had to give *650ID, but did not feel free to refuse.657 After he produced the ID, Floyd reached into his pocket to get his cell phone or a pen to write down the officers’ identification. The white male officer jumped toward him and Floyd immediately stopped and put his hands up and said, “it’s a cellphone.” The officer said that it made him nervous when people put their hands in their pockets. The officer asked Floyd if he had a weapon, and Floyd said he did not and that he did not consent to being searched. The officer proceeded to pat down Floyd’s entire body including pushing the cell phone, a BlackBerry, up out of Floyd’s pocket with his finger.658
The frisk lasted about thirty seconds and then the officer holding Floyd’s ID told him it was illegal for him not to have a New York City license — Floyd’s ID was out-of-state. The officers then got back in the van. Floyd asked for names and badge numbers and two of the officers identified themselves as Rodriguez and Goodman and provided badge numbers. The officers then left.659 Floyd did not file a CCRB complaint in connection with this stop.660
ii. Mixed Finding of Fact and Law
Without testimony from the stopping officers, I have insufficient evidence to determine whether they had reasonable suspicion to approach Floyd and question him. However, I credit Floyd’s testimony and find that there was no basis for the frisk. The officers did not appear to believe that Floyd was armed and dangerous when they approached him, because they did not immediately frisk him. Nothing in Floyd’s behavior gave the officers reason to reconsider whether Floyd was “armed and dangerous.” He reached for his cell phone and promptly identified it as such. The officer asked Floyd if he had a weapon and Floyd said he did not. I find that there was no credible basis for believing Floyd was armed and dangerous. Therefore, the frisk was unconstitutional.
c. David Floyd — February 27, 2008 Stop
i. Findings of Fact
On February 27, 2008, Floyd resided on Beach Avenue in a three-family home with a separate cottage. Floyd’s Godmother owned the property and lived on the top floor, and tenants occupied the ground floor and the basement. Floyd lived in the cottage. Around 3:00 p.m., Floyd left the cottage carrying a backpack, with his wallet, cell phone, keys, and some change in his pocket. Before Floyd got to the street, the basement tenant, also a black male, told Floyd that he was locked out of his apartment and asked for Floyd’s assistance because Floyd had access to the spare keys.661
Floyd retrieved seven to ten keys on separate key rings from his Godmother’s apartment and went to the door of the basement apartment with the tenant. Because the keys were not marked, both Floyd and the tenant tried five or six different keys for a minute or two.662 At *651that point, three plainclothes officers, since identified as Officers Cormac Joyce and Eric Hernandez, and Sergeant James Kelly, approached and told Floyd and the tenant to stop what they were doing and put their hands up.663 Floyd obeyed.664 Officer Joyce patted Floyd down and searched Floyd’s pockets without his consent.665 He did not remove anything from Floyd’s pockets.666
After frisking the men, the officers remained calm throughout the encounter.667 The officers asked the men for ID, and Floyd showed his Louisiana drivers’ license. The tenant did not have ID on him. After additional inquiries, Floyd produced an electric bill with his name and address on it, and the tenant went inside his apartment and got ID.668 Floyd asked why he had been stopped and the officers informed him that there had been a pattern of burglaries in the area. Floyd asked for the officers’ names and badge numbers and the officers provided them. The officers then left.669
Earlier that day the officers had been patrolling Beach Avenue in response to reports of robberies and burglaries of private homes in the area — specifically the area of the 43rd Precinct near the Cross-Bronx Expressway. Proximity to the Cross-Bronx Expressway was significant because it provided easy access for a vehicle to get away from the area quickly.670 The majority of the burglaries in the pattern identified during trial occurred in January 2008, with the last occurring on February 2.671
The officers observed Floyd and the tenant for about two minutes before approaching them.672 Sergeant Kelly observed them playing with the door knob, and also saw a bag on the ground next to Floyd.673 When he approached the men, he saw that they were trying numerous keys, which was consistent with his belief that they might be burglars, because burglars sometimes have master keys.674 The basis for the frisk was the belief that Floyd and the tenant were in the process of committing a violent felony.675
Officer Joyce filled out a UF-250 in connection with this stop. He checked the box for Furtive Movements based on the jostling of the doorknob and the keys.676 *652He also checked time of day corresponding to criminal activity.677
ii. Mixed Findings of Fact and Law
Floyd was stopped when the officers told him to stop what he was doing and raise his hands. The totality of the circumstances established reasonable suspicion to stop Floyd and his neighbor. The officers observed the men jostling the door knob and trying numerous keys in the door, and also observed a backpack on the ground. Beach Avenue is near the Cross Bronx Expressway, which makes it a target for burglary. The officers also had knowledge of a specific burglary pattern in the area of Beach Avenue. Although the last reported burglary was over three weeks before Floyd was stopped, the totality of the circumstances just recounted justified the officers’ belief that the men might be in the process of committing a daytime burglary.
Furthermore, because burglary is often a violent crime, the officers were justified in promptly telling the men to put their hands up and frisking their outer garments for weapons before further investigating.678 The stop was also reasonable in duration because Floyd’s ID was out of state and the basement tenant did not have ID on him. Therefore, the officers were justified in continuing to investigate the possibility that the men were burglars. However, Officer Joyce did not testify that he felt anything that might be a weapon or anything that was clearly contraband. Therefore, Officer Joyce violated Floyd’s Fourth Amendment rights when he felt inside his pockets.
d. Clive Lino — February 24, 2011 Stop
i. Findings of Fact
On the night of February 24, 2011, Lino was at a party at his mother’s apartment at 102nd Street and Third Avenue. He left the apartment to take the subway home, wearing a red leather Pelle Pelle brand jacket and carrying a white plastic grocery bag full of Tupperware containing leftover food.679
Lino entered the subway at 103rd Street and Lexington Avenue and noticed two male police officers, since identified as Officer Daniel Leek, who is white, and Officer Edgar Figueroa, who is Hispanic.680 Lino went down to the uptown platform.681 As an uptown-bound train was pulling into the station, the officers entered the uptown platform. Lino stepped back on the platform to let them pass but instead the officers surrounded him.682
Officer Figueroa immediately put his hand in Lino’s right jacket pocket, and Lino pushed it away while stepping aside. Lino asked what the problem was, to *653which the officers responded that he needed to wait with them. Lino missed the train. Lino asked why he was being stopped and stated his belief that his race was the reason. Officer Leek said, “If you shut the fuck up, we’ll tell you why we stopped you.”683
The officers told Lino to put the bag of food on the platform, which he refused to do because it was filthy. Officer Figueroa said “just put the [fuck]ing bag down” and reached for the bag. Lino placed the bag on the bench. The officers asked for ID, which Lino produced.684
Officer Leek asked Lino if he had anything on him that he shouldn’t have and Lino said no. Officer Leek then said, “Do you mind if we check?” 685 While Lino may have initially consented, he clearly did not agree to be searched and at one point said, “you can’t search me.” 686 Yet Officer Leek patted his waist and front pockets, and Officer Figueroa reached into his back pockets. When Lino looked back at what Officer Figueroa was doing, Officer Figueroa asked if Lino had a “fucking problem.” Lino said he did have a problem because the officer “was in his pockets.” 687 He had not yet been told why he had been stopped. Officer Leek eventually informed Lino that he was being stopped because there were reports of a shooting suspect wearing a jacket similar to Lino’s.688
After frisking Lino, the officers forced Lino to go upstairs with them without explaining why. At no point was Lino free to walk away.689 Outside the subway station, Officer Figueroa stayed with Lino while Officer Leek took Lino’s ID to the police van where his Sergeant was waiting. Officer Leek attempted, unsuccessfully, to locate the wanted poster describing the shooting suspect. He did not run Lino’s ID because Lino had not, to Officer Leek’s knowledge, committed a crime.690 Officer Leek returned and the officers walked Lino back down to the platform, without his having to pay another fare, and returned his ID.691
Lino never received a ticket or summons. The interaction lasted about twenty minutes.692 Following the interaction, Lino filed a complaint with the CCRB. The CCRB complaint was substantiated with respect to the stop and charges were recommended against the officers. The allegations regarding the search and the offi*654cers’ use of rude or obscene language were unsubstantiated.693
The officers stopped Lino because he matched a description of a homicide suspect from a wanted poster given to them by their sergeant at the beginning of their tour that day. The poster stated that the crime occurred on February 10, 2011, two weeks earlier, at 108th Street and Madison Avenue, two avenues and five blocks away from where Lino was stopped. The poster described a black male approximately five foot nine to six feet tall and showed a security photo of the suspect leaving the shooting wearing a red leather Pelle Pelle jacket, and two stock photos of such a jacket.694 From the photo, Officer Leek discerned the suspect’s height, weight, and an age range of eighteen to thirty.695
Officer Leek believed that Lino’s jacket, height and complexion matched the crime scene photo. In addition, he noticed that the jacket was loose-fitting and bulky.696 Officer Leek frisked Lino because he matched the description of a murder suspect, had a bulky jacket that could conceal a weapon, and was in a high crime area.697
ii. Mixed Findings of Fact and Law
Lino was stopped when the officers surrounded him on the subway platform. Although the murder was two weeks old, the distinctiveness of the red leather Pelle Pelle jacket, along with the three-inch height range — which was verifiable from the suspect photo — and the suspect’s race, “afford[ed] a sufficient basis for ‘selective investigative procedures’ visa-vis a universe made up of all [potential suspects] of the crime in question.” 698 It was reasonable to assume that the murder suspect might have lived in the neighborhood and remained or returned within an eight-block radius.
Murder is a violent crime that would justify a frisk. Although the crime was two weeks old, it was reasonable to believe that the suspect in a shooting might still be armed and, if approached and questioned about the murder, would pose a danger to the officers.699 Because the jacket was loose and a train was approaching, the officers could not readily verify whether Lino was the right person or whether he was armed before stopping and frisking him.
Although both the stop and the frisk were justified at their inception, the frisk was not “reasonably related in scope to the justification for [its] initiation.” 700 Officer Figueroa immediately put his *655hands in Lino’s pockets without Lino’s consent, and the officers later conducted a search, to which they acknowledge Lino did not consent. The frisk, particularly in combination with the abusive manner in which the stop was conducted, violated Lino’s Fourth Amendment rights.
e. Deon Dennis
i. Findings of Fact
Deon Dennis was raised in Harlem and resides in South Carolina, but visits New York frequently.701 Around 8:00 p.m. on January 12, 2008, Dennis arrived at the apartment of his then-girlfriend at 122nd Street and Seventh Avenue to help set up for her birthday party, which was scheduled to begin at 11:00 p.m.702 While setting up he drank a beer and some brandy in a plastic cup.703 Later, Dennis went to smoke a cigarette on the sidewalk outside the building on Seventh Avenue, while his girlfriend went to the store. The street was empty. Dennis was wearing a jacket and jeans, and had his cell phone, keys, wallet and cigarettes in his pocket.704
After Dennis finished his cigarette, an NYPD van pulled up and two uniformed officers, Angelica Salmerón and Luis Pichardo, both of whom are Hispanic, approached Dennis.705 Dennis was not free to leave.706 Officer Salmerón pointed to a cup several feet to the right of Dennis and asked if it was his cup. Dennis said it was not. Officer Pichardo asked if Dennis had been drinking, and Dennis said he had been drinking earlier.707 The officers claim that Dennis had a cup in his hands and was drinking from it when they approached him, and that there was a bottle of Hennessy on the sidewalk.708
The officers asked Dennis for ID and he produced his driver’s license from the wallet in his pocket. Officer Salmerón took Dennis’s ID to the van while Officer Pichardo stayed and searched Dennis’s jacket and pants pockets without consent. During the search, Dennis was standing with his wallet in his hand and his hands raised at chest level.709
Officer Salmerón returned from the van having discovered an active warrant against Dennis. The officers then handcuffed Dennis and put him in the back of the van based on the open container violation and the active warrant.710 Dennis’s girlfriend returned and saw Dennis handcuffed.711 No summons was issued and Dennis was never prosecuted for drinking in public. No alcohol was vouchered or put in evidence.712 Neither officer recorded the incident in his or her memo book.713 *656Dennis’s girlfriend filed a CCRB complaint about Dennis’s arrest, and Dennis spoke with a CCRB officer about the incident.714 Dennis believes that he was stopped because of his race, because “only blacks and Hispanics get stopped in Harlem.” 715
ii. Mixed Findings of Fact and Law
Based on the conflicting testimony about what the officers saw when they approached Dennis, and because plaintiffs bear the burden of proof, I cannot find by a preponderance of the evidence that the officers lacked reasonable suspicion that Dennis was violating New York’s open container law.716 However, even if the officers saw Dennis drinking from a plastic cup, they would only have had reasonable suspicion to approach him and investigate further.717 They did not have reasonable suspicion to frisk or search Dennis.
The officers acknowledge that when they took Dennis’s ID to the van, they intended to issue him a summons, not to arrest him. Even assuming the officers were justified in issuing Dennis a summons for violating open container laws, and even with knowledge of an active warrant, there was no basis for Officer Pichardo to frisk, let alone search, Dennis.718 The authority to conduct a protective frisk when issuing a summons for a violation is limited to occasions where officers have reason to believe that the suspect is armed and dangerous.719 Absent any indication that Dennis posed a threat to the officers’ safety, there was no basis to conduct a frisk or search prior to arresting Dennis. Officer Pichardo’s search violated Dennis’s Fourth Amendment rights.720
3. Failure of Proof
a. John Doe Stops of Nicholas Peart in Spring 2008 and February 2010 and David Ourlicht in February and June 2008
In the John Doe stops discussed below, plaintiffs’ testimony did not provide sufficiently detailed credible information for me to find, by a preponderance of the evidence, that the stop and, in some cases, frisk, lacked reasonable suspicion.721 Nicholas Peart testified that he was stopped and frisked in the Spring of 2008 in Brook*657lyn, based , on a burglary pattern. Although the existence of a burglary pattern alone would not provide reasonable suspicion, I cannot conclude that there was not reasonable suspicion to stop and frisk Peart without knowing what additional information the officers had.722 Peart also testified that, in September of 2010, he was stopped and frisked at 144th Street, between Seventh and Eighth Avenues. Because I have no information about the basis for this stop, I cannot find that the stop and frisk lacked reasonable suspicion.723
David Ourlicht testified that on February 21, 2008, he and a friend were stopped and frisked near the subway station at 168th Street in the Bronx.724 1 do not fully credit Ourlicht’s version of the events, and without the officers’ version of what occurred, I cannot find that Ourlicht’s Fourth or Fourteenth Amendment rights were violated. Ourlicht also testified that in June 2008, he was stopped and frisked at an apartment complex at 115th Street and Park Avenue based on reports of a gun in the area.725 Without information about the officers’ knowledge about the gun in the area, I cannot find that the' stop and frisk lacked reasonable suspicion.
b. Kristianna Acevedo Stop
Kristianna Acevedo is a thirty-year-old Hispanic female resident of Staten Island and works as a recruiter of home health aides. In 2007 she lived in Queens.726 On Tuesday, May 29, 2007, Acevedo was walking on 43rd Street in a desolate area when she noticed two men, since identified as Detectives Louis DeMarco and Damian Vizcarrondo, in a minivan.727 Detective DeMarco spoke to Acevedo to obtain information about drug activity.728 Acevedo did not believe the men were police, so she kept walking and then began to run.729
Acevedo stopped at a UPS truck parked up the block.730 The van reversed and stopped, and three officers — Detectives DeMarco, Vizcarrondo, and a female officer, Detective Michele Hawkins — got out, approached Acevedo, and identified themselves as police.731 The officers wanted to assure Acevedo that she was not in danger because there had been recent media reports of individuals impersonating police officers.732 The detectives did not stop Acevedo based on reasonable suspicion.733 After a brief exchange, the officers left.
Acevedo called 911 to report what happened and filed a CCRB complaint.734 The *658CCRB substantiated the charges that the detectives abused their authority by stopping Acevedo, and failing to record the incident in their memo books. As discipline, the officers were docked one vacation day. Detective DeMarco was exonerated of the charges relating to his questioning of Acevedo and the other charges were unsubstantiated.735
Acevedo clearly felt free to leave when the detectives first spoke to her because she continued walking and eventually ran.736 What occurred after the detectives exited the van is unclear. Acevedo’s version of the events is irreconcilable with the officers’ testimony. Although the CCRB found that the officers stopped Acevedo, I did not find her story sufficiently credible to conclude by a preponderance of the evidence that a forcible stop or frisk occurred in violation of her Fourth Amendment rights.737
c. Clive Lino — August 3, 2008
Lino testified about an August 3, 2008 interaction with Officers Jose Colon and Mohamed Hassan in the lobby of his apartment, a NYCHA building at 102nd Street and Third Avenue.738 Lino’s testimony about what occurred is incompatible with the officers’ testimony.739 Because I credit the officers’ testimony, I do not find that Lino’s Fourth or Fourteenth Amendment rights were violated by this interaction.
V. CONCLUSIONS OF LAW
A. The City Is Liable for Violations of Plaintiffs’ Fourth Amendment Rights
Plaintiffs established the City’s liability for the NYPD’s violation of their Fourth Amendment rights under two theories, either of which is adequate under Monell: first, plaintiffs showed that senior officials in the City and at the NYPD were deliberately indifferent to officers conducting unconstitutional stops and frisks; and second, plaintiffs showed that practices resulting in unconstitutional stops and frisks were sufficiently widespread that they had the force of law.
1. Deliberate Indifference
There is no dispute that the primary concern of a police department can and should be combating crime. At the same time, section 1983 limits the lack of concern that any municipal agency may show toward constitutional violations by its employees. The NYPD’s senior officials have violated section 1983 through their deliberate indifference to unconstitutional stops, frisks, and searches. They have received both actual and constructive notice since at least 1999 of widespread Fourth Amendment violations occurring as a result of the NYPD’s stop and frisk practices. Despite this notice, they deliberately maintained and even escalated policies and practices that predictably resulted *659in even more widespread Fourth Amendment violations.740 Moreover, while the NYPD is an acknowledged leader in the use of data collection and analysis to improve the effectiveness of policing, it has hindered the collection of accurate data concerning the constitutionality of its stops, and made no effective use of the limited data that is available. The NYPD has repeatedly turned a blind eye to clear evidence of unconstitutional stops and frisks.
Further evidence of deliberate indifference is found in the City’s current positions as expressed at trial. The City continues to argue that no plaintiff or class member was subjected to an unconstitutional stop or frisk — not Downs, Almonor, McDonald, Sindayiganza, or any of the other plaintiffs.741 The City defends Officer Dang’s stops in the third quarter of 2009 as unproblematic,742 despite the fact that he stopped 120 black people and 0 white people during that period. Officer Dang relied on a routine set of vague and unreliable stop justifications, and in only 5.5% of his stops made an arrest or summons.743 The City also defends the contents of the tape recordings quoted above, arguing that they “provide no basis whatsoever from which any reasonable inference can be drawn that ... pressure for activity existed that drove officers to make unconstitutional stops.” 744 In addition, the City recognizes the impossibility of tracking unconstitutional stops through UF-250s,745 but disclaims the need to develop a better, more adequate system of documentation and review.746 Indeed, the City continues to believe that there is no need to alter the status quo.747 Confronted with the persuasive statistical evidence that stops frequently lack reasonable suspicion, the City argues that even if “18% ... of the 4.43 million stops” were legally unjustified, that “is not necessarily a widespread pattern,” and would not require a remedy.748
Throughout the class period, the need for better supervision, monitoring, training, and discipline to protect against constitutional violations was obvious, but senior officials at the NYPD “ ‘fail[ed] to make meaningful efforts to address the risk of harm to plaintiffs.’ ” 749 Even if “deliberate indifference” were not the standard for liability, it would still perfectly describe the attitude of senior officials at the NYPD toward the risk of officers conducting stops, frisks, and searches in violation of the Fourth Amendment.
2. Widespread Practice
Despite the NYPD’s deliberate failure to collect accurate data regarding stops that violate the Fourth Amendment, there is sufficient evidence of such stops to establish Monell liability based on “practices so persistent and widespread as to *660practically have the force of law.” 750 As described above, the likely number of stops lacking reasonable suspicion was far higher than the roughly 200,000 “apparently unjustified” stops identified by Dr. Fagan. In particular, this conclusion is supported by the number of UF-250s that do not identify a suspected crime (36% of all forms in 2009), the problems inherent in the two most commonly checked stop factors (Furtive Movements and High Crime Area), and the fact that only 6% of all stops result in an arrest for any crime.751 The NYPD’s practice of making stops that lack individualized reasonable suspicion has been so pervasive and persistent as to become not only a part of the NYPD’s standard operating procedure, but a fact of daily life in some New York City neighborhoods.
Likewise, the pervasiveness of unconstitutional frisks was established by the uncontested fact that over half of all people stopped are frisked, while only 1.5% of frisks reveal a weapon, as well as the institutional evidence of inaccurate training regarding when to frisk, testimony by officers who did not know the constitutional standard for a frisk, and anecdotal evidence of routine unconstitutional frisks in this case. “The security of one’s privacy against arbitrary intrusion by the police— which is at the core of the Fourth Amendment — is basic to a free society.” 752 Far too many people in New York City have been deprived of this basic freedom far too often.
B. The City Is Liable for Violations of Plaintiffs’ Fourteenth Amendment Rights
Plaintiffs have established the City’s liability for the NYPD’s violation of plaintiffs’ Fourteenth Amendment rights under two theories, either of which is adequate under Monell. First, plaintiffs showed that the City, through the NYPD, has a policy of indirect racial profiling based on local criminal suspect data. Second, plaintiffs showed that senior officials in the City and at the NYPD have been deliberately indifferent to the intentionally discriminatory application of stop and frisk at the managerial and officer levels.
1. Policy of Indirect Racial Profiling
Throughout this litigation the City has acknowledged and defended the NYPD’s policy of conducting stops based in part on criminal suspect data, of which race is a primary factor. The NYPD implements this policy by emphasizing to officers the importance of stopping “the right people.” In practice, officers are directed, sometimes expressly, to target certain racially defined groups for stops.
“The Constitution prohibits selective enforcement of the law based on considerations such as race.” 753 The Second Circuit has admonished that courts should “not condone racially motivated police behavior” and must “take seriously an allegation of racial profiling.”754 Racial profiling constitutes intentional discrimina*661tion in violation of the Equal Protection Clause if it involves any of the following: an express classification based on race that does not survive strict scrutiny; 755 the application of facially neutral criminal laws or law enforcement policies “in an intentionally discriminatory manner;” 756 or a facially neutral policy that has an adverse effect and was motivated by discriminatory animus.757 The City’s policy of targeting “the right people” for stops clearly violates the Equal Protection Clause under the second method of proof, and, insofar as the use of race is explicit, the first.
a. Intentionally Discriminatory Application of a Facially Neutral Policy
In order to establish an equal protection violation based on an intentionally discriminatory application of a facially neutral policy, plaintiffs “must prove that the defendants’ actions had a discriminatory effect and were motivated by a discriminatory purpose.” 758 In this case, plaintiffs’ statistical evidence of racial disparities in stops is sufficient to show a discriminatory effect.759 In particular, plaintiffs showed that: (1) the NYPD carries out more stops where there are more black and Hispanic residents, even when other relevant variables are held constant; (2) NYPD officers are more likely to stop blacks and Hispanics than whites within precincts and census tracts, even after controlling for other relevant variables; (3) NYPD officers are more likely to use force against blacks and Hispanics than whites, after controlling for other relevant variables; and (4) NYPD officers stop blacks and Hispanics with less justification than whites.760 In addition to their statisti*662cal evidence of a racially disproportionate impact, plaintiffs provided significant anecdotal evidence, such as the stark racial disparities in the UF-250s prepared by Officers Dang and Gonzalez, and the fact that Officer French chose to stop McDonald, rather than similarly situated non-blacks nearby, based in part on generalized crime complaints about black males.
To establish discriminatory intent, plaintiffs must show that those responsible for the profiling did so “at least in part ‘because of,’ not merely ‘in spite of,’ its adverse effects upon” the profiled racial groups.761 Plaintiffs are not required to prove that race was the sole, predominant, or determinative factor in a police enforcement action.762 Nor must the discrimination be based on “ill will, enmity, or hostility.”763
The NYPD has directed officers to target young black and Hispanic men because these groups are heavily represented in criminal suspect data — the reliability of which is questionable 764 — in those areas where the NYPD carries out most of its stops. Under the NYPD’s policy, targeting the “right people” means stopping people in part because of their race. Together with Commissioner Kelly’s statement that the NYPD focuses stop and frisks on young blacks and Hispanics in order to instill in them a fear of being stopped, and other explicit references to race discussed in the next section, there is a sufficient basis for inferring discriminatory intent.765
The fact that the targeted racial groups were identified based on crime victim complaints does not eliminate the discriminatory intent. Just as it would be impermissible for a public housing agency to adopt a facially race-neutral policy of disfavoring applications from any group that is disproportionately subject to tenant complaints, and then apply this policy to disfavor applications from a racially defined group, so it is impermissible for a police department to target its general enforcement practices against racially defined groups based on crime suspect data.
*663b. Express Classification
Plaintiffs have readily established that the NYPD implements its policies regarding stop and frisk in a manner that intentionally discriminates based on race. While it is a closer call, I also conclude that the use of race is sufficiently integral to the policy of targeting “the right people” that the policy depends on express racial classifications. When an officer is directed to target “male blacks 14 to 21” for stops in general based on local crime suspect data — a practice that the City has defended throughout this litigation — the reference to “blacks” is an express racial classification subject to strict scrutiny.766 Chief Esposito’s concession that the NYPD has targeted young blacks and Hispanics for stops confirms that explicit references to race are not limited to a few rogue supervisors.767 The City has not attempted to defend — nor could it defend — the proposition that the targeting of young black males or any other racially defined group for stops is narrowly tailored to achieve a compelling government interest. Because the use of express racial classifications in the City’s policy of indirect racial profiling cannot withstand strict scrutiny, the policy violates the Equal Protection Clause.768
This policy far exceeds the permissible use of race in stopping suspects as set forth in Brown v. City of Oneonta, New York. There, the Second Circuit held that when the police carry out stops as part of a “search[ ] for a particular perpetrator,” the use of racial information from the victim’s description of the suspect is not an express racial classification subject to strict scrutiny.769 The court explained that *664the Oneonta police department’s “policy was to investigate crimes by interviewing the victim, getting a description of the assailant, and seeking out persons who matched that description” and, as such, “was race-neutral on its face.”770
The NYPD’s policy of targeting “the right people” for stops, by contrast, is not directed toward the identification of a specific perpetrator.771 Rather, it is a policy of targeting expressly identified racial groups for stops in general. There is no dispute that it would violate equal protection for a police department to adopt an express policy of targeting members of one race for stops or other enforcement activities— such as an express policy of only pulling over speeding drivers who are Hispanic. Similarly, the following hypothetical police department policy would surely be subject to strict scrutiny, despite its failure to mention any specific race at the outset: “No one is to be stopped except the members of whatever race participated at the highest rate in violent crime during the previous month, based on suspect descriptions.” Such a policy would be especially deserving of strict scrutiny if its drafters knew that the same race would be targeted every month, and managers implementing the policy were responsible for expressly directing officers to stop members of that race. The NYPD’s policy of indirect racial profiling is closer to this hypothetical policy than it is to the race-neutral policy in Brown.
c. Conclusion
Whether through the use of a facially neutral policy applied in a discriminatory manner, or through express racial profiling, targeting young black and Hispanic men for stops based on the alleged criminal conduct of other young black or Hispanic men violates bedrock principles of equality. Two young men in the 81st Precinct who are similarly situated in every way, except that one is black and the other white, are similarly situated for the purposes of equal protection and must be treated alike. Brown establishes the common-sense principle that if a description of a specific criminal suspect includes the fact that the suspect is black, then the police need not focus equal attention on individuals of other races in pursuit of that suspect. But Brown specifically rejects the use of racial profiling as a basis for enforcement activity. The Equal Protection Clause does not sanction treating similarly situated members of different racial groups differently based on racial disparities in crime data. Indeed, such treatment would eviscerate the core guarantees of the Equal Protection Clause. If equal protection means anything, it means that individuals may not be punished or rewarded based on the government’s views regarding their racial group, regardless of the source of those views.772
*6652. Deliberate Indifference
In a case alleging that a municipality bears Monell liability based on senior officials’ deliberate indifference to equal protection violations by subordinates, it is not necessary for plaintiffs to provide direct evidence that the senior officials were motivated by a discriminatory purpose. Rather, it is sufficient if plaintiffs show that: (1) subordinates followed a course of action in part because of its adverse effects on an identifiable group, and (2) senior officials were deliberately indifferent to those adverse effects in such a way that a reasonable inference can be drawn that those officials intended those adverse effects to occur.773
Plaintiffs in this case did provide direct evidence of discriminatory intent, as discussed above. But plaintiffs also showed that senior officials in the City and at the NYPD have been deliberately indifferent to the discriminatory application of stop and frisk at the managerial and officer level such that a reasonable inference of discriminatory intent can be drawn. Despite frequent and ongoing notice of troubling racial disparities in stops,774 the *666NYPD has long shown its lack of concern for racial profiling through the failure of NYPD officials and managers to discuss racial profiling among themselves or at Compstat meetings, and through the numerous failures of supervision, monitoring, training, and discipline discussed above.775 In addition, senior NYPD officials such as Deputy Commissioner Farrell have adopted an attitude of willful blindness toward statistical evidence of racial disparities in stops and stop outcomes.776 During trial this indifference was further demonstrated by many officials’ apparent belief that racial profiling is a myth created by the media,777 as well as by the testimony describing and defending the targeting of “the right people” for stops.778
The City and the NYPD’s highest officials also continue to endorse the unsupportable position that racial profiling cannot exist provided that a stop is based on reasonable suspicion.779 This position is fundamentally inconsistent with the law of equal protection and represents a particu*667larly disconcerting manifestation of indifference. As I have emphasized throughout this section, the Constitution “prohibits selective enforcement of the law based on considerations such as race.” 780 Thus, plaintiffs’ racial discrimination claim does not depend on proof that stops of blacks and Hispanics are suspicionless.781 A police department that has a practice of targeting blacks and Hispanics for pedestrian stops cannot defend itself by showing that all the stopped pedestrians were displaying suspicious behavior. Indeed, the targeting of certain races within the universe of suspicious individuals is especially insidious, because it will increase the likelihood of further enforcement actions against members of those races as compared to other races, which will then increase their representation in crime statistics. Given the NYPD’s policy of basing stops on crime data, these races may then be subjected to even more stops and enforcement, resulting in a self-perpetuating cycle.782
The Equal Protection Clause’s prohibition on selective enforcement means that suspicious blacks and Hispanics may not be treated differently by the police than equally suspicious whites. Individuals of all races engage in suspicious behavior and break the law. Equal protection guarantees that similarly situated individuals of these races will be held to account equally.
VI. CONCLUSION
For the foregoing reasons, the City is liable for the violation of plaintiffs’ Fourth and Fourteenth Amendment rights. In a separate opinion, I will order remedies, including immediate changes to the NYPD’s policies, a joint-remedial process to consider further reforms, and the appointment of an independent monitor to oversee compliance with the remedies ordered in this case. I conclude with a particularly apt quote: “The idea of universal suspicion without individual evidence is what Americans find abhorrent and what black men in America must constantly fight. It is pervasive in policing policies — like stop-and-frisk, and neighborhood watch — regardless of the collateral damage done to the majority of innocents. It’s like burning down a house to rid it of mice.”783
APPENDIX A
Blank OF-250
*668[[Image here]]
. See generally U.S. Const, amend. IV; Terry v. Ohio, 392 U.S. 1, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968).
. See generally U.S. Const, amend. XIV § 1; Whren v. United States, 517 U.S. 806, 813, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996).
. District of Columbia v. Heller, 554 U.S. 570, 636, 128 S.Ct. 2783, 171 L.Ed.2d 637 (2008).
. The law requires plaintiffs to produce evidence that at least some class members have been victims of unconstitutional stops. See U.S. Const, art. III.
. See Monell v. New York City Dep’t of Soc. Sens., 436 U.S. 658, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978) (establishing the standards under 42 U.S.C. § 1983 for municipal liability for constitutional torts by employees).
. Teny, 392 U.S. at 14 n. 11, 88 S.Ct. 1868.
. See Ligón v. City of New York, 925 F.Supp.2d 478 (S.D.N.Y.2013).
. Cash v. County of Erie, 654 F.3d 324, 333 (2d Cir.2011), cert. denied, -U.S.-, 132 S.Ct. 1741, 182 L.Ed.2d 528 (2012) (quoting Connick v. Thompson, — U.S.-, 131 S.Ct. 1350, 1359, 179 L.Ed.2d 417 (2011)).
. Connick, 131 S.Ct. at 1359.
. See infra Part III.B.
. United States v. Swindle, 407 F.3d 562, 566 (2d Cir.2005) (quoting United States v. Sokolow, 490 U.S. 1, 7, 109 S.Ct. 1581, 104 L.Ed.2d 1 (1989)) (some quotation marks omitted).
. United States v. Bayless, 201 F.3d 116, 133 (2d Cir.2000).
. See Florida v. Bostick, 501 U.S. 429, 111 S.Ct. 2382, 115 L.Ed.2d 389 (1991).
. United States v. Lopez, 321 Fed.Appx. 65, 67 (2d Cir.2009) (quoting Arizona v. Johnson, 555 U.S. 323, 326-27, 129 S.Ct. 781, 172 L.Ed.2d 694 (2009)).
. Hayden v. Paterson, 594 F.3d 150, 163 (2d Cir.2010).
. See infi'a Part IV.A.
. See infra Part IV.B.
. See infra Part IV.B.2.
. See infra Part IV.B.3.
. See infra Part IV.C.
. See infra Part IV.C.2.
. 2010 Memorandum of Chief of Patrol James Hall, Plaintiffs’ Trial Exhibit ("PX”) 290 at *0096.
. See infra Part IV.C.3.
. 4/18 Trial Transcript ("Tr.”) at 4047-4049 (emphasis added).
. 5/9 Tr. at 6431-6433.
. See infra Part IV.D.
. "Throughout the country, police work diligently every day trying to prevent crime, arrest those who are responsible, and protect victims from crimes that undermine their dignity and threaten their safety. They work for relatively low pay for the risks that they take, and although in some communities their role is respected and admired, in other communities they are vilified and treated as outcasts.” Charles Ogletree, The Presumption of Guilt 125 (2012).
. I note again that based on the uncontested statistics, see infra Part IV.A, the NYPD’s current use of stop and frisk has not been particularly successful in producing arrests or seizures of weapons or other contraband.
. For example, as discussed at length in this Opinion, race is a permissible consideration where there is a specific suspect description that includes race. See, e.g., Brown v. City of Oneonta, New York, 221 F.3d 329, 340 (2d Cir.2000).
. Cf. City of Indianapolis v. Edmond, 531 U.S. 32, 41-42, 121 S.Ct. 447, 148 L.Ed.2d 333 (2000) (holding that while suspicionless stops at a highway checkpoint may be constitutional under the Fourth Amendment when "designed primarily to serve purposes closely related to the problems of policing the border or the necessity of ensuring roadway safety," highway stops that lack "some measure of individualized suspicion" and “whose primary purpose [is] to detect evidence of ordinary criminal wrongdoing” contravene the Fourth Amendment).
. Heck v. Humphrey, 512 U.S. 477, 483, 114 S.Ct. 2364, 129 L.Ed.2d 383 (1994) (quoting Memphis Cmty. Sch. Dist. v. Stachura, 477 U.S. 299, 305, 106 S.Ct. 2537, 91 L.Ed.2d 249 (1986)).
. Interpreting the language of section 1983 and the legislative history surrounding its passage in the Civil Rights Act of 1871, the Court in Monell held that local governing bodies could be held liable either on the basis of formally approved policies or on the basis of “ 'customs’ ” or “ 'usages.' " Monell, 436 U.S. 658, 690-91, 98 S.Ct. 2018, 56 L.Ed.2d 611 (1978) (quoting Adickes v. S.H. Kress & Co., 398 U.S. 144, 167-68, 90 S.Ct. 1598, 26 L.Ed.2d 142 (1970)).
. Cash, 654 F.3d at 333 (quoting Connick, 131 S.Ct. at 1359, in turn quoting Monell, 436 U.S. at 691, 98 S.Ct. 2018 (quotation marks omitted)). Cases after Monell “considerably broadened the concept of official municipal action.” Amnesty Am. v. Town of W. Hartford, 361 F.3d 113, 125 (2d Cir.2004) (Soto-mayor, J.).
. Cash, 654 F.3d at 333 (quoting Connick, 131 S.Ct. at 1359) (some quotation marks omitted).
. Connick, 131 S.Ct. at 1359 (citing Monell, 436 U.S. at 694, 98 S.Ct. 2018; Pembaur v. Cincinnati, 475 U.S. 469, 479, 106 S.Ct. 1292, 89 L.Ed.2d 452 (1986); Adickes, 398 U.S. at 167-68, 90 S.Ct. 1598).
. Cash, 654 F.3d at 334.
. Id. (quoting Amnesty, 361 F.3d at 126).
. Connick, 131 S.Ct. at 1359 (quoting Board of Comm’rs of Bryan Cnty. v. Brown, 520 U.S. 397, 410, 117 S.Ct. 1382, 137 L.Ed.2d 626 (1997)).
. Cash, 654 F.3d at 334 (quoting Connick, 131 S.Ct at 1360; Amnesty, 361 F.3d at 128).
. Id. (quoting Vann v. City of New York, 72 F.3d 1040, 1049 (2d Cir.1995); Reynolds v. Giuliani, 506 F.3d 183, 192 (2d Cir.2007)).
. Connick, 131 S.Ct. at 1359 (citing Oklahoma City v. Tuttle, 471 U.S. 808, 822-23, 105 S.Ct. 2427, 85 L.Ed.2d 791 (1985) (plurality opinion)).
. Id. (citing Biyan Cnty., 520 U.S. at 407, 117 S.Ct. 1382).
. Cash, 654 F.3d at 334 (citing Walker v. City of New York, 974 F.2d 293, 297-98 (2d Cir. 1992)).
. Walker, 974 F.2d at 298 (quoting City of Canton, Ohio v. Harris, 489 U.S. 378, 390, 109 S.Ct. 1197, 103 L.Ed.2d 412 (1989)). In order to establish Monell liability based on the Walker test, plaintiffs must also, of course, show that the training or supervision was in fact inadequate and that this inadequacy caused plaintiffs’ constitutional injuries. See Reynolds, 506 F.3d at 193.
. Connick, 131 S.Ct. at 1360 (quoting Bryan Cnty., 520 U.S. at 409, 117 S.Ct. 1382). By contrast, "once a municipal policy is established, 'it requires only one application ... to satisfy fully Monell's requirement that a municipal corporation be held liable only for constitutional violations resulting from the municipality's official policy.’ ” Pembaur, 475 U.S. at 478 n. 6, 106 S.Ct. 1292 (quoting Tuttle, 471 U.S. at 822, 105 S.Ct. 2427) (emphasis added).
. See Maryland v. Pringle, 540 U.S. 366, 369, 124 S.Ct. 795, 157 L.Ed.2d 769 (2003) (citing Mapp v. Ohio, 367 U.S. 643, 81 S.Ct. 1684, 6 L.Ed.2d 1081 (1961)).
. U.S. Const. amend. IV.
. Brigham City, Utah v. Stuart, 547 U.S. 398, 403, 126 S.Ct. 1943, 164 L.Ed.2d 650 (2006). Accord Pennsylvania v. Mimms, 434 U.S. 106, 108-09, 98 S.Ct. 330, 54 L.Ed.2d 331 (1977) (per curiam) ("The touchstone of our analysis under the Fourth Amendment is always 'the reasonableness in all the circumstances of the particular governmental invasion of a citizen’s personal security.’ ” (quoting Terry, 392 U.S. at 19, 88 S.Ct. 1868)).
. Swindle, 407 F.3d at 566 (quoting Sokolow, 490 U.S. at 7, 109 S.Ct. 1581) (some quotation marks omitted).
. See Davis v. City of New York, 902 F.Supp.2d 405, 411 (S.D.N.Y.2012) (citing Terry, 392 U.S. at 1, 88 S.Ct. 1868).
. 501 U.S. at 436, 111 S.Ct. 2382. The "free to terminate the encounter” standard is a *566more general formulation of Justice Potter Stewart’s “free to leave” standard in United States v. Mendenhall, 446 U.S. 544, 554, 100 S.Ct. 1870, 64 L.Ed.2d 497 (1980) ("[A] person has been ‘seized’ within the meaning of the Fourth Amendment' only if, in view of all of the circumstances surrounding the incident, a reasonable person would have believed that he was not free to leave.”). See also Terry, 392 U.S. at 16, 88 S.Ct. 1868 ("It must be recognized that whenever a police officer accosts an individual and restrains his freedom to walk away, he has ‘seized’ that person.”).
. Bostick, 501 U.S. at 437, 111 S.Ct. 2382, 115 L.Ed.2d 389 (quoting Michigan v. Chestemut, 486 U.S. 567, 569, 108 S.Ct. 1975, 100 L.Ed.2d 565 (1988)). Bostick also notes that "the ‘reasonable person’ test presupposes an innocent person.” Id. at 438, 111 S.Ct. 2382. For a comprehensive summary of the "free to leave” test as it has been interpreted and applied, see 4 Wayne R. LaFave, Search & Seizure § 9.4(a) (5th ed. 2012) ("LaFave”).
. See Bostick, 501 U.S. at 437, 111 S.Ct. 2382, 115 L.Ed.2d 389.
. See id. at 434-35, 111 S.Ct. 2382.
. Id. (collecting cases) (emphasis added and citations omitted). Accord INS v. Delgado, 466 U.S. 210, 216, 104 S.Ct. 1758, 80 L.Ed.2d 247 (1984) ("[P]olice questioning, by itself, is unlikely to result in a Fourth Amendment violation ... \u\nless the circumstances of the encounter are so intimidating as to demonstrate that a reasonable person would have believed he was not free to leave if he had not responded.” (emphasis added)). These cases confirm that the manner and context of police conduct are relevant to the inquiry into whether a reasonable person would have felt free to terminate the encounter. As the Second Circuit has noted, this inquiry is essentially "an objective assessment of the overall coercive effect of the police conduct.” United States v. Lee, 916 F.2d 814, 819 (2d Cir.1990) (citing Chestemut, 486 U.S. at 573-74, 108 S.Ct. 1975).
. Bostick, 501 U.S. at 438, 111 S.Ct. 2382, 115 L.Ed.2d 389 (quoting Bostick, 501 U.S. at 447, 111 S.Ct. 2382, 115 L.Ed.2d 389 (Marshall, J., dissenting)) (emphasis omitted).
. United States v. Glover, 957 F.2d 1004, 1008 (2d Cir.1992) (citing Lee, 916 F.2d at 819). Accord United States v. Drayton, 536 U.S. 194, 203-04, 122 S.Ct. 2105, 153 L.Ed.2d 242 (2002) (concluding, under Bostick framework, that a reasonable passenger on a bus would feel free to leave, where "officers gave the passengers no reason to believe that they were required to answer the officers’ questions,” and the officer asking questions of passengers “did not brandish a *567weapon or make any intimidating movements,” “left the aisle free so that respondents could exit,” and "spoke to passengers one by one and in a polite, quiet voice”).
. Ligon, 925 F.Supp.2d at 534-36 (citing United States v. Simmons, 560 F.3d 98, 101, 105-06 (2d Cir.2009); Brown, 221 F.3d at 340) .
. Id. at 536 n. 410 (citing Brown, 221 F.3d at 341) .
. Simmons, 560 F.3d at 105 (citing Swindle, 407 F.3d at 572).
. Terry, 392 U.S. at 30, 88 S.Ct. 1868.
. United States v. Place, 462 U.S. 696, 702, 103 S.Ct. 2637, 77 L.Ed.2d 110 (1983) (citing Terry, 392 U.S. at 22, 88 S.Ct. 1868). Although the Court in Terry did not explicitly refer to crimes that are "about to be” committed, the stop upheld in Terry was based on a police officer's suspicion that two men were about to carry out a "stick-up.” Terry, 392 U.S. at 6, 88 S.Ct. 1868. It has long been recognized that arrests may be based on probable cause to believe that a crime is about to be committed. The New York stop and frisk statute, New York Criminal Procedure Law (“CPL”) § 140.50(1), allows stops when an officer "reasonably suspects that such person is committing, has committed or is about to commit either (a) a felony or (b) a misdemeanor defined in the penal law.”
. Alabama v. White, 496 U.S. 325, 329, 110 S.Ct. 2412, 110 L.Ed.2d 301 (1990) (quoting Sokolow, 490 U.S. at 7, 109 S.Ct. 1581) (some quotation marks omitted). Courts are divided over whether reasonable suspicion must be of a particular crime, or may be of criminality in general. See 4 LaFave § 9.5(c).
. United States v. Elmore, 482 F.3d 172, 178-79 (2d Cir.2007) (quoting Terry, 392 U.S. at 21, 88 S.Ct. 1868).
. Bayless, 201 F.3d at 133.
. See Edmond, 531 U.S. at 37, 121 S.Ct. 447 (citing Chandler v. Miller, 520 U.S. 305, 308, 117 S.Ct. 1295, 137 L.Ed.2d 513 (1997)).
. See id. at 37-40, 121 S.Ct. 447 (summarizing exceptions); id. at 34, 41-44, 121 S.Ct. 447 (distinguishing between suspicion-less stops at highway checkpoints "for the purposes of combating drunk driving and intercepting illegal immigrants,” which are constitutional; and suspicionless stops at checkpoints that primarily aim to advance " 'the general interest in crime control,’ ” which are unconstitutional (quoting Delaware v. Prouse, 440 U.S. 648, 659 n. 18, 99 S.Ct. 1391, 59 L.Ed.2d 660 (1979))).
. United States v. Arvizu, 534 U.S. 266, 273, 122 S.Ct. 744, 151 L.Ed.2d 740 (2002) (quoting United States v. Cortez, 449 U.S. 411, 417-18, 101 S.Ct. 690, 66 L.Ed.2d 621 (1981)).
. Lee, 916 F.2d at 819.
. Illinois v. Wardlow, 528 U.S. 119, 123, 120 S.Ct. 673, 145 L.Ed.2d 570 (2000).
. Adams v. Williams, 407 U.S. 143, 146, 92 S.Ct. 1921, 32 L.Ed.2d 612 (1972) (citing Terry, 392 U.S. at 24, 88 S.Ct. 1868).
. Ybarra v. Illinois, 444 U.S. 85, 93, 100 S.Ct. 338, 62 L.Ed.2d 238 (1979) (citing Adams, 407 U.S. at 146, 92 S.Ct. 1921).
. Lopez, 321 Fed.Appx. at 67 (quoting Johnson, 555 U.S. at 326-27, 129 S.Ct. 781) (holding that the following behavior provided reasonable suspicion that a suspect was carrying a weapon: "[w]hen [the suspect] saw [the officer] approaching him, [the suspect] transferred [a] cup from his right to his left hand and dropped his right hand to his right side”). Accord 4 LaFave § 9.6(a) (noting that "assuming a proper stopping for investigation, a protective search is permissible when, at the time the frisk itself is commenced, there is reason to believe that the suspect may be armed and dangerous”).
. 4 LaFave § 9.6(a) (citing United States v. Thaipe, 536 F.2d 1098 (5th Cir.1976)). Trial courts “have been inclined to view the right to frisk as being 'automatic' whenever the suspect has been stopped upon the suspicion [of a criminal activity] for which the offender would likely be armed,” such as "robbery, burglary, rape, assault with weapons, car theft, homicide, and dealing in large quantities of narcotics.” Id.
I note that the New York stop and frisk statute authorizes an officer to conduct a frisk *569whenever, after a stop, he "reasonably suspects that he is in danger of physical injury.” CPL § 140.50(3). This standard is not the constitutional standard. It would allow an officer to conduct a frisk even when she lacks reasonable suspicion that the stopped person is armed and dangerous. As the Supreme Court has made clear, New York "may not ... authorize police conduct which trenches upon Fourth Amendment rights.” Sibron v. New York, 392 U.S. 40, 61, 88 S.Ct. 1889, 20 L.Ed.2d 917 (1968). The Fourth Amendment, and not New York law, establishes the requirements for a constitutional frisk in this case.
. Adams, 407 U.S. at 146, 92 S.Ct. 1921.
. Id.
. Minnesota v. Dickerson, 508 U.S. 366, 373, 113 S.Ct. 2130, 124 L.Ed.2d 334 (1993) (quoting Terry, 392 U.S. at 26, 88 S.Ct. 1868).
. Id. at 375, 113 S.Ct. 2130 (emphasis added).
. Ybarra, 444 U.S. at 93-94, 100 S.Ct. 338.
. See People v. Collins, 1 Cal.3d 658, 83 Cal.Rptr. 179, 463 P.2d 403, 406 (1970) (holding “that an officer who exceeds a pat-down without first discovering an object which feels reasonably like a knife, gun, or club must be able to point to specific and articulable facts which reasonably support a suspicion that the particular suspect is armed with an atypical weapon which would feel like the object felt during the pat-down”).
. People v. De Bour, 40 N.Y.2d 210, 386 N.Y.S.2d 375, 352 N.E.2d 562 (1976). See, e.g., Ligón, 925 F.Supp.2d at 532-40. I note that the NYPD’s policies and training materials also draw from New York’s stop and frisk statute. Compare, e.g., Patrol Guide 212-11: Stop and Frisk, PX 98, at 1, with CPL § 140.50. As noted earlier, the New York statutory standard for a frisk is not the Fourth Amendment standard as defined by the Supreme Court.
. See Davis v. City of New York, No. 10 Civ. 0699, 959 F.Supp.2d 324, 342 n. 75, 2013 WL 1288176, at *7 n. 75 (S.D.N.Y. Mar. 28, 2013).
. Sibron, 392 U.S. at 61, 88 S.Ct. 1889 (reversing conviction for failure to suppress evidence seized in an unlawful stop).
. U.S. Const. amend. XIV § 1.
. City of Cleburne v. Cleburne Living Ctr., Inc., 473 U.S. 432, 439, 105 S.Ct. 3249, 87 L.Ed.2d 313 (1985).
. See Washington v. Davis, 426 U.S. 229, 239-40, 96 S.Ct. 2040, 48 L.Ed.2d 597 (1976). I note that the parties sometimes use the phrase "pattern and practice” in referring to plaintiffs' constitutional claims. See, e.g., 6/12/13 Post-Trial Memorandum of Law in Support of Plaintiffs’ Claims ("PL Mem.”) at i-ii; 6/12/13 Defendant's Post-Trial Memorandum of Law ("Def. Mem.”) at 6 n. 11. However, pattern or practice analysis does not govern equal protection claims. See Chavez v. Illinois State Police, 251 F.3d 612, 638 n. 8 (7th Cir.2001). Rather, the term "pattern or practice” appears in civil rights statutes such as Title VII of the Civil Rights Act of 1964, as well as several "statutes authorizing the Attorney General to bring suits to remedy discrimination.” Marshall Miller, Police Brutality, 17 Yale L. & Pol'y Rev. 149-151, 169 & n. 124 (1998) (discussing Title XXI of the Violent Crime Control and Law Enforcement Act of 1994, codified at 42 U.S.C. § 14141, which allows the Attorney General to sue law enforcement agencies that "engage in a pattern or practice” of unconstitutional conduct).
I also note that despite the occasional use of the terms "disparate treatment” and "disparate impact” by the parties’ experts, see infra Part IV.B.3, these terms of art are generally applied to Title VII and other statutory claims, not equal protection claims. See, e.g., Ricci v. DeStefano, 557 U.S. 557, 577, 129 S.Ct. 2658, 174 L.Ed.2d 490 (2009) (“Title VII prohibits both intentional discrimination (known as 'disparate treatment') as well as, in some cases, practices that are not intended to discriminate but in fact have a disproportionately adverse effect on minorities (known as 'disparate impact’).”). The New York City Council recently passed a bill that would create a private right of action for claims of "bias-based profiling” based on discriminatory intent or disparate impact. See N.Y. City Council Introductory No. 1080 of 2013 § 2.
. Brown, 221 F.3d at 337. Accord Pyke v. Cuomo ("Pyke II”), 567 F.3d 74, 76 (2d Cir. 2009) (citing Pyke v. Cuomo {"Pyke I”), 258 F.3d 107, 110 (2d Cir.2001)).
. Brown, 221 F.3d at 337 (quoting Hayden v. County of Nassau, 180 F.3d 42, 48 (2d Cir. 1999)). An express racial classification is "subject to strict judicial scrutiny.” Pyke II, 567 F.3d at 77 (citing Loving v. Virginia, 388 U.S. 1, 11, 87 S.Ct. 1817, 18 L.Ed.2d 1010 (1967); lohnson v. California, 543 U.S. 499, 505, 125 S.Ct. 1141, 160 L.Ed.2d 949 (2005) (holding that “all racial classification” im*571posed by government "must be analyzed by a reviewing court under strict scrutiny”)). Accord Fisher v. University of Texas at Austin, - U.S. -, 133 S.Ct. 2411, 2419, 186 L.Ed.2d 474 (2013) {“[A\ny official action that treats a person differently on account of his race or ethnic origin is inherently suspect.” (quotation marks and citation omitted, and emphasis added)). "In order to satisfy strict scrutiny, a classification must further a compelling state interest and be narrowly tailored to accomplish the purpose.” Pyke II, 567 F.3d at 77 (citing Shaw v. Hunt, 517 U.S. 899, 908, 116 S.Ct. 1894, 135 L.Ed.2d 207 (1996)).
. Brown, 221 F.3d at 337 (citing Yick Wo v. Hopkins, 118 U.S. 356, 373-74, 6 S.Ct. 1064, 30 L.Ed. 220 (1886)).
. Id. (citing Village of Arlington Heights v. Metropolitan Hous. Dev. Corp., 429 U.S. 252, 264-65, 97 S.Ct. 555, 50 L.Ed.2d 450 (1977); Johnson v. Wing, 178 F.3d 611, 615 (2d Cir. 1999)).
. Pyke I, 258 F.3d at 110. An exception exists for plaintiffs alleging a selective prosecution in violation of the Equal Protection Clause. In order to prevail on this claim, plaintiffs "must plead and establish the existence of similarly situated individuals who were not prosecuted; that is because courts grant special deference to the executive branch in the performance of the 'core' executive function of deciding whether to prosecute.” Pyke I, 258 F.3d at 109 (citing United States v. Armstrong, 517 U.S. 456, 465, 116 S.Ct. 1480, 134 L.Ed.2d 687 (1996)).
. Paterson, 594 F.3d at 163 (quoting Arlington Heights, 429 U.S. at 265, 97 S.Ct. 555).
. Arlington Heights, 429 U.S. at 265, 97 S.Ct. 555.
. Id. at 265-66, 97 S.Ct. 555 (emphasis added). Accord Personnel Adm’r of Mass. v. Feeney, 442 U.S. 256, 277, 99 S.Ct. 2282, 60 L.Ed.2d 870 (1979) ("Discriminatory intent is simply not amenable to calibration. It either is a factor that has influenced the [governmental action] or it is not.”).
. Paterson, 594 F.3d at 163 (quoting Feeney, 442 U.S. at 279, 99 S.Ct. 2282 (citation and footnote omitted)) (some quotation marks omitted).
. Id. (quoting Arlington Heights, 429 U.S. at 266, 97 S.Ct. 555).
. Feeney, 442 U.S. at 279 n. 24, 99 S.Ct. 2282. "An invidious discriminatory purpose may often be inferred from the totality of the relevant facts, including the fact, if it is true, that the [practice] bears more heavily on one race than another.” Washington, 426 U.S. at 242, 96 S.Ct. 2040.
. United States v. City of Yonkers, 96 F.3d 600, 612 (2d Cir.1996) (quoting United States v. Yonkers Bd. of Educ., 837 F.2d 1181, 1217 (2d Cir.1987)).
. Id. (quoting Yonkers, 837 F.2d at 1217).
. Plaintiffs filed the case on January 31, 2008. On August 31, 2011, I granted in part and denied in part defendants' motion for partial summary judgment. See Floyd v. City of New York, 813 F.Supp.2d 417 (S.D.N.Y. 2011) , partial reconsideration granted, 813 F.Supp.2d 457 (S.D.N.Y.2011). On April 14, 2012, I granted in part and denied in part defendants’ motion to exclude the testimony of plaintiffs’ liability expert. See Floyd v. City of New York, 861 F.Supp.2d 274 (S.D.N.Y. 2012) . On May 16, 2012, I granted plaintiffs’ motion to be certified as a class. See Floyd v. City of New York, 283 F.R.D. 153 (S.D.N.Y. 2012). In an order entered March 8, 2013, I approved the parties’ stipulation withdrawing plaintiffs’ damages claims, dismissing plaintiffs’ claims against individual officers, and altering the caption to reflect the remaining parties. See Stipulation and Order of Withdrawal of Individual Damage Claims (3/8/13). Plaintiffs have not pursued the state law claims in their Second Amended Complaint. See Second Amended Complaint (10/20/08); PL Mem. at i-ii; Def. Mem. at 2 n. 2. Plaintiffs’ only claims are section 1983 claims against the City for Fourth and Fourteenth Amendment violations.
. "To establish by a preponderance of the evidence means veiy simply to prove that something is more likely than not so." Duke Labs., Inc. v. United States, 222 F.Supp. 400, 406 (D.Conn.1963), aff'd, 337 F.2d 280 (2d Cir.1964).
. "In an action tried on the facts without a jury ..., the court must find the facts specially and state its conclusions of law separately." Fed.R.Civ.P. 52(a)(1).
. See infra App. A ("Blank UF-250”).
. See Report of Jeffrey Fagan, Ph.D. (Oct. 15, 2010), PX 411 (“Fagan Rpt.”); Supplemental Report of Jeffrey Fagan, Ph.D. (Dec. 3, 2010), PX 412 ("Fagan Supp. Rpt.”); Second Supplemental Report of Jeffrey Fagan, Ph.D. (Nov. 29, 2012), PX 417 ("Fagan 2d Supp. Rpt.").
. See infra Part IV.B. The City's liability experts concede that Dr. Fagan's "descriptive statistics are not a source of contention.” Report of Dennis C. Smith, Ph.D. and Robert M. Purtell, Ph.D. in Response to the Second Supplemental Report of Jeffrey Fagan, Ph.D. (Feb. 1, 2013), Defendant's Trial Exhibit ("DX”) H13, at 7 ("Smith Rpt.”). I rely on the Smith Report for most of my citations to Dr. Smith’s and Dr. Purtell's written opinions, although they had prepared prior written opinions. See, e.g., Report of Dennis C. Smith, Ph.D. (Nov. 15, 2010), DX T8; Declaration of Dennis C. Smith (Dec. 19, 2011), DX 08.
. Of course, if an officer decides that he has not made a Terry stop, or simply fails to complete a UF-250 after a stop, then the encounter will not be reflected in the UF-250 database. See infra Part IV.C.4-5 (evidence of inadequate training regarding what constitutes a Terry stop, and of failures to fill out UF-250s).
. These dates are the class period in this case. Dr. Fagan’s October 2010 and December 2010 reports analyzed data from January 1, 2004 through December 31, 2009, while his November 2012 report analyzed data from January 1, 2010 through June 30, 2012. See Fagan 2d Supp. Rpt. at 1.
. The City counts a total of 4,431,414 UF-250s, while plaintiffs count a total of 4,430,-140. See DX V14-A; PX 417D; Fagan 2d Supp. Rpt. at 10 tbl. 1. While I use the 4.4 million figure throughout this Opinion, the actual number of stops during the class period is likely higher, because officers do not always prepare a UF-250 after a stop. See, e.g., infra Part IV.D.l.a (no UF-250 for Leroy Downs stop); infra Part IV.D.l.h (no UF-250 for Clive Lino stop); Ligón, 925 F.Supp.2d at 498-99 (noting that the officer who stopped named plaintiff Charles Bradley failed to fill out a UF-250); id. at 512 & n. 241 (noting that the CCRB has reported on a systematic problem with officers failing to complete UF-250s after stops). It is impossible to determine how often officers fail to fill out UF-250s for Terry stops, because the NYPD has no reliable mechanism for monitoring this failure. See infra Part IV.C.4.
. See DX V14-A.
. See id. (2,284,246 total frisks); DXV14-C (33,882 weapons found). This assumes that weapons were only found in stops that involved frisks. If weapons were found in stops that did not involve frisks, the percentage of frisks leading to the discovery of weapons would be even lower.
. See DX V14-B; DX V14-C; DX V14-D; Fagan Rpt. at 63 (defining contraband).
. These statistics are sometimes referred to as "hit rates.” See Fagan Rpt. at 64 tbls. 14 & 15 (5.37% of stops resulted in arrest and 6.26% of stops resulted in a summons between 2004 and 2009); Fagan 2d Supp. Rpt. at 34 tbl. 14 (6.26% of stops resulted in arrest *574and 6.25% of stops resulted in a summons between 2010 and June 2012).
. See Fagan Rpt. at 22 tbl. 3 (1,445,472 stops of blacks between 2004 and 2009); Fagan 2d Supp. Rpt. at 11 tbl. 3 (843,684 stops of blacks between January 2010 and June 2012). In Dr. Fagan’s studies, the category “black” encompasses two checkboxes in the "Race” section on the UF-250, "Black” and "Black Hispanic.” See Fagan Rpt. at 20; Blank UF-250. The term black is used throughout this opinion rather than African-American as that term is used on the UF-250 form.
. See Fagan Rpt. at 22 tbl. 3 (841,755 stops of Hispanics between 2004 and 2009); Fagan 2d Supp. Rpt. at 11 tbl. 3 (520,171 stops of Hispanics between January 2010 and June 2012). As in Dr. Fagan's studies, when the term "Hispanic” appears in (his Opinion, it refers only to the "White Hispanic” checkbox on the UF-250. See Fagan Rpt. at 20; Blank UF-250. "Hispanic” is an imperfect and contested term, like most racial and ethnic classifications. The term "Hispanic” is used rather than "Latino” because "Hispanic” appears on the UF-250. See Blank UF-250. It is unfortunately necessary to reduce New York's rich demographic diversity to a few simple racial categories for the purposes of this case.
. See Fagan Rpt. at 22 tbl. 3 (286,753 stops of whites between 2004 and 2009); Fagan 2d Supp. Rpt. at 11 tbl. 3 (148,283 stops of whites between January 2010 and June 2012). 90% of all stops were of men, and 69% were of people between the ages of 16 and 34. See Fagan Rpt. at 22 tbl. 3; Fagan 2d Supp. Rpt. at 11 tbl. 3.
. See Raymond W. Kelly, Police Commissioner, NYPD, Crime and Enforcement Activity in New York City (Jan[J 1-Dec[J 31, 2012) app. B (citing 2010 census).
. See Fagan Rpt. at 64 tbl. 14; Fagan 2d Supp. Rpt. at 34 tbl. 14.
. See Fagan Rpt. at 64 tbl. 15; Fagan 2d Supp. Rpt. at 35 tbl. 15. Because guns were seized in only 0.1% of stops, it is difficult to draw meaningful inferences from the statistics regarding gun seizures. See Fagan Rpt. at 63-64 & tbls. 14-15; Fagan 2d Supp. Rpt. at 35 & tbl. 15.
. See Fagan Rpt. at 64 tbl. 15; Fagan 2d Supp. Rpt. at 35 tbl. 15.
. Dr. Fagan did not update the following statistics for the years 2010 through 2012.
. See Fagan Rpt. at 51-52 & tbl. 11; Fagan Supp. Rpt. at 41. In a stop based on a radio run, as opposed to a self-initiated stop, the suspicion leading to the stop involves information received by the officer over the radio. *575See Fagan Rpt. at 50. 78% of stops during the class period were self-initiated. See PX 417D.
I also note that the number of stop factors indicated on UF-250s increased from 2004 to 2009. In 2004, the average was 1.01 factors on the front of the form ("What Were Circumstances Which Led To Stop?”) and 1.53 factors on the back (“Additional Circumstances/Factors”). By 2009, the average was 1.47 factors on the front of the form and 1.93 factors on the back. "Furtive Movements” appears on the front, and "High Crime Area” on the back. See Fagan Supp. Rpt. at 40 tbl. S6; Blank UF-250.
. See Fagan Rpt. at 52.
. See Fagan Supp. Rpt. at 39; DXV14-A. On the forms without a specific suspected crime, the "Specify Which Felony/P.L. Misdemeanor Suspected” field was either left empty, contained a text string that does not describe a penal law category or a violation, or contained the following generic text strings: "FEL,” "FELONY,” "MISD,” or "MISDEMEANOR.” See Fagan Supp. Rpt. at 39 n. 89.
. 6/12/13 Defendant’s Proposed Findings of Fact and Conclusions of Law ("Def. Findings”) ¶ 88.
. Blank UF-250. See also Patrol Guide 212-11: Stop and Frisk (7/18/13) ("Patrol Guide: Stop and Frisk”), PX 98, at 1.
. In fact, by far the most commonly charged summons offenses in the City are public consumption of alcohol, a violation under New York City Administrative Code § 10-125(b), and disorderly conduct, a violation under N.Y. Penal Law ("PL”) § 240.20. See, e.g., Hon. Fern A. Fisher, Deputy Chief Administrative Judge, Criminal Court of the City of New York: Annual Report 2010, at 35 (recording 140,425 summonses for public consumption and 81,036 summonses for disorderly conduct in 2010).
. See, e.g., infra Part IV.D.l.g (David Ourlicht's post-stop summons for disorderly con*576duct based on conduct beginning after the stop).
. See, e.g., infra Part IV.D.2.e (stop in which officers approached Deon Dennis for drinking in public, then arrested him based on an active warrant).
. See, e.g., infra Part IV.D.l.g (Ourlicht's post-stop summons was dismissed); Fisher, Criminal Court of the City of New York: Annual Report 2010, at 16 (stating that 42% of all summonses in 2010 resulted in either dismissal or adjournment in contemplation of dismissal).
. See, e.g., Ligón, 925 F.Supp.2d at 491-96 (discussing the Bronx ADA's decision to decline prosecution of some trespass arrests based on stops outside Clean Halls buildings).
. See New York Civil Liberties Union, NYPD Stop-and-Frisk Activity in 2012, at 17 (2013) (noting that 16% of total arrests following stops are for marijuana possession,' making marijuana the most common arrest offense arising out of stops). However, marijuana possession can only lead to arrest when the marijuana is "in a public place ... burning or open to public view." PL § 221.10(1).
. NYPD, Operations Order 49 (9/19/11), ¶ 2.
. The NYPD’s marijuana possession arrest practices recently became the subject of a separate lawsuit filed by the Bronx Defenders. See Felix v. City of New York, No. 13 Civ. 2941(JMF).
. See Floyd v. City of New York, 861 F.Supp.2d 274, 279 (S.D.N.Y.2012) (summarizing Dr. Fagan's qualifications, including his role in the production of the 1999 report from *577the New York State Office of the Attorney General on the NYPD’s stop and frisk practices).
. See Curriculum Vitae of Jeffrey Fagan (Oct. 2012), PX417A.
. See Smith Rpt. at 3, 5.
. See Floyd v. City of New York, 910 F.Supp.2d 506, 509-10 (S.D.N.Y.2012) (summarizing Dr. Smith’s and Dr. Purtell’s qualifications).
. See Smith Rpt. at 5-6; 5/2 Tr. at 5724, 5728-5729, 5839.
. Indeed, Dr. Purtell testified that he had never read a study of racial disparities in police stops other than Dr. Fagan’s study in this case. See 5/2 Tr. at 5861.
. To take two examples: First, in his October 2010 report, Dr. Fagan made a coding error, later corrected, that resulted in the misclassification of a large number of “apparently unjustified” stops as "indeterminate.” See 4/4 Tr. at 2295-2296; 4/5 Tr. at 2445-2448. Second, table 10 in the Smith Report contains a partly mislabeled column of data that was later corrected. See 5/13 Tr. at 6839; Smith Rpt. at 70 tbl. 10; PX 574 (corrected and expanded table 10).
. See Fagan Rpt. at 33 tbl. 5, 42 tbl. 7; Fagan 2d Supp. Rpt. at 18 tbl. 5, 20 tbl. 7.
. 5/2 Tr. at 5764. Based on this assumption, Dr. Purtell claimed that Dr. Fagan’s Table 5 showed that “the chance of a black person over a white person being stopped is ... [0.22%] above random chance.” Id.
. In response to Dr. Purtell's testimony, plaintiffs' counsel objected that "nowhere in table 5 is there anything about the likelihood of a black or a white person being stopped.” Id. at 5765. Dr. Purtell defended his analysis of the numbers in Table 5, stating: "This is a standard interpretation of these numbers.” Id. at 5767. The following day, the City and Dr. Purtell conceded plaintiffs’ objection. Dr. Purtell stated that his exhibit had been mistitled, and explained that he "wrote that the night before,” while he had another of Dr. Fagan’s tables on his mind. 5/3 Tr. at 5903.
. Even if there were no time constraints, after-the-fact testimony from interested parties is an imperfect source of information, as the individual stop testimony in this case shows.
. See, e.g., 4/10 Tr. at 3207 (McHugh); PL Findings ¶ 118. Accord. Def. Mem. at 7.
. See Fagan Rpt. at 48-50; 2/2/12 Declaration of Jeffrey Fagan, PX 415, ¶¶ 14-18; Floyd, 861 F.Supp.2d at 283-84; Blank UF-250.
. See Fagan Rpt. at 48-50; Blank UF-250.
. See Fagan 2d Supp. Rpt. at 24-25. The revised version of the October 2010 report uses different language to describe these three categories, see Fagan Rpt. at 50, but I use the updated language in the Fagan 2d Supp. Rpt. for the sake of simplicity.
. See Fagan Rpt. at 50; Fagan 2d Supp. Rpt. at 25.
. See id. at 50-52, 56 tbl. 12.
. The parties have generally referred to the latter factor as "High Crime Area” and I do the same, although everyone agrees that the abbreviation is not a perfect reflection of the full label, "Area Has High Incidence of Reported Offense Of Type Under Investigation.” The same applies to the shorthand for other stop factors, including "Time of Day” for "Time Of Day, Day Of Week, Season Corresponding To Reports Of Criminal Activity,” on Side 2 of the UF-250. See Blank UF-250.
. See supra Part IV.A.
. The Supreme Court has "recognized that nervous, evasive behavior is a pertinent factor in determining reasonable suspicion.” Ward-low, 528 U.S. at 124, 120 S.Ct. 673 (emphasis added) (citing numerous cases). But "furtive behavior absent additional indicia of suspicion generally does not suffice to establish reasonable suspicion.” United States v. Bellamy, 592 F.Supp.2d 308, 318-19 (E.D.N.Y. 2009) (collecting cases). Accord Ligón, 925 F.Supp.2d at 530-31.
. United States v. Broomfield, 417 F.3d 654, 655 (7th Cir.2005).
. See Ligón, 925 F.Supp.2d at 530-31.
. Id. (footnote omitted).
. See, e.g., Jerry Kang & Mahzarin R. Banaji, Fair Measures: A Behavioral Realist Revision of “Affirmative Action”, 94 Cal. L.Rev. 1063 (2006) (illustrating relevance of implicit social cognition studies to issues of discrimination). Kang and Banaji quote the following apt observation from a dissent by Justice Ruth Bader Ginsburg: “Bias both conscious and unconscious, reflecting traditional and unexamined habits of thought, keeps up barriers that must come down if equal opportunity and nondiscrimination are ever genuinely to become the country’s law and practice.” Adarand Constructors, Inc. v. Pena, 515 U.S. 200, 274, 115 S.Ct. 2097, 132 L.Ed.2d 158 (1995) (Ginsburg, J., dissenting).
. As I noted in a related context in Ligón, "this is an area in which further training may be highly beneficial.” Ligón, 925 F.Supp.2d at 530 n. 374. A study of police officers in Savannah, Georgia found evidence that minority suspects were more likely than white suspects to be viewed suspiciously by the officers for nonbehavioral reasons — even when the officers knew they were being closely observed by social scientists while on patrol. See Geoffrey P. Alpert et ah, Police Suspicion and Discretionary Decision Making During *581Citizen Stops, 43 Criminology 407, 417-19 (2005).
. See Fagan 2d Supp. Rpt. at 23 app. tbl. Dl. These are the numbers for 2010 through June 2012. In 2004 through 2009, the numbers were 46% for stops of black people, 42% for stops of Hispanics, and 37% for stops of white people. See Fagan Rpt. app. D tbl. DI.
. See, e.g., Wardlow, 528 U.S. at 124, 120 S.Ct. 673 (citing Brown v. Texas, 443 U.S. 47, 99 S.Ct. 2637, 61 L.Ed.2d 357 (1979); Adams, 407 U.S. at 144, 147-48, 92 S.Ct. 1921).
. See, e.g., 4/1 Tr. at 1687 (Officer Fernando Guimaraes testifying that when he worked in the 43rd Precinct, the entire precinct was a "high crime area”); 4/17 Tr. at 3717-3721 (Officer Edward French testifying that "two robberies with similar circumstances in Queens could constitute a robbery pattern,” and that this pattern could "encompass all of Queens,” and that when he was in the anti-crime unit, there was a crime pattern that encompassed all of Queens).
. See Fagan Rpt. at 53-54 & fig. 13 (summarizing data for 2004 through 2009); Fagan 2d Supp. Rpt. at 32-34 & fig. 13 (summarizing data for 2010 through June 2012). The former analysis was based on precincts, and the latter on census tracts. The fact that the results of the analysis were unchanged despite a change in specifications and time period indicates the robustness of the results.
In otherwise low-crime areas, there may be small areas of concentrated criminality where the NYPD often deploys officers. The City appropriately notes that a particular building where a stop takes place might have a.high incidence of the suspected crime for which the stop is made. See 4/5 Tr. at 2359-2361, But, as Dr. Fagan noted at trial, see id. at 2360-2361, it is simply not plausible that 55% of all stops take place in an “area” whose particular crime characteristics objectively justified checking the High Crime Area box, regardless of the crime rate of the census tract or the precinct where the stop takes place. The unvarying rate of checking High Crime Area across locations and times is more likely a product of reflexive box-checking that is unrelated to any defined crime condition.
. This reveals the fallacy of the critique by Drs. Smith and Purtell that Fagan’s analysis failed to address the "steady improvement in NYPD use of ‘Terry stops.’ ” Smith Rpt. at 49. It may well be that officers simply learned how to fill out the UF-250s to better indicate reasonable suspicion — even when it did not exist.
. See Fagan Supp. Rpt. at 39-47. "Evasive, False Or Inconsistent Response To Officer’s Questions” ("Evasive Response”) is a Side 2 additional circumstance. See Blank UF-250. In theory, High Crime Area could be checked based on an analysis of crime *582data. In practice, however, it is often checked based on inconsistent, subjective impressions or scripts, as discussed above.
. See PX557, 557-D.
. See DX L12, L14; 6/12/13 Plaintiffs' Proposed Findings of Fact and Conclusions of Law ("PL Findings”) ¶ 8.
. Without analyzing the "Other” text strings, 198,000 or 5.7% of the non-radio-run stops between 2004 and June 2012 were “apparently unjustified” according to Dr. Fagan's definition. See PX 417B (updating Fagan Rpt. at 56 tbl. 12 to show 6.45% of 2,233,027 non-radio-run stops between 2004 and 2009 to be "apparently unjustified”); Fagan 2d Supp. Rpt. at 26 tbl. 12 (categorizing 4.43% of 1,215,846 non-radio-run stops between 2010 and June 2012 as "apparently unjustified”). The summary numbers in PX 417D appear to include the results of the "Other” text string analysis. PX 417D states that 219,-814 or 6.4% of the 3,448,873 non-radio-run stops .between 2004 and June 2012 were "apparently unjustified.”
. See, e.g., infra Part IV.D.l.a (Downs); infra Part IV.D.l.h (Lino); Ligón, 925 F.Supp.2d at 499-99; id. at 512 & n. 241. It remains likely that there are " ‘many unlawful searches ... of innocent people which turn up nothing incriminating, in which no arrest is made, about which courts do nothing, and about which we never hear.’ ” Washington v. Lambert, 98 F.3d 1181, 1193 n. 1 (9th Cir. 1996) (quoting Brinegar v. United States, 338 U.S. 160, 181, 69 S.Ct. 1302, 93 L.Ed. 1879 (1949) (Jackson, J., dissenting)).
. In fact, as discussed supra in Part IV.A, the true "hit rate” is likely below 12%.
. Organization for Economic Cooperation and Development Statistical Glossary, available at http://stats.oecd.org/glossary/detail. asp?ID=7228. Accord Black's Law Dictionary (defining benchmark as "a standard unit used as a basis for comparison”).
. Greg Ridgeway, RAND, Analysis of Racial Disparities in the New York Police Department’s Stop, Question, and Frisk Practices xi (2007) ("RAND Report”), DX K6. See also Smith Rpt. at 11 ("Benchmarks help social scientists demonstrate statistically if there is an unjustifiable cause-and-effect relationship between membership in a group ... and a particular practice....”).
. RAND Report at 19.
. Fagan Rpt. at 16-17.
. Id. at 17. See also Fagan Rpt. at 33 tbl. 5; Pi. Mem. ¶ 22. To be clear: Dr. Fagan includes local crime rate data because the *584police are more likely to be deployed to places with higher crime rates, and stops are more likely to take place in areas where the police are more heavily deployed.
. See 5/6 Tr. at 6121, 6135-6139 (Smith). Dr. Ayres’ report on the LAPD is discussed in 3/4/13 Amicus Curiae Brief of Communities United for Police Reform at 6. Drs. Smith and Purtell cited Dr. Fagan’s and Dr. Ayres' scholarship in a short list of "[n]otable articles” in the literature on benchmarking in disparate treatment claims. See Smith Rpt. at 11 n. 5. The City concedes that ”[t]here is no prevailing benchmark for racial disparity regression analysis.” Def. Findings ¶ 69.
. Smith Rpt. at 19 (quotation marks omitted).
. Because crime victims are often unable to provide race information — especially for nonviolent crimes — the City’s experts eventually attempted to supplement their crime suspect race data with data showing the races of arrestees. Dr. Fagan offered persuasive criticisms of the resulting “Merge File” — such as its reliance on arrestee data that may obscure rather than reveal racial bias — as well as of the City’s experts' earlier reliance on extrapolations from incomplete crime suspect data. See PI. Findings ¶ 20 (collecting sources).
. Def. Findings ¶ 70 (citations and emphasis omitted).
. See supra Part IV.A.
. See Fagan 2d Supp. Rpt. at 23 app. tbl. Dl; Fagan Rpt. app. D tbl. Dl. The City’s speculation that the low hit rate for stops overall could be explained by "multiple people [being] stopped in connection with the description provided in a single radio run” neglects the fact that only 13% of stops are based on matching a suspect description. Def. Findings ¶ 90. In addition, the City’s theory is weakened by its failure to collect evidence from the OF-250 database concerning how often multiple stops are made in connection with a single event or a single suspect description. See id.
. 5/6 Tr. at 6155-6156.
. Id. at 6114.
. Id. at 6151, 6154. Accord Def. Findings ¶ 70 ("Crime suspect description data estimates the available pool of persons exhibiting suspicious behavior that could be observed by the police — while population merely estimates the potential number of persons in a given area.” (citations omitted)).
. Drs. Smith and Purtell state:
Obviously, if particular racial or ethnic groups in New York City participate in crime at a rate disproportionate to their share of the population, we would expect officers to conduct Terry stops for such groups at rates higher than each groups’ respective share of the City’s population. The benchmark of suspect race description allows us to measure if [the] NYPD’s officers are stopping minorities at a rate over and above what could be explained by the racial composition of the criminally active population in New York.
Smith Rpt. at 20. Accord id. at 14-15 (suggesting that criminal participation of various races must be incorporated into benchmark, because certain races may commit crimes at a higher rate than others, “and so are more likely to be observed by police engaging in suspicious activity that would justify a Terry stop” (emphasis added)).
. See 4/19 Tr. at 4310-4314 (Assistant Commissioner Philip McGuire stating that "nobody knows” what portion of people who are stopped but not arrested or summonsed "were probably about to or might have committed a crime,” and that "the number could be zero to fifty percent,” but that he does not know); 4/9 Tr. at 2915-2916, 2983-2984 (Chief Esposito stating "we're not really able to tell” how often a stop prevents a crime, then imagining what such a stop might look like). I also note that none of the testifying plaintiffs — even those whose stops were based on reasonable suspicion — were engaged in criminal activity.
. Of course, if the real purpose of the stops is not to investigate suspected criminal activity based on individualized suspicion, but instead to deter criminals from carrying weapons or contraband by stopping people who fit a general profile of criminal suspects in an area, then criminal suspect data would in a sense be the appropriate benchmark for measuring racial disparities in stops. But the City does not make this argument, and if it did, the argument would fail for reasons related to plaintiffs’ Fourth and Fourteenth Amendment claims. First, to the extent that such "deterrence” stops were based solely on a person's resemblance to a general profile of the criminals in an area, the stop would not be based on individualized reasonable suspicion of criminal activity. Second, to the extent that “the general profile of criminal suspects in an area” contains a racial description, the use of that description as a basis for stops would constitute racial profiling. On the issue of racial profiling, see infra Part V.B.l.
. 5/6 Tr. at 6158.
. See Smith Rpt. at 9 (asking rhetorically: "In view of the tens or even hundreds of thousandth] of persons an officer observes in the course of each month on duty, when the overwhelming majority are Blacks and Hispanics, is it plausible to believe that the several persons actually stopped were selected based on their race?”). Plaintiffs and the City dispute the average number of stops per officer per month, with the City claiming it is "only” two to three, and the plaintiffs claiming it is far higher. A review of the numerous *587monthly activity reports in the record suggests that most officers who are actively making stops conduct between 10 and 20 per month. See, e.g., PX 15, 178, 219, 229, 236, 309, 310. But the number of stops is not particularly relevant. It is the quality of the stops that is at issue here — that is, whether the stops were based on reasonable suspicion and free of racial discrimination.
. Similarly, by controlling for various other factors, such as patrol strength, socioeconomic factors, and other local characteristics, Dr. Fagan persuasively undermined a number of other potential race-neutral explanations for the racial disparities in stops.
. For an analysis touching on the prevalence of this stereotype, and the consequences related to it, see generally Michelle Alexander, The New Jim Crow: Mass Incarceration in the Age of Colorblindness (2010).
. 7/19/13 Remarks by President Barack Obama on Trayvon Martin, White House Press Briefing Room.
. Ekow N. Yankah, Op-Ed., The Truth About Trayvon, N.Y. Times, July 16, 2013, at A23.
. The suspect’s race was unknown in 70% of crime complaints in 2005 and 2006. See Fagan Rpt. at 76 tbl. 18. Even after merging crime suspect data and arrestee data, the race of the perpetrator is only known for roughly 63% of crimes. See Fagan 2d Supp. Rpt. app. B tbl. 1-2 (using data from 2010 to 2011); Smith Rpt. at 34. In addition, based on this "merged” data, while the suspect’s race is known for a high percentage of certain types of crime — such as felony violent crimes (86%), weapons crimes (98%), and drug offenses (99%) — race is known in only 22% of felony property crimes, which are the basis for 25% of all stops. See id. app. B tbl. 2. Dr. Fagan persuasively showed that using data where almost 40% of the information is missing would introduce sample selection bias, and is not a reliable approach to drawing conclusions about the criminal suspect population. See id.; 4/3 Tr. at 2148-2150 (Fagan); 5/6 Tr. at 6160-6161 (Dr. Smith acknowledging that he had not found any scholarly support for estimating the demographics of crime suspects based on data that is nearly 40% incomplete).
. See infra Part IV.C.3.
. See 4/3 Tr. at 2029.
. See PL Findings ¶ 15 (citing, among other sources, Dr. Fagan’s testimony; Fagan Rpt. at 4, 40-45 & tbls. 7-10; Fagan 2d Supp. Rpt. at 19-21 & tbl. 7). See also 4/3 Tr. at 2030.
. See 4/3 Tr. at 2031-2032. See also Fagan Rpt. at 66.
. See PL Findings ¶ 17; Fagan Rpt. at 66, 68 fig. 14. I note two reservations about this data: First, it is difficult to control reliably for the suspected crime when by 2009, 36% of UF-250s stated no suspected crime. See Fagan Supp. Rpt. at 39. Second, these "use of force” figures refer to stops in which any use of force was indicated, including "Hands On Suspect.” See Fagan Rpt. at 63-69; Blank UF-250.
. See Pl. Findings ¶ 16 (citing, among other sources, Fagan Rpt. at 66-67 & tbl. 16); 4/3 Tr. at 2123.
. The investigation was also prompted by the shooting and killing of Amadou Diallo, an unarmed twenty-two-year-old West African immigrant, by members of the NYPD's now-disbanded Street Crimes Unit. Diallo was shot during an incident that apparently began as a stop. See The New York City Police Department’s Stop and Frisk Practices: A Report to the People of the State of New York from the Office of the Attorney General (1999) ("1999 AG Report”), PX 333, at 4-5. The 1999 AG Report was admitted only for the purpose of showing that the NYPD had notice of various issues raised in this case, not for the truth of its contents. See 4/19 Tr. at 4283-4284.
. See 1999 AG Report at v-xv, 1, 12. The UF-250s analyzed in the report were the old version of the form. See id. at 89; PX 449 (old UF-250).
. See 1999 AG Report at vii-xv.
. The Report quotes Safir as stating: " ‘The racial/ethnic distribution of the subjects of ‘stop’ and frisk reports reflects the demographics of known violent crime suspects as reported by crime victims. Similarly, the demographics of arrestees in violent crimes also correspond with the demographics of known violent crime suspects.’ ” Id. at 120 n. 26. The Report also highlights an issue discussed later in this Opinion: because only 13% of stops resulted from the stopped person fitting the description of a known criminal suspect, suspect descriptions for violent criminals cannot be the primary driver for NYPD stop activity, which suggests that such descriptions cannot provide a race-neutral explanation for racial disparities in stops overall. See id. at 122 n. 30.
. Id. at 121.
. See id. at 130-131.
. “It is now for the [NYPD] and others interested in a constructive dialogue to review the data and offer their perspectives.” Id. at 175.
. Assistant Commissioner Philip McGuire, who has been in charge of Crime Analysis and Program Planning (CAPPS) in the Office of Management, Analysis and Planning (OMAP) since 1994, disagreed with the Report’s findings because the benchmark it used was the arrest data from 1997, instead of data from 1998 and 1999, the years in which the UF-250s were prepared. Deputy Commissioner Michael Farrell has overseen OMAP and the Quality Assurance Division (QAD) since 2002. He disagreed with the findings in the Report because Dr. Fagan used arrest data instead of suspect description data as his benchmark. See 5/14 Tr. at 7081-7082, 7090-7091. The use of arrest data rather than suspect data would likely hide rather than exaggerate the overstepping of minorities.
. See 4/9 Tr. at 2804.
. See id. at 2807; DX V14-A. The City speculates that this increase may only reflect increased documentation of stops, but offers no evidence for its speculations. See 4/11/13 Defendants] Memorandum of Law in Opposition to Plaintiffs’ Requested Injunctive Relief ("Def. Inj. Mem.”) at 1.
. See, e.g., New York State Division of Criminal Justice Services, Index Crimes Reported to Police By Region: 2003-2012 (2013) (showing 17% drop in index crime reports between 2003 and 2012, and 30% drop in reported murders). I emphasize again that this Opinion takes no position on whether stop and frisk contributed to the decline in crime. The point here is that the dramatic increase in the number of stops cannot be explained by an increase in the number of people displaying suspicious behavior on the streets. To the contrary, the fall in crime reports suggests that stops based on suspicious behavior rose even as the amount of suspicious behavior declined.
. See generally PL Findings ¶¶ 56-85 (collecting many of the sources discussed below).
. See John A. Eterno & Eli B. Silverman, The NYPD’s Compstat: Compare Statistics or Compose Statistics?, 12 Int’l J. Police Sci. & Mgmt. 426-29 (2010). Redacted versions of two tables from the Eterno and Silverman article were admitted into the record as PX 291 and 292. See 5/31/13 Endorsed Letter [Dkt. No. 306].
. See 4/9 Tr. at 2868-2869 (Esposito); 3/29 Tr. at 1511 (testimony that Chief Esposito chaired weekly Compstat meetings). See also 3/22 Tr. at 1030-1031 (Diaz).
. See, e.g., 2008 Compstat Meeting Notes Part A, PX 281, at *7017 (Chief Esposito: "Your [enforcement] numbers are way down.... If you look at [the] raw [number] of 250s[,] you are down 50%.” An executive officer responds that he "[w]ill look at it.”); id. at *7026 (Chief Esposito: "How many C summonses are given out per officer on straight time? What should [the] average be?” A commanding officer responds: "2.3 sir and 2.3 for 250s as well.”); id. at *7080 (Chief Esposito: "I have to go but 9 robberies compared to none[,] I don’t think we are doing enoughf] in that zone. You have 4 C[’]s and 5 250s in [a] 28 day period.”); 2008 Compstat Meeting Notes Part B, PX 283, at *7959 (Chief Esposito noting that there were only "1250 and 4 C [summonses] for [a] whole daytour.” A commanding officer responds: "I[']m on top of it[,] I saw it.”); id. at *8045 (Chief Esposito: "Everyone is working hard, just trouble with the violence, 250s[] down, C[s] down, arrests down. OCCB collars up 16%[,] doing [a] great job.”); id. at *8144 (Chief Esposito: "In and around housing you are down in C's, 250s.” A command*593ing officer responds: “We have called on the [borough] for resources.... ’’). The quotations come from Compstat meeting notes.
. See 3/29 Tr. at 1556. Accord Deposition of Chief of Patrol Robert Giannelli (“Giannelli Dep.”), PX 157, at 268-269 (testifying that at Compstat meetings, Chief Esposito might criticize an inordinately low number of stops in relation to a crime pattern).
. 3/29 Tr. at 1555.
. See, e.g., 4/9 Tr. at 2881 (Chief Esposito testifying that "[i]n the context of CompStat, we always stress quality”); 3/22 Tr. at 925 (Deputy Chief Michael Marino testifying that "[t]he main thing that you hear a lot at CompStat is they talk about quality over quantity. Nobody from the top on down has ever said they want more numbers for numbers’ sake.”); 3/22 Tr. at 1030 (similar testimony from Assistant Chief Diaz).
. See, e.g., Giannelli Dep. at 268-269; 2008 Compstat Meeting Notes Part B at *7958 (Chief Esposito considering redistribution of resources, in part based on low number of OF-250s in area with spike in robberies). Accord 3/22 Tr. at 925 (Deputy Chief Marino illustrating the emphasis on quality at Comps-tat meetings by stating: "They can do things like they can put up the computer maps and show robberies up in this area. And then they will show a lot of activity in this area. No, it should be here. You’re not taking proper steps to stop the conditions.”).
. 2008 Compstat Meeting Notes Part B at *8025.
. 4/9 Tr. at 2893-2894. See also 4/9 Tr. at 2879-2881, 2893-2894 (Chief Esposito credibly clarifying that 2008 Compstat Meeting Notes Part B at *8051-8052 does not question the constitutionality of stops generating a low hit rate).
. See id. at 2895. See also 4/2 Tr. at 1838 (Deputy Inspector Steven Mauriello stating that he did not recall any discussion at a Compstat meeting of "whether the stops and frisks that would be recorded in [a commander’s] 250s are ... legal or constitutional”); 4/16 Tr. at 3544 (Deputy Inspector Charles Ortiz stating that the review of UF-250s at Compstat meetings consisted only of aggregate statistics, not the review of individual forms). For further evidence of NYPD offi*594ciáis' and managers' failure to discuss racial profiling at Compstat meetings, or among themselves, see PL Findings ¶¶ 186-190 (collecting sources).
. Def. Findings ¶ 47 (citing 5/15 Tr. at 7348-7352).
. See 5/15 Tr. at 7349. See also id. at 7348-7350; 5/16 Tr. at 7623.
. Chief Hall ultimately conceded that "[p]rimarily,” the purpose of his and his staff's reviews of UF-250s "would probably be more related to the crime condition than the actual stop.” 5/16 Tr. at 7626; id. at 7621-7626. When asked whether there is "ever an analysis of the circumstances which led to [a] stop” at a Compstat meeting, Assistant Chief Diaz replied "No”: "The CompStat process is basically to look at criminal activity and to see what strategies are in place to address that criminal activity, not so much to look at the quality ... of the stop.” 3/29 Tr. at 1518.
. These surveys were pre-tested and approved by the Institutional Review Board of Molloy College, where Dr. Eterno is an associate dean and professor of criminal justice.
. See 4/5 Tr. 2470-2481 (Silverman testimony regarding 2008 survey); id. at 2503-2509 (same regarding 2012 survey); PX 300 (blank copy of questionnaire used in 2008 survey). Dr. Silverman testified that the original purpose of the 2008 survey was to determine whether NYPD personnel were experiencing pressure to misrepresent crime numbers in order to improve their performance measurements at Compstat meetings. See 4/5 Tr. at 2470. After the publication of the survey results, Commissioner Kelly appointed a three-person committee, including two former federal prosecutors, to prepare a report on "whether the NYPD's internal-auditing and quality-control processes are sufficient to ensure the accuracy” of Compstat. David N. Kelley & Sharon L. McCarthy, The Report of the Crime Reporting Review Committee to Commissioner Raymond W. Kelly Concerning Compstat Auditing i (Apr. 8, 2013). The report finds support for Dr. Silverman’s concerns, see id. at 49-50, and recommends a number of reforms, including officer interviews to complement QAD’s document-focused audits, more accountability for egregious misclassifications of crimes, and formalized periodic external assessments of the Compstat audit process. See id. at 55-58.
. PX 300.
. See PX 291. The mean responses with regard to pressure to increase summonses and arrests were 1.5 points and 1.8 points higher for Compstat-era personnel, and both of these results were also statistically significant. See id.
. The increases in the degrees of reported pressure to increase summonses and arrests were similarly stark. See PX 443; PX 441 (with regard to summonses, reports of low pressure fell from 28% to 9%, while reports of high pressure rose from 26% to 45%); PX 442 (with regard to arrests, reports of low pressure fell from 31% to 8%, while reports of high pressure rose from 17% to 42%); 4/5 Tr. at 2496-2500.
. See PX 293.
. See PX 446 (stops); PX 444 (summonses); PX 445 (arrests).
. See PX 292, 447; 4/5 Tr. at 2518-2519.
. See Def. Findings ¶ 68 (arguing, for example, that the 2008 survey is not "representative” because it "included only retirees who opted to join [a] union mailing list,” without offering any grounds for believing that retirees who joined this mailing list are more or less likely than other retirees to perceive pressure to increase stops).
. See Chief of Patrol, Memorandum Regarding Quota Bill (Oct. 22, 2010), PX 290, at *0096 ("Quotas are not addressed as part of any managerial training....”). The lack of citation to any prior written policy prohibit-mg quotas strongly suggests that no such policy existed.
. To the extent that the officers chose what to record, the recordings may present an incomplete picture. But for the most part, the context of the recordings and the meaning of the supervisors' words were plain.
. See 3/19 Tr. at 425 (Polanco).
. See, e.g., PX 289T (10/30/08 at 4.20-6.30) ("If you see something just do some, uh, 250’s, get all the fucking riff-raff off the corners.”).
. Id. (6/12/08 at 7.13-8.10).
. Id. (6/12/08 at 12.10-13.28).
. Id. (11/8/08 at 13.09-14.36).
. Id. (11/1/08 at 2.12-3.50).
. See 5/9 Tr. at 6458-6459 (Inspector Juanita Holmes, who became the Commanding Officer of the 81st Precinct after the release of Officer Schoolcraft’s tapes in 2010, testifying that the 81st Precinct is "77 percent African-American”); NYC.gov, Community Snapshot 2011, CD3: Bedford Stuyvesant (2012) (stating that Bedford Stuyvesant as a whole is 59% black).
. PX 289T (3/13/09 at 4.32-5.20) (emphasis added). "The hounds” apparently is a reference to superiors, such as commanders, who are monitoring enforcement activity, pushing for higher activity, and have the power to transfer an officer as punishment for low enforcement activity. See, e.g., id. (6/12/08 at 14.58- 16.40); id. (12/8/08 at 12.20-15.00); 4/2 Tr. at 1897-1898.
For another example of Sergeant Stukes' directives at roll call, see PX 289T (1/7/08 at 6.58- 8.00) ("Be an asshole. They going to do something, shine a light in their face whatever the occasion, inconvenience them.”). When asked to explain Sergeant Stukes' statement "[b]e an asshole,” Deputy Inspector Steven Mauriello offered the following interpretation: "It means be a police officer. You have a footpost. You walk your footpost. And be omnipresent.” 4/2 Tr. at 1900.
. PX 289T (11/23/08 at 5.46-6.28) (emphasis added); 4/2 Tr. at 1825. When asked to interpret the phrase "[y]ou can always articulate later,” Deputy Inspector Mauriello at first evaded the question by discussing the NYPD’s commitment to "CPR” (Courtesy, Professionalism, and Respect). Id. at 1903-1904. He eventually responded: "I guess he means do your paperwork later. Articulate. Online booking sheet.” Id. at 1903.
. PX 289T (10/31/08 at 9.05-9.50).
. See 4/2 Tr. at 1912-1913.
. This tactic, of course, contravenes the requirement that an officer must have individualized and articulable reasonable suspicion before making a stop.
. PX 289T (12/8/08 at 1.20-1.38). According to one officer's testimony, there are three types of summonses: A summonses for parking violations ("parkers”), B summonses for moving violations, and C summonses for quality of life offenses (also known as "criminal court summonses”). See 5/10 Tr. at 6765 (Herran).
. PX 289T (11/8/08 at 15.34-15.45). Even if it were an effective gang-suppression strategy to stop every person wearing known gang paraphernalia, it would not be a constitutional strategy, because neither carrying beads nor flaunting a bandana is a crime.
. Id. (12/8/08 at 12.20-15.00). Accord id. (12/8/08 at 7.07-7.42) (suggesting that officers with low activity will be transferred); id. (12/12/18 at 2.20-4.30) (encouraging more OF-250s and C summonses, and stating "your evaluations are based on your activity”); id. (1/28/09 at 23.24-24.10) (“[T]hey’re looking at those numbers and people are gonna be moved.”); id. (1/29/09 at 6.56-9.03) (“You get ... your numbers, and everybody leaves you alone.”).
. PX 284T track 1; 3/19 Tr. at 423-425.
. PX 284T track 1-2. See also 3/20 Tr. at 469-470 (Officer Polanco interpreting Officer Herran as encouraging officers to carry out an arrest on Friday night so that the City will have to pay overtime the following day); 3/21 Tr. at 734-735 (Officer Serrano describing distinction between a "zero” with low activity and a "hero” with high activity).
. See 5/10 Tr. at 6765-6766, 6768.
. See PX 284T track 5; 3/19 Tr. at 417.
. PX 284T track 5. Officer Polanco credibly interpreted this as an instruction to "[s]top and frisk anybody who is on a bike carrying a bag” in the area of the crime pattern. 3/20 Tr. at 487-488.
. Lieutenant Barrett credibly testified that "we were looking for fives, that was any combination of summonses or UF-250s.” 5/7 Tr. at 6272.
. PX 297.
. See PL Findings ¶¶ 61-62 (collecting sources).
. See 4/2 Tr. at 1842.
. Previously, section 215 only addressed quotas for traffic violations. See PX 290 at *0096.
. Id. at *0097.
. Id. (emphasis added).
. See 4/15 Tr. at 3364 (Deputy Commissioner of Labor Relations John Beirne, one of the designers of Quest).
. OO 52, PX 285 ¶¶ 1, 3. For the many ways of referring to low activity numbers, see 3/22 Tr. at 966 (Lieutenant Rafael Mascol testifying that he would tell an underperforming officer: "Listen, you need to be a little more proactive out there, be a little bit busier about doing your assignments out there, handling the conditions that you have out there, you know, you seem to have fallen off. for a few months there, is there something going on?” (emphases added)).
. 4/15 Tr. at 3368-3369.
. 00 52 ¶ 15.
. See Interim Order 49 (Oct. 24, 2011) (‘TO 49”), PX 315 (suspending Patrol Guide 205-57), at 2; Police Officer’s Monthly Conditions Impact Measurement Report, PX 205 (blank form).
. See PX 205. The form provides another example of the NYPD systematically collecting information relevant to evaluating the "quality” of stops in the sense of their effectiveness, but not in the sense of their constitutionality.
. See 10 49 at 2-5; PX 205; 3/22 Tr. at 892-893 (Deputy Chief Marino confirming that officers’ success or failure in meeting performance goals is monitored exclusively through the monthly, quarterly, and annual reports). The monthly and quarterly evaluations also play a significant part in an officer’s annual performance evaluation. See, e.g., 4/15 Tr. at 3410 (Beirne).
. This fact is evident from the form itself, see PX 205, but was also corroborated by testimony. See, e.g., 3/27 Tr. at 1178 (Lieutenant Jonathan Korabel acknowledging “there’s no substantive information [on a monthly activity report] about [the] stops and frisks, arrests, or summonses” tallied on the report).
. For example, the following are notes from supervisors on the back of officers’ monthly report forms in the field labeled "Officer's Impact on Declared Conditions,” which invites supervisors to "[d'[escribe in detail why [the officer] was effective/ineffective.” All of the evaluations are from 2012: "Officer showed improvement from last month and was proactive in combating conditions which resulted in 24 summonses to address conditions.” PX 234 (marked "Effective”). "PO ... was effective for the month with 1 grand larceny ... arrest and 20 UF-250s.” Id. at *1255 (marked “Effective”). "PO ... was effective for the month [with] 1 arrest and 20 UF-250s in target areas.” Id. at * 1257 (marked “Effective”).
. 4/15 Tr. at 3370-3372. In contrast to Deputy Commissioner Beirne, Chief Esposito and Deputy Chief Marino evaded plaintiffs’ questions about numerical enforcement activity goals, and insisted that officers are evaluated on "takfing] care of the condition,” 4/9 Tr. at 2957 (Esposito). Accord 3/22 Tr. at 881 (Marino). Chief Esposito and Deputy Chief Marino offered this testimony even though the Quest program contains no metrics for measuring "addressing conditions” other than the tallies of an officer's activities. I also note that an arbitration proceeding found that Deputy Chief Marino himself imposed quotas on officers when he was the Commanding Officer of the 75th Precinct. See 4/9 Tr. at 2954-2955 (Chief Esposito acknowledging Marino arbitration findings).
. See OO 52 ¶¶ 1, 3.
. 4/15 Tr. at 3369 (distinguishing quota and performance goal).
. See id. at 3399-3400 (Beirne); 5/10 Tr. at 6790-6791 (Herran).
. PX 290 at *0096. The original includes "quota" rather than "performance goal,” but the meaning is the same.
. Terry, 392 U.S. at 14 n. 11, 88 S.Ct. 1868 (quoting President’s Commission on Law Enforcement and Administration of Justice, Task Force Report: The Police (1967)). See also id. at 14-15, 88 S.Ct. 1868 (noting the problem of ”[t]he wholesale harassment by certain elements of the police community, of which minority groups, particularly Negroes, frequently complain”).
. Lambert, 98 F.3d at 1188 (collecting scholarly sources).
. See, e.g., Interim Order 20 (5/16/12), PX 183 ¶¶ 2-4 (current racial profiling policy, emphasizing that stops must be based on reasonable suspicion); Operations Order 11 (3/13/02), PX 184 ¶¶ 1-2 (previous racial profiling policy, emphasizing same).
. PX 332T at 20. NYPD personnel of diverse ranks repeated variations on this phrase throughout the trial. See PI. Findings ¶¶ 49-55 (collecting sources); 4/10 Tr. at 3035 (Chief Esposito agreeing that the NYPD looks for “the right people, at the right place, at the right time”).
. See infra Part V.B.l (conclusions of law regarding racial profiling).
. 4/10 Tr. at 3034. To be clear, there is nothing constitutionally problematic about targeting "the right place at the right time,” where this means that deployment should "mirror the time of the crime and the place of the crimes that are being committed.” PX 157 at 268-269 (Giannelli).
. 4/10 Tr. at 3027-3029 (emphasis added). Chief Esposito later testified that stop activity targets "the people that are committing the crimes.” Id. at 3029. While I find that the NYPD has defined "the right people” to be stopped in terms of race, gender, and age, I note that it would be equally problematic if the NYPD instructed officers to target "the right people” or "the people that are committing the crimes” without defining these categories. This would invite officers to fill in the undefined terms with their own stereotypes and biases regarding what a criminal looks like.
. Chief Esposito does not view targeting “the right people” as a form of racial profiling, because in his view, racial profiling cannot exist provided that stops are based on reasonable suspicion:
As I think about, since this has been going on since '08, yeah, I think if you look at that form, if it's filled out properly, it gives you reasonable suspicion, and if you have reasonable suspicion established, then you do not have racial profiling. It’s as simple as that.
4/9 Tr. at 2824. Of course, it is also erroneous to suggest that properly filling out a UF-250 necessarily establishes reasonable suspicion.
. PX 332T at 21.
. Id. at 23-24. Mott Haven is a mostly black and Hispanic housing development in the Mott Haven neighborhood in the Bronx. See id. at 22-23; 5/13 Tr. at 7014 (McCormack).
. See 5/13 Tr. at 7016-7017.
. See PX 332T at 23. I also note that Deputy Inspector McCormack does not dis*605play the contempt or hostility toward the local population that appears in many of the Schoolcraft recordings. To the contrary, Deputy Inspector McCormack emphasizes his belief "that 99 percent of these people in this community are great, hardworking people,” and makes clear that he has no interest in targeting blacks or Hispanics as such: his goal is to focus stop activity on "whatever group” is committing targeted crimes. Id. at 8, 11.
. Similarly, Deputy Inspector Stephen Cirabisi testified that at Compstat meetings, attention is paid to whether the people being stopped are the same as “the people that are suspected of committing the crimes.” 5/2 Tr. at 5696-5697.
. See 4/17 Tr. at 3743; Def. Findings ¶ 13.
. See infra Part IV.D.l.c.
. See PX 557; PX 557-D; 5/7 Tr. at 6327-6328 (Lieutenant Charlton Telford). "Radio Run/Sprint # ” is a field on the front of the UF-250, while "Report From Victim/Witness” and "Ongoing Investigations, e.g., Robbery Pattern” are boxes in the Additional Circumstances section on the back of the form. See Blank UF-250.
. 5/7 Tr. at 6327-6328.
. Id. at 6341.
. PX 557-D. Similarly, Officer Dang stopped 120 blacks and 0 whites during a sample quarter in 2009, despite the fact that he was patrolling a 43% black precinct. See *606DX L14; Raymond W. Kelly, Police Commissioner, Reasonable Suspicion Stops: Precinct Based Comparison by Stop and Suspect Description, DX Y8, at *4974.
. DX Y8 at *4974.
. 4/1 Tr. at 1589. Defendants did not object to this out of court statement.
. Id. at 1585-1589.
. Mayor Bloomberg Delivers Address on Public Safety to NYPD Leadership (4/30/13) ("April 30, 2013 Bloomberg Address”), PX 583 (emphasis added).
. 3/18 Tr. at 44 (Heidi Grossman).
. Id. at 45.
. See Def. Mem. at 7; supra Part IV.B.2 (discussing weaknesses of UF-250s for verifying constitutionality of stops); Fagan Rpt. at 53-55.
. Stop and Frisk Report, PX 449. As one supervisor stated on a recording: “We used to have to write a two-page story on the damn thing.... You had to write the whole story.... Now it’s easy. You just check a couple boxes off.” PX 289T (1/29/09 at 6.56-9.03).
. See Blank UF-250.
. See Operations Order 44 (9/11/08), PX 96, at 1 (emphasizing the importance of activity logs, stating that supervisors must inspect subordinates’ activity logs for "accuracy and completeness at regular and unspecified intervals,” and noting that failure to make required entries may result in discipline); id. at 2 (stating that "IT IS IMPERATIVE THAT A DETAILED ENTRY BE MADE AS INDICATED” for each stop, such as: "susp. male randomly looking in apt. windows”); Patrol Guide: Stop and Frisk at 1 (noting that officer is required to enter details of stop in activity log); 3/27 Tr. at 1099 (B. Dennis) (stating that all officers are required to carry a memo book, which is also called an activity log).
. 3/27 Tr. at 1099-1100. Accord id. at 1141-1142 (B. Dennis).
. Id. at 1101-1102; PX 19 (memo book).
. See 3/27 Tr. at 1103.
. See id. at 1100-1101. Similarly, Officer French did not believe he needed to state the reasons for a stop in his memo book, as long as he noted the suspected crime. See 4/17 Tr. at 3737. For numerous other examples of credible testimony concerning inadequate memo book entries, supervisors’ failures to discuss these deficiencies, and superiors' failures to address the problem, see PL Findings ¶¶ 97, 99 (collecting sources).
. See 4/17 Tr. at 3796-3797.
. 4/9 Tr. at 2912. Accord, id. (Chief Esposito testifying that "[i]n a perfect world, I think this is a hundred percent acceptable. But we're not in a perfect world out there.”); id. at 2929 (Chief Esposito testifying that memo book entries providing more detail than a UF-250 would be "[flor the most part" redundant, and that "if an officer checked off furtive movement,” that would be "enough” for him).
. See PX 58, 71, 89 (Worksheets 802 and 802-A, and instructions for completing them). Reviewers are instructed that an activity log entry does not need to include any information about the reasons for a stop beyond what already appears on the UF-250. See 4/23 Tr. at 4645 (Cronin).
. See 4/23 Tr. at 4624.
. See PL Findings ¶ 197 (citing PX 450; DX G6); 4/23 Tr. at 4651 (testimony that any score below a three is considered failing).
. See, e.g., 4/10 Tr. at 3213-3219 (Inspector Donald McHugh, Commanding Officer of 41st Precinct); 4/16 Tr. at 3527-3531 (Deputy Inspector Charles Ortiz, Commanding Officer of 43rd Precinct); Pl. Findings ¶ 198 (collecting sources).
. PX 155 at 93-94, 118-120 (Dale deposition).
. 4/24 Tr. at 4845. Accord id. (Reiter testifying that "[fit’s like everybody sticks their head in the sand and hopes that passing [a] memo up through the chain of command and back down will somehow change it.”).
. See 2013 Memorandum of Chief of Patrol James Hall, DX J13. The nine categories of information are: “Date/time of stop; Location of stop; Suspect’s Last name, First name; Suspect's pedigree; Suspected crime or offense {felony or penal law misdemeanor); Explanation of suspicion (looking into windows, pulling on doorknobs, eic[.]); Whether or not the suspect was frisked; Sprin1/Job number; Disposition of stop {96, 92C, 93Q, etc.).” Id.
. See Ligón, 925 F.Supp.2d at 519 n. 291.
. See Def. Findings ¶ 54 (citing testimony of Chief Hall and Chief William Morris). This evidence included the review by NYPD personnel of a sample of forty UF-250s. According to the City's witness, all forty were accompanied by memo book entries, but sixteen contained "either no entry or an inadequate entry regarding the reason for the stop” — the essential element in determining whether a stop was legally justified. 5/9 Tr. at 6578-6580. In addition, because plaintiffs were not given an opportunity to review the forty UF-250s and accompanying memo book entries, see id., it is possible that more than sixteen lacked adequate entries regarding the reason for the stop.
. Cf. United States v. Oregon State Med. Soc., 343 U.S. 326, 333, 72 S.Ct. 690, 96 L.Ed. 978 (1952) ("It is the duty of the courts to beware of efforts to defeat injunctive relief by protestations of repentance and reform, especially when abandonment seems timed to anticipate suit, and there is probability of resumption.”).
. See 9/24/03 Stipulation of Settlement, Daniels v. City of New York, No. 99 Civ. 1695, PX 114, ató.
. See PX 58, 71, 89 (Worksheets 802 and 802-A, and instructions); Finest Message Regarding Compliance with Operations Order 11 § 2 (12/26/02), PX 350 (ordering self-inspection based on Worksheets 802 and 802-A); 4/23 Tr. at 4630; PI. Findings ¶¶ 108-121; Def. Findings ¶ 24.
. See PX 89 at 2, 4.
. See 4/23 Tr. at 4650 (Cronin).
. See, e.g., 3/22 Tr. at 1053-1055 (Assistant Chief Diaz conceding that officers engaged in racial profiling are unlikely to write on a UF-250 " 'I stopped this person because they’re black’ ”). As early as 1999, Commissioner Safir recognized the inadequacy of relying solely on paperwork prepared by an officer to monitor the constitutional adequacy of the officer’s stops. See PX 46 at 48. In addition, NYPD commanders and supervisors testified at trial that UF-250s do not provide enough information to determine whether reasonable suspicion existed for a stop. See, e.g., 4/10 Tr. at 3207 (McHugh); PI. Findings ¶ 118. Accord Def. Mem. at 7.
. Defending the current paperwork-only audit process, Deputy Commissioner Farrell rejected a proposal to include field observations of stops in the audit. See 5/15 Tr. at 7297-7299. QAD is aware of practical methods for evaluating police conduct without relying purely on paperwork, such as the methods it uses to audit the treatment of crime complaints, and complaints about courtesy, professionalism, and respect. In both cases, QAD makes phone calls to a sample of civilians to evaluate their encounters with the police. See 4/24 Tr. at 4792 (Cronin); 5/15 Tr. at 7291-7293 (Farrell).
. 4/23 Tr. at 4640-4641, 4720-4721.
. Def. Findings ¶ 47. Accord PI. Findings V 86; 4/9 Tr. at 2841-2842, 2919 (Chief Esposito testifying that supervisory review mitigates the risk posed by incomplete or inaccurate documentation of stops); id. at 2929-2930 (Chief Esposito testifying that he relies on supervisors as his "main ... way of determining” whether the NYPD’s officers are acting in accordance with the policy against racial profiling and the policy of only making stops based on reasonable suspicion).
. See, e.g., Def. Findings ¶ 47 (collecting some sources).
. Id.
. As noted above, 78% of the stops during the class period were self-initiated rather than based on a radio run. See PX 417D.
. Sergeant Stephen Monroe, when asked directly how he ensures "that officers under [his] supervision are conducting lawful stops based on reasonable suspicion,” responded: "I usually ... observe the stops looking for safety and their approach of the suspect." 4/29 Tr. at 5266-5267 (emphasis added). When asked what he discusses with officers regarding their stops, Sergeant Monroe again emphasized officer safety and effectiveness, making no mention of constitutionality. See id. at 5267; 4/17 Tr. at 3778. In addition, the supervisors who witnessed the unconstitutional stops in this case failed to recognize or respond to their unconstitutionality. See, e.g., 3/27 Tr. at 1145-1147, 1176-1179 (Sergeant Korabel testifying that he was supervising Officer Dennis in the anticrime unit during the stop of Devin Almonor); infra Part IV.D.l.b (concluding that Almonor’s stop and frisk was unconstitutional).
Sergeant Loria testified that during his twelve years as an anticrime sergeant, he was present "[v]ery frequently” for stops by his officers. 4/17 Tr. at 3755-3756, 3789-3792. Nevertheless, he testified that he had never told a single officer that a stop or frisk was improper. See id. at 3778. Sergeant Loria also testified at his deposition that as a sergeant, he did not review his officers’ UF-250s for reasonable suspicion, and that he could not recall anyone in the NYPD ever telling him to do so. See id. at 3770; 4/8 Tr. at 2706 (Officer Leek testifying that his supervising Sergeant never exited the police van during the February 24, 2011 stop of Clive Lino). See infra Part IV.D.2.d.
. Patrol Guide 212-11, Stop and Frisk (7/18/03) ("P.G. 212-11”), PX 98 at 1; Patrol Guide 212-11, Stop and Frisk (5/24/02), Ex. 4 to Giannelli Dep.
. See Giannelli Dep. at 166-167. Chief Giannelli also testified that he did not require desk officers to ask police officers about the facts of a stop. See id. at 168.
. See P.G. 212-11 ¶ 10.
. See Lesson Plan, Sergeants Leadership Course (4/30/09), DX R3; 4/29 Tr. at 5168-5170 (Chief James Shea unable to explain why, contrary to his testimony that sergeants are trained to review every street stop for reasonable suspicion, lesson plan does not instruct newly promoted sergeants to do this); 4/2 Tr. at 1937, 1939, 1951-1953 (Sergeant Richard Hegney testifying that he does not recall receiving any training regarding stop and frisk after he was promoted to sergeant in 2000). I also note that the Quest for Excel*612lence program, as discussed above, does not require supervisors to review the constitutionality of officers’ stops. See supra Part IV.C.4.a; 4/15 Tr. at 3385-3390 (Deputy Commissioner Beirne conceding that nothing in Orders 49, 50, or 52 directs a supervisor to review stops for constitutionality).
. See, e.g., PL Findings ¶ 94 (collecting sources); 4/17 Tr. at 3763 (Sergeant Loria's deposition testimony that when he reviews UF-250s, he ”make[s] sure that it's signed and all the boxes are filled in,” but does not review it for anything else); 3/20 Tr. at 634 (Officer Edward Velazquez testifying that supervisors had only returned UF-250s to him based on incompleteness, such as failing to record the date); 3/22 Tr. at 986-990 (Sergeant Julio Agron ultimately conceding that he does not believe an Operation Impact sergeant is required to review his officers’ UF-250s at the end of a tour).
. See, e.g., Pl. Findings ¶ 100 (collecting sources); 4/16 Tr. at 3522 (Deputy Inspector Ortiz testifying that he did not require supervisors to review activity logs alongside UF-250s).
. See, e.g., PL Findings ¶¶ 95-96 (collecting sources); 3/20 Tr. at 606-607 (Officer Victor Marrero testifying that when a sergeant reviews a UF-250, the officer is usually present, but that the only question the sergeant asks is the time of the occurrence); 3/20 Tr. at 634 (Officer Velazquez testifying that no supervisor had ever questioned him about the circumstances of a stop based on a UF-250). In addition, supervisors do not notice stop patterns that suggest unconstitutionality. See Pl. Findings ¶ 87. For example, Lieutenant Tel-ford testified that as an anticrime sergeant, he reviewed his officers' UF-250s. However, he was unaware that two of his officers were among the four top issuers of UF-250s in the NYPD in the third quarter of 2009 — including Officer Gonzalez, who checked the same four boxes on 99% of his UF-250s. See 5/7 Tr. at 6314-6315. Sergeant Joseph Marino was unaware that Officer Dang, one of the anticrime officers under his supervision, was another of the officers who conducted the most stops in the third quarter of 2009. See 4/30 Tr. at 5555-5556.
. See Def. Findings ¶ 48.
. Giannelli Dep. at 47. See generally 4/16 Tr. at 3581-3582 (Lieutenant Enno Peters, ICO of the 28th Precinct, stating that ICOs are usually lieutenants, and assistant ICOs are usually sergeants).
. For example, ICOs observe officers in the field responding to radio runs, and at least some ICOs review officers’ memo books.
. For example, Lieutenant Cosmo Palmieri, ICO of the 43rd Precinct, testified that despite his regular field observations of officers, during his five years as an ICO he has never observed an officer perform a stop. See 4/17 Tr. at 3673. Lieutenant Peters testified at a deposition that as ICO of the 28th Precinct, he could not remember having ever discussed reasonable suspicion with members of the precinct. See 4/16 Tr. at 3583-3584.
. See Patrol Guide 202-15, Command Integrity Control Officer ("P.G. 202-15”), DX *613F5. Paragraph 16 of P.G. 202-15 requires ICOs to inspect and sign the activity logs of sergeants, but not of officers, and not specifically in relation to stops. I note that the burden of carrying out the thirty-one duties listed on P.G. 202-15 make it highly unlikely that ICOs would have the time or resources to review the constitutionality of stops unless specifically trained and directed to do so.
. See Giannelli Dep. at 48-50, 106.
. See 4/17 Tr. at 3675 (Lieutenant Palmieri testifying that as an ICO, he has not reviewed UF-250s).
. See, e.g., 5/2 Tr. at 5692-5693 (Cirabisi); PL Findings ¶¶ 102-107. The City solicited lengthy testimony concerning the duties and activities of ICOs. See, e.g., Def. Findings ¶ 48 (collecting sources). As with so much of the City’s evidence concerning institutions at the NYPD, however, this testimony was largely irrelevant because it did not relate to the constitutionality of stops.
. See, e.g., Pl. Findings ¶¶ 89, 91 (collecting some sources showing that high-level managers, such as precinct commanders, did not direct supervisors and ICOs to monitor the constitutionality of stops).
. See supra Part III.
. See Def. Findings ¶¶ 43, 45 & nn. 46, 48 (collecting sources).
. The distinction between official policy and operational policy is common in the study of law enforcement. See 4/24 Tr. at 4834-4835 (Reiter).
. See generally Pl. Findings ¶¶ 122-134.
. Police Student’s Guide, Policing Legally: Street Encounters (July 2012) ("Student’s Guide: Street Encounters"), DX Qll, at 17.
. See infra Part IV.D.l.g.
. 4/18 Tr. at 4042 (Moran).
. Id. at 4046-4049.
. See infra Part IV.D.l.g. See also infra Part IV.D.l.b (unconstitutional Almonor stop based in part on "furtive movements”); 5/9 Tr. at 6431-6433 (Officer Dang explaining that "usually” a furtive movement is someone "hanging out in front of [a] building, sitting on the benches or something like that” and then making a “quick movement,” such as "bending down and quickly standing back up,” "going inside the lobby ... and then quickly coming back out,” or "all of a sudden becoming] very nervous, very aware”).
. See Detective Benito Gonzalez, Characteristics of Armed Subjects (“Armed Subjects Powerpoint”), DX C8, at 44-57.
. See 4/25 Tr. at 5040-5043 (Shea).
. See Armed Subjects Powerpoint at 4; Lesson Plan, Firearms and Tactics Section (3/12) (“Firearms Lesson Plan”), DXW3, at 1, 20; 4/29 Tr. at 5176-5177 (Shea).
. See infra Part IV.D.l.a-d (Leroy Downs frisked based on presence of keys, wallet and bag of cookies in his pocket; Officer Brian Dennis checked "suspicious bulge” on the UF-250 after the fact, when Devin Almonor had only his cell phone in his pocket; Cornelio McDonald stopped based on "suspicious bulge” that turned out to be cell phone; Nicholas Peart stopped in August 2006 based on anonymous suspect description corroborated only by "suspicious bulge” that turned out to be cell phone).
. See Fagan Rpt. 57, 64 tbl. 15.
. See Johnson, 555 U.S. at 326-27, 129 S.Ct. 781.
. Firearms Lesson Plan at 2. If an officer’s fear for her safety is unrelated to any suspicion that the stopped person may be armed, there is no reason to conduct a protective frisk for weapons, and such a frisk would be unjustified. For example, an officer may not frisk a stopped person simply because the stopped person is physically imposing, or because the stop took place in a dangerous neighborhood. The only justification for a protective frisk is to discover weapons. See generally supra Part III.B.3 (noting that the frisk standard in CPL § 140.50 does not reflect the constitutional standard for a frisk).
. See, e.g., 4/17 Tr. at 3867 (Officer Mahoney, who frisked Downs, testifying that the standard for whether a stopped person may be frisked is "[w]hether it was a violent crime or you have reasonable ... fear for your safety or the safety of ... [a] civilian”); 4/10 Tr. at 3117-3120 (Officer Luke White testifying, although somewhat unclearly, that his frisk of Dominique Sindayiganza was justified based on- reasonable suspicion that Sindayiganza had committed aggravated harassment, even though this is not a violent crime and Officer White did not suspect that Sindayiganza had a weapon). See also infra Parts IV.D.l.a, IV.D.2.a (detailing Downs's and Sindayiganza’s stops and frisks).
. See infra Parts IV.D.1-2.
. See DX V 14-A; DX V 14-C.
. See PI. Findings ¶ 126; 4/25 Tr. at 5161-5162 (Shea).
. See PX417D.
. See generally infra Part V.B.l (conclusions of law regarding racial profiling). For the NYPD’s racial profiling training, see, for example, Police Student’s Guide, Policing Impartially, DX VI1, at 3-7, 12-14 (appropriately drawing attention to latent bias even among well-intentioned officers, but offering an inadequately narrow definition of racial profiling: "when a police officer decides to stop and question a person when the sole rationale for the contact is the race, ethnicity, or national origin of the person being stopped” (emphasis added)).
. See Ligón, 925 F.Supp.2d at 519-20, 532-40; PL Findings ¶¶ 130-131. I stated in that opinion that more than three thousand officers had attended the training since its development in 2012. See Ligón, 925 F.Supp:2d at 519-20. More than six thousand officers have now attended. See 4/25 Tr. at 5121 (Shea).
. See Ligón, 925 F.Supp.2d at 537-38.
. See 4/25 Tr. at 5126, 5151-5154 (Shea). The individual stops in this case show that officers frequently fail to fill out OF-250s after encounters that were clearly stops. See, e.g., infra Part IV.D.l.a (no OF-250 for Downs stop); infra Part IV.D.l.h (no OF-250 for February 5 Lino stop). See also Ligón, 925 F.Supp.2d at 498-99, 512 & n. 241.
. See 4/29 Tr. at 5212 (Detective Damian Vizcarrondo testifying that an encounter rises to the level of a stop not based on whether the stopped person is free to leave, but based on whether the officer has reasonable suspicion); 4/25 Tr. at 4971 (Lieutenant James McCarthy testifying to the same). The Rodman’s Neck training materials, which are reflected in Detective Vizcarrondo’s and Lieutenant McCarthy’s testimony, teach the following lesson: if an officer has an encounter with a civilian and is wondering whether to fill out a OF-250, the officer should ask himself whether the encounter was based on reasonable suspicion — not whether a reasonable person would have felt free to leave. This approach is legally incorrect and will predictably lead to officers not filling out OF-250s after encounters that began based on something less than reasonable suspicion, but that in retrospect involved a level of coercion amounting to a stop. Chief Shea argued that it makes more sense to train officers never to conduct stops without reasonable suspicion than it does to train officers to prepare a OF-250 even when a stop was not based on reasonable suspicion. See 4/25 Tr. at 5119-5124 (Shea). But officers can and should be trained to do both. That is, they should be trained never to conduct stops without reasonable suspicion and trained to recognize when an encounter that was intended to be a "Level 1” request for information or "Level 2” inquiry under De Bour in fact resulted in a Terry stop. Whenever this occurs, whether intentionally or unintentionally, the encounter must be recorded as a stop.
. Allegations of improper stops or frisks fall within the category of "abuse of authority” complaints. See 4/15 Tr. at 3284 (Joan Thompson, Executive Director of CCRB); CCRB, January-June 2012 Report ("CCRB June 2012 Report”), DX V13, at 4 (defining "Abuse of Authority”).
. See Def. Findings ¶ 25.
. See infra Part IV.D.l.a. Downs took extraordinary measures to identify the police officers despite the lack of any official record of the stop.
. See Def. Findings ¶ 25 & n. 34 (noting that the NYPD refers all FADO complaints to the CCRB, including "[s]earch and seizure allegations relating to stop, question, and frisk,” and refers "the remainder of the complaints” to the OCD); id. ¶ 32 (noting that the OCD processes a small number of "complaints featuring] allegations related to” stop and frisk).
. See id. ¶ 25.
. Def. Findings ¶ 26.
. CCRB June 2012 Report at 4.
. See id. An officer is "exonerated” if she committed the alleged acts, but the acts “were determined to be lawful and proper,” and an allegation is "unfounded” if there is sufficient evidence that the officer did not *618commit the alleged act. Id. Accord 4/15 Tr. at 3272 (Thompson).
. See Def. Findings ¶ 26; CCRB, 2011 Annual Report ("CCRB 2011 Annual Report”), DX PI2, at 17-18 (describing prosecutorial function of the DAO). In the first half of 2012, the CCRB closed 2,518 cases. Of the 27% of cases closed after full investigations, 11 % resulted in a finding of substantiated misconduct. See CCRB June 2012 Report at 10.
. 4/22 Tr. at 4485-4486.
. Id. at 4486-4487.
. Id. at 4508.
. See id. at 4484, 4486-4487, 4508.
. See id. at 4508-4511.
. Deputy Commissioner Schwartz's later testimony also conflicted with her deposition testimony in another case. See id. at 4511-4512.
. See ¿d. at 4512-4513.
. See id. at 4481-4482, 4485-4486, 4496; 4/15 Tr. at 3289, 3292 (Thompson).
. Deputy Commissioner Schwartz also testified that the law governing search and seizure, and especially De Bour, is not "clearly established, well-articulated, and understandable,” and when an officer violates an unclear law like De Bour unintentionally, the CCRB should consider the officer's "good faith” in its credibility determination. 4/22 Tr. at 4513-4521. In addition, she testified that when the CCRB has failed to identify an officer by name and has not conducted a show-up or a photo array, the DAO will dismiss the case rather than attempting to determine the identity of the officer based on specific identifying information in the CCRB’s findings. See id. at 4484-4485.
. See CCRB 2011 Annual Report at 17; CCRB, Police Department Discipline, DX U13. These figures exclude the small number of cases where the statute of limitations had expired or the officer resigned before the NYPD could take action. Several officers testified that they were never disciplined even after the CCRB substantiated complaints against them. See PL Findings ¶¶ 139-141 (collecting sources).
. See 4/15 Tr. at 3294 (Thompson); CCRB 2011 Annual Report at 18, tbl. 30. Instructions as the most common form of discipline in every year. Chief Hall testified that an officer who engages in stops that are not based on reasonable suspicion could be subjected to "incredibly severe discipline,” up to and including termination. 5/16 Tr. at 7628. When asked on cross-examination whether he had personal knowledge of any officer being subjected to such discipline for a stop lacking reasonable suspicion, Chief Hall said that he did not. See id. at 7629.
. See 3/19 Tr. at 345 (Nicholas Peart testifying that after his first CCRB complaint was not substantiated, he did not file any CCRB complaints in response to later stops because he did not believe that the CCRB would do anything). See also Pl. Findings ¶ 155 (collecting sources indicating the DAO’s awareness of longstanding public concerns that the NYPD does not take the CCRB’s recommendations seriously); id. ¶ 156 (noting that even after the NYPD agreed in 2012 to allow CCRB attorneys to prosecute a small category of serious police misconduct cases, they can only do so at the discretion of the Police Commissioner). The City’s proposed findings describe at great length the NYPD’s performance monitoring system, which tracks officers who have received multiple substantiated CCRB complaints. See Def. Findings ¶¶ 36-42; Pl. Findings ¶ 138. However, this system is ineffective for monitoring unjustified stops because of the low likelihood that a wrongfully stopped person will have the knowledge and take the time to file a CCRB complaint. The likelihood of multiple complaints against the same officer is even lower.
. See Pl. Findings ¶¶ 142-147 (collecting sources).
. See 4/18 Tr. at 3962-3968, 3981, 4013 (Inspector McAleer). In addition, there was evidence that complaints of racial profiling made during stops are not conveyed to the OCD. See Tr. 5/7 at 6307, 6332-6333 (Lieutenant Telford testifying that as an anticrime supervisor in 2009, he was not concerned after Officer Gonzalez submitted a UF-250 indicating that a black male he had stopped had asked: "Why don’t you stop other people?’’); PX 557 at *15999-16000.
. See 4/17 Tr. at 3822 (Rothenberg); infra Part IV.D.l.f (detailing Provost stop and frisk).
. See 4/17 Tr. at 3820, 3825 (Rothenberg).
. See 4/30 Tr. at 5385 (stipulation).
. See, e.g., 4/18 Tr. at 3962-3963 (Inspector Helen McAleer, OCD, testifying that CCRB forwards non-FADO complaints to OCD, including racial profiling complaints, of which there have been ”[v]ery few”); 4/15 Tr. at 3271 (Thompson testifying that the CCRB sometimes forwards racial profiling complaints to the OCD).
. 4/24 Tr. at 4878.
. See 4/10 Tr. at 3023-3027. Chief Esposito also testified that he has been aware of public controversies concerning street encounters between NYPD officers and blacks and Hispanics since at least 2000, when he became Chief of Department. See 4/9 Tr. at 2794-2796.
. 3/22 Tr. at 931.
. See 4/30 Tr. at 5410, 5429-5430, 5434-5435.
. See 5/10 Tr. at 6633.
. See 4/25 Tr. at 5081.
. Id. at 5081-5082.
. See, e.g., 3/22 Tr. at 1055-1056 (Diaz testifying that he is not sure); 4/10 Tr. at 3025-3026 (Esposito); 4/30 Tr. at 5434 (Lehr); 5/9 Tr. at 6511 (Holmes); 5/10 Tr. at 6633 (Morris); Def. Findings ¶ 23.
. See, e.g., 5/14 Tr. at 7174 (Deputy Commissioner Farrell testifying that he was aware in early 2007 of "concerns raised by [some] members of the public that the NYPD may be engaging in racially biased policing”).
. When asked why he submitted the letter anonymously, Officer Polanco explained that he was worried by the adverse consequences Officer Schoolcraft suffered after reporting the alleged manipulation of crime statistics. See 3/20 Tr. at 453.
. 3/20 Tr. at 450, 453, 508-509. The letter does not appear in the record. See id. at 450, 661; 3/19 Tr. at 435-436.
. 3/20 Tr. at 450-451, 509. See also id. at 460-461 (Officer Polanco testifying that his letter to IAB stated the address, date, and other information concerning the encounter with the thirteen-year-old).
. Id. at 457-458. See also id. at 459-460 (Officer Polanco testifying that after an unjustified stop, supervisors instructed him about which boxes to check on the UF-250, including High Crime Area and Furtive Movements).
. See id. at 455-456.
. See id. at 462-464. For a summary of some of Officer Polanco’s recordings, see supra Part IV.C.2. Eventually, Officer Polanco abandoned his anonymity and participated in a televised interview. See 3/20 Tr. at 515-516. After the interview aired in March 2010, IAB charged him with perjury. The charge was based on an incident described in Officer Polanco’s complaint to IAB. Officer Polanco stated that he was once instructed to issue a summons to someone for walking a dog without a license, even though he had not seen a dog, and that he obeyed the instruction. See id. at 451-452, 540. While Officer Polanco was charged with perjury, the supervisor who issued the instruction was promoted. See id. at 540.
I note that Officer Craig Matthews has alleged in another case that the NYPD retaliated against him for complaining to superiors about " 'a system of quotas mandating numbers of arrests, summonses, and stop-and-frisks’ ” that was allegedly introduced in 2008. See Matthews v. City of New York, No. 12 Civ. 1354, 957 F.Supp.2d 442, 445-46, 2013 WL 3879891, at *1 (S.D.N.Y. July 29, 2013) (quoting Complaint ¶ 2). Officer Matthews' allegations of retaliation echo Officer Polanco’s allegations. Compare 3/20 Tr. at 575-576 (Officer Polanco testifying that foot-post assignment was punitive for someone with his seniority), and 3/21 Tr. at 727, 789 (Officer Serrano testifying to same), with Complaint ¶21, Matthews, 957 F.Supp.2d 442, 2013 WL 3879891 (describing footpost assignment as punitive). While recognizing that "as a matter of fact, Officer Matthews’ speech had undeniable value to the public,” the Court granted summary judgment to the City based on First Amendment grounds unrelated to the substance of Officer Matthews’ allegations. Matthews, 957 F.Supp.2d at 465, 2013 WL 3879891 at *19.
. See 3/28 Tr. at 1424-1425 (Officer Hernandez testifying regarding CCRB complaints about his stop activity).
. See Daniels, No. 99 Civ. 1695. See also 4/9 Tr. at 2800-2801 (Chief Esposito testifying that, as of 1999, he was aware of the allegations in Daniels of racial profiling and stops and frisks that lacked reasonable suspicion).
. I note that plaintiffs originally intended to pursue this case as a challenge to the NYPD’s non-compliance with the Daniels settlement.
. See Davis, No. 10 Civ. 0699; Ligón, No. 12 Civ. 2274.
. Def. Findings ¶ 20. See RAND Report. I note that the City only relies on the RAND Report as proof that it lacked notice of racially motivated stops, but not that it lacked notice of stops made without reasonable suspicion. See Def. Findings ¶ 20. I also note that the RAND Report was admitted only for the purpose of notice, not for the truth of its conclusions.
. See supra Part IV.B.3.
. See 4/19 Tr. at 4302 (Assistant Commissioner McGuire); Fagan Rpt. apps. C4-C6 (noting that 15% of stops from 2004 to 2009 were based on suspicion of violent crime); Fagan 2d Supp. Rpt. app. B tbl. 2 (noting that 23% of stops from 2010 to 2011 were based on suspicion of violent crime). The earlier chart shows that 15% of all UF-250s from 2004 to 2009, including those that failed to state a specific suspected crime, were based on suspicion of a violent crime; by contrast, the later table appears to show that 23% of the UF-250s from 2010 to 2011 containing a codable suspected crime were based on suspicion of a violent crime. Compare Fagan Rpt. app. C4-C5 noting that the suspected crime on 18% of UF-250s was uncodable or otherwise erroneous), with Fagan 2d Supp. Rpt. app. B tbl. 2 (not including UF-250s with erroneous crime codes in any category). Thus, the 15% figure appears to be a more accurate indication of the percentage of stops that are based on suspicion of a violent crime.
. See RAND Report at 19 (noting that "black suspects were described in 69 percent of all violent-crime suspect descriptions”); Fagan Rpt. at 76 tbl. 18 (revealing, apparently based on updated data, that the suspect's race was black in 73% of the violent crime reports in 2006 indicating a suspect's race, versus 54% for non-violent crimes). The RAND Report’s claim that black suspects were described in 69% of all violent-crime suspect descriptions in 2006 is not strictly accurate, because no race was indicated in 47% of those complaints. See Fagan Rpt. at 76 tbl. 18, 77 (noting that contrary to the language in the RAND Report, black was in fact "identified as the suspect’s race in only 38.50% of all violent crime complaints ... in 2005,” once the complaints lacking race data are included). As Dr. Fagan testified, the lack of race information in 47% of the complaints further undermines the reliability of the 2006 violent crime suspect data as a benchmark. See 4/4 Tr. at 2266-2268; Fagan Rpt. at 75-77.
. RAND Report at 19 (emphasis added).
. Id. at 44.
. See 4/9 Tr. at 2830-2831 (Esposito); 5/14 Tr. at 7121 (Deputy Commissioner Farrell testifying that "it was our conclusion that the differences that did exist were quite small”). Throughout his testimony, Deputy Commissioner Farrell refused to acknowledge the validity of statistical evidence of racial disparities in stops and stop outcomes, no matter how large or persistent the disparities. See, e.g„ 5/15 Tr. at 7248-7256.
. See RAND Report at 44-46. For example, the NYPD began a pilot project to distribute information cards to stopped pedestrians, as recommended by the RAND Report. See 5/14 Tr. at 7106-7107; DX A8 (tear-off card used in pilot project). Although Deputy Commissioner Farrell testified that the pilot project had been expanded citywide, none of the individuals stopped in this case testified to receiving a card, and the card itself does not provide a number for making complaints to the CCRB or the NYPD. Instead, the card states: "For more information ... www.nyc. gov/nypd,” and directs recipients of the card to call a toll-free number or 311 with tips about illegal handguns. DX A8. See also PL Findings ¶¶ 179-182 (collecting sources regarding the NYPD's failure to implement the RAND Report’s recommended early warning system for officers who overstop minorities).
. See Terry, 392 U.S. at 22, 88 S.Ct. 1868 (“The scheme of the Fourth Amendment becomes meaningful only when it is assured that at some point the conduct of those charged with enforcing the laws can be subjected to the more detached, neutral scrutiny of a judge who must evaluate the reasonableness of a particular search or seizure in light of the particular circumstances.”).
. See 4/19 Tr. at 4094-4095 (Downs). The following account is largely taken from Downs’s testimony, which I found credible despite minor inconsistencies.
. See id. at 4095-4097.
. See id. at 4097-4098.
. Officer Giacona was wearing a black t-shirt and Officer Mahoney was wearing a New York Jets jersey that said “Favre” on the back.
. See id. at 4098-4102.
. See id. at 4101-4106.
. See id. at 4106-4108.
. See id. at 4108-4111.
. See id. at 4110-4111, 4113; PX 166-D, 166-CL. Officer Mahoney was still wearing his Brett Favre jersey. Officer Giacona’s records confirm that he arrested two young, black men that evening.
. See 4/19 Tr. at 4111-4112; PX 166-C.
. See 4/19 Tr. at 4114-4115. Downs later spoke with a Deputy Inspector at the 120th Precinct, who told him the stop "was probably an isolated incident,” and invited Downs to attend the citizens' police academy, to better understand the NYPD’s practices. Downs attended the academy for eleven months. At the end, he still believed the August 20 stop violated his rights. See id. at 4116-4119, 4124-4126.
. See id. at 4116; PX 166 (CCRB file for Downs’s complaint).
. See 4/17 Tr. at 3847-3850, 3852-3853, 3865, 3870 (Mahoney); id. at 3876-3879 (Giacona); Tr. 4/18 at 3895 (Giacona). I refer to Sergeant Mahoney as Officer Mahoney because that was his rank at the time of the stop.
. See Tr. 4/17 at 3849-3850 (Officer Mahoney); id. at 3875-3875 (Officer Giacona); Tr. 4/18 at 3892-3896 (Officer Giacona) ("According to CCRB I guess or whatever their investigation, they found that I did this stop that I don't remember or have any knowledge of.... So that’s — I guess that’s what I was disciplined for, not having memo book entries for the stop that I don't remember doing.”).
. See 4/19 Tr. at 4119 (Downs).
. See Dickerson, 508 U.S. at 373, 377-78, 113 S.Ct. 2130.
. See United States v. Jones, 269 F.3d 919, 924 (8th Cir.2001) (“Fourth Amendment intrusion 'must be temporary and last no longer than is necessary to effectuate the purpose of the stop' and ... the officer should employ the least intrusive means available to dispel the officer’s suspicion in a timely fashion.”) (quoting Florida v. Royer, 460 U.S. 491, 500, 103 S.Ct. 1319, 75 L.Ed.2d 229 (1983)).
. Because Officers Mahoney and Giacona disclaim any memory of the incident, I cannot draw any inferences in their favor.
. See 3/18 Tr. at 111-114 (Almonor). Minor inconsistencies between Almonor's deposition and trial testimony, such as whether he began to walk his friend to the bus stop at 8 p.m. versus 8:45 p.m., see id. at 137, or whether the bodega was at 141st or 142nd Street, see id. at 140, do not undermine his credibility.
. See id. at 143-144.
. See id. at 115-117.
. See id. at 119-124. Malik texted Almonor that he was at the bodega.
. See id. at 134-135. Neither officer saw Almonor with a group of males.
. Korabel is now a Lieutenant, see 3/27 Tr. at 1068 (B. Dennis), but I will refer to him as Sergeant Korabel, as that was his title at the time of the stop.
. See id. at 1085-1087, 1115. The officers arrived at Hamilton Place about twenty minutes after the last 911 call. See id. at 1096.
. See id. at 1117. Accord id. at 1189 (Korabel).
. See id. at 1086 (B. Dennis).
. See id. at 1117; id. at 1150 (Korabel).
. See id. at 1119.
. See 3/18 Tr. at 125-126 (Almonor).
. See id. at 127; 3/27 Tr. at 1069, 1087 (B. Dennis).
. 3/27 Tr. at 1088, 1121 (B. Dennis). Officer Dennis thought Almonor was twisting his body as if to keep his right side away from the officers — a behavior called "blading.” See id. at 1144.
. See id. at 1121-1122. Sergeant Korabel testified that the reason for handcuffing Almonor was that his conduct rose to the level of disorderly conduct. See id. at 1162 (Korabel). However, he testified that he did not observe any disorderly conduct prior to the frisk. See id. at 1164, 1169. Moreover, Officer Dennis testified that Almonor was not yelling or screaming when he was being questioned and frisked. See id. at 1090 (B. Dennis). While Almonor may have struggled in response to being grabbed by the arm and pushed up against a car, I do not find credible the assertion that Almonor’s behavior constituted disorderly conduct.
. Id. at 1089-1090. Officer Dennis testified that he frisked Almonor after handcuffing him and that the basis for the frisk was that Almonor's "actions as he was walking down the street and as I continued to observe him by holding his waist, I suspected that he may have had a weapon at that point, and that’s the area that I frisked.” Id. at 1093. See also id. at 1161 (Korabel) (Officer Dennis frisked Almonor within a few seconds of exiting the car). I do not find credible that Almonor was walking as if he had a weapon in his waist.
. See id. at 1091 (B. Dennis).
. See id. at 1205 (Korabel) (the name of the individual Sergeant Korabel tried to call was Ms. Al); 3/18 Tr. at 145-146 (Almonor) (acknowledging that he initially identified himself as Devin Al).
. See 3/18 Tr. at 128-129. Almonor’s age was relevant because if he was sixteen or older, the officers could issue a summons for disorderly conduct, but if he was younger than sixteen, he would have to be released to his parents. See 3/27 Tr. at 1123 (B. Dennis).
. See 3/27 Tr. at 1090 (B. Dennis).
. See id. at 1124. Accord 3/18 Tr. at ISO-131, 143 (Almonor). Almonor disputes that he resisted efforts by the police to search, cuff or put him in the squad car. See id. at 157 (Almonor).
. See 3/27 Tr. at 1125 (B. Dennis).
. See id. at 1070-1077.
. See id. at 1103.
. Id. at 1081. Officer Dennis acknowledged that the officers did not have probable cause to arrest Almonor. See id. at 1088. I *630do not believe Sergeant Korabel's testimony that they had probable cause based on Almonor’s alleged jaywalking in violation of New York laws. In any event, Sergeant Korabel acknowledged that the officers did not intend to arrest Almonor for jaywalking. See id. at 1152-1155 (Korabel).
. See id. at 1125-1126 (B. Dennis). Officer Dennis acknowledged that he did not see a suspicious bulge on Almonor. See id. at 1078-1079.
. See Lambert, 98 F.3d at 1190-91 ("If the general descriptions relied on here can be stretched to cover [plaintiffs] then a significant percentage of [black] males walking, eating, going to work ... might well find themselves subjected to similar treatment.”).
. See El-Ghazzaivy v. Berthiaume, 636 F.3d 452, 457-58 (8th Cir.2011) (investigative stop unconstitutional where there was no indication that plaintiff was armed and dangerous, and officer "failed to conduct even the most basic investigation into the facts prior to handcuffing and frisking”); Lambert, 98 F.3d at 1188 (stating that "handcuffing substantially aggravates the intrusiveness of an otherwise routine investigatory detention”) (quotations omitted).
. See Singleton v. United States, 998 A.2d 295, 302 (D.C.2010) (A "generic bulge in a pocket can be explained by too many innocent causes to constitute 'reasonable' suspicion.”).
. See Terry, 392 U.S. at 28-29, 88 S.Ct. 1868 ("The Fourth Amendment proceeds as much by limitations upon the scope of governmental action as by imposing preconditions upon its initiation.”).
. See 4/17 Tr. at 3677 (McDonald). McDonald is employed at Cavalry Staffing where he drives cars. See id.
. See id. at 3678.
. See id. at 3679-3680.
. See id. at 3678-3679. McDonald testified that about 80% of the residents on McDonald’s side of the block are white, while about 80% of the residents on his mother’s side of the block are black. The basis for this testimony was McDonald’s own perception rather than documented statistical information. See id. at 3701.
. See id. at 3679. It was December 19 when he left.
. See id. at 3681-3683, 3696.
. See id. at 3680-3681.
. Id. at 3683.
. Although French has been promoted to detective, I refer to him by his rank at the time of the stop. See id. at 3739 (French).
. See id. at 3685, 3709 (McDonald).
. See id. at 3686. Accord id. at 3750 (French) (testifying that he frisked the outermost jacket and the pocket in which he observed the suspicious bulge, including placing his hand inside the pocket, because once McDonald removed his hands from his pockets, there was still a suspicious bulge). I do not believe that a cell phone created a "suspicious” bulge.
. See id. at 3706-3707 (McDonald).
. See id. at 3687.
. See id. at 3689.
. See id. at 3726 (French).
. See id. at 3737.
. See id. at 3726, 3743. The only suspect description was black male. See id. at 3743.
. See id. at 3747.
. The testimony about a "suspicious bulge” appears to have been an after-the-fact justification and is unsupported by either contemporaneous documents or the objects McDonald had in his pockets.
. See id. at 3688-3689, 3701 (McDonald). McDonald has brought several lawsuits based on racial discrimination in the past. Specifically, he alleged that the postal service falsified the records of African Americans and sued a white male for racial discrimination. See id. at 3691.
. See id. at 3726.
. Wardlow, 528 U.S. at 124, 120 S.Ct. 673 (citing Texas, 443 U.S. at 47, 99 S.Ct. 2637).
. See 4 LaFave § 9.5(h)("the time and spatial relation of the 'stop' to the crime is an important consideration in determining the lawfulness of the stop”) (collecting cases).
. See 3/19 Tr. at 300-302 (Peart).
. See id. at 319-320.
. See id. at 320, 347-348. Peart testified that he only recently found the shorts he was wearing that day while doing laundry. See id. at 320.
. See id. at 320. Officer White, who has since been promoted to detective, was in a vehicle with Officer Fontanez. See 5/7 Tr. at 6211-6213 (White). On August 5, Officer White responded as backup to a radio run reporting crime in progress involving a man with a gun in the vicinity of 96th and Broadway. See id. at 6214. Three cars responded. See id. at 6215-6216.
. See 3/19 Tr. at 320-321 (Peart). Accord 5/7 Tr. at 6224 (White).
. 5/7 Tr. at 6224 (White). Officer White testified that he ordered Peart and his companions to get on the ground because they did not initially comply with the order to raise their hands and because he believed one or more of them might be armed.
. See id. at 6225 (Officer White stating that he had to ask two or three times before the men got on the ground, during which time they asked "Why? For what? What did I do? I didn’t do nothing.”).
. See 3/19 Tr. at 322-323, 326-327 (Peart).
. See 5/7 Tr. at 6226 (White).
. See id. at 6227; 3/19 Tr. at 324-326 (Peart).
. See 3/19 Tr. at 324-326 (Peart); 5/7 Tr. at 6256-6258 (White) (discussing the radio run description). The officers asked central command to repeat the description several times in an effort to convince the three individuals *634that they were stopped was because they fit the suspect description and location provided in the call. See 5/7 Tr. at 6219-6220, 6227 (White).
. See DX Z8 (radio transmissions).
. See 3/19 Tr. at 324-326 (Peart).
. See 5/7 Tr. at 6227 (White).
. See 3/19 Tr. at 326 (Peart). In his formal interview with the CCRB on August 26, 2006, Peart said that he had received a laceration on his lip when he dropped to the ground in response to the officers' commands. However, he corrected this testimony at trial, explaining that the reason he lied was that he was eighteen and wanted to be taken seriously. See id. I find this explanation plausible and it does not undermine Peart’s generally credible trial testimony.
. See id. at 370.
. See id. at 345. Peart did not file any CCRB complaints concerning his later stops because he did not feel that the CCRB would do anything, given that his initial complaint had not been substantiated. See id.
. See DX Z8; 5/7 Tr. at 6219 (White).
. See 5/7 Tr. at 6219-6221. Although Officer White could not recall exactly what the men were wearing, he testified that ''[t]hey were wearing exactly what central had put over which [] was very unique.” Id. at 6222.
. See id. at 6228. The officers attempted to locate the caller but were unsuccessful. See id.
. See id. at 6230. Officer White acknowledged that he couldn't stop someone based on the description “male blacks with a gun,” or solely based on an anonymous call reporting criminal conduct. He explained that he would need additional contributing factors and that, in this case, those factors were the bulges in the waistband and proximity to the crime location. See id. at 6234.
. See id. at 6231. Officer White did not check violent crime suspected on all three forms, which he attributed to "officer oversight.” Id. at 6231-6232.
. See id. at 6232.
. 529 U.S. 266, 268, 120 S.Ct. 1375, 146 L.Ed.2d 254 (2000). The report described three black males, one of whom was wearing a plaid shirt. See id.
. Id. at 272, 120 S.Ct. 1375 (explaining that "[a]n accurate description of a subject's readily observable location and appearance is of course reliable in th[e] limited sense [that it] will help the police correctly identify the person whom the tipster means to accuse. Such a tip, however, does not show that the tipster has knowledge of concealed criminal activity”). Under the reasoning in J.L., Officer White’s belief that he stopped the right individuals, although they were not armed, does not establish reasonable suspicion.
. Id.
. See Simmons, 560 F.3d at 103.
. See 5/7 Tr. at 6223 (White).
. Accord Singleton, 998 A.2d at 302 ("[A] generic bulge in a pocket can be explained by too many innocent causes to constitute 'reasonable' suspicion.”) (distinguishing cases in which only a "noticeable bulge” was cited from cases in which the officer identified with specificity the characteristics that led him to believe it was a firearm).
. No officers were identified in connection with the April 13, 2011 stop of Peart and no UF-250 or other form exists. My findings are *636based entirely on Peart’s testimony, which was both detailed and consistent in identifying which officer did what.
. See 3/19 Tr. at 303-307. Peart was not certain of the second officer’s race but testified that the officer was not black. See id. at 305.
. See id. at 307-311.
. See id. at 312, 397-398. The officers asked if he was coming from building 129, and he explained that he was coming from building 125. See id. at 409.
. See id.
. See id. at 313-314.
. See id. at 315-316, 399. Peart testified that he was upset and concerned that the other officer was inside the building with his younger siblings. See id. at 316.
. See id. at 317-318.
. See id. at 399-400. The door of Peart’s building had recently been replaced, requiring all residents to obtain a new key, which was an involved process requiring proof of residency in the building. See id. at 391-394.
. See 4 LaFave § 9.5(h) (discussing requisite specificity for suspect descriptions).
. See El~Ghazzawy, 636 F.3d at 457 (Terry stop must be conducted in a manner " 'reasonably related in scope to the circumstances which justified the interference in the first place.' ”) (quoting Terry, 392 U.S. at 19-20, 88 S.Ct. 1868); Allen v. City of Los Angeles, 66 F.3d 1052, 1057 (9th Cir.1995) (“'The relevant inquiry is always one of reasonableness under the circumstances.' " (quoting United States v. Sanders, 994 F.2d 200, 206 (5th Cir.1993))).
. See Dickerson, 508 U.S. at 378, 113 S.Ct. 2130 ("The officer's continued exploration of [plaintiffs] pocket after having concluded that it contained no weapon was unrelated to the sole justification of the search under Terry: the protection of the police officer and others nearby.”) (original alterations and quotations omitted). Because the initial frisk provided no evidence of a weapon or immediately perceptible contraband, it could not have justified a further search of his pockets or wallet.
. 10/10/12 Deposition of Ian Provost ("Provost Dep.”), PX 584, at 17-19. Provost did not testify at trial. His testimony was received through his pretrial deposition.
. See id. at 33-36, 82-83. Both Officer Rothenberg and Sergeant Houlahan testified that crowds sometimes gathered and became violent in response to police activity in the area.
. See id. at 33-36.
. See id. at 38-41.
. See id. at 34-35, 39, 42.
. See id. at 42-46. Provost testified that he had seen the officers from the window of his girlfriend's apartment before he left. See id. at 44. See also 1/8/13 Deposition of Sergeant Daniel Houlahan ("Houlahan Dep.”), DX Q14, at 41.
. See Provost Dep. at 45-46. Accord 4/17 Tr. at 3807 (Rothenberg).
. See Provost Dep. at 46-48.
. See id. at 49.
. 4/17 Tr. at 3807 (Rothenberg).
. Id. Officer Rothenberg claims that he was using suspicion of criminal trespass and questioning Provost about where he was going to divert attention from the fact that he knew Provost was carrying a knife. See id. at 3809. In fact, the officer had no basis to suspect that Provost had committed criminal trespass.
. See Provost Dep. at 48-51. Provost could not recall whether his cell phone was in his hand or in his pocket but noted that he was not inclined to reach into his pockets while talking to police because he believed police had a tendency to shoot people who reach into their pockets and pull things out while being questioned. See id. at 51-52.
. See id. at 58.
. See id. at 59, 60.
. See id. at 58.
. See id.
. See id. at 64-65.
. See id. at 59.
. See Houlahan Dep. at 46-49.
. See 4/17 Tr. at 3814, 3832 (Rothenberg). Rothenberg admitted that Provost was not being disorderly prior to being stopped and that he was not engaging in "tumultuous behavior” as defined in the disorderly conduct statute.
. See id. at 3802; PX 277 (memo book). Rothenberg testified that he mistakenly wrote criminal trespass because he was rushed. See *6394/17 Tr. at 3833. See also Houlahan Dep. at 85-86 (stating that if Officer Rothenberg engaged Provost because of a knife in plain view, he should have recorded it in his activity log). Criminal possession of a weapon was also noted in the memo book but not as a basis for the stop and no further details were recorded. See PX 277.
. See 4/17 Tr. at 3804.
. See id. at 3805.
. See Provost Dep. at 92-93.
. See id. at 74.
. As I have recognized, memo books, while imperfect, are the only contemporaneous documents by which a judge or supervisor can evaluate the basis for the stop — that is their primary purpose. Therefore, I have no choice but to credit them over after-the-fact testimony of what occurred, which often incorporates post hoc justifications for the stop.
. Because Provost was not arrested prior to Sergeant Houlahan's arrival, the search of Provost’s person and cell phone cannot be justified as part of a search incident to arrest. The stated purpose of calling Sergeant Houlahan was to determine whether or not Provost should be arrested.
. See 4/19 Tr. at 4173-4177 (Ourlicht).
. See id. at 4175-4176, 4185.
. Moran was promoted to Sergeant in 2010, but I will refer to him by the position he held at the time of the stop. See 4/18 Tr. at 4027 (Moran).
. See 4/19 Tr. at 4177-4180 (Ourlicht); 4/18 Tr. at 4051, 4055 (Moran).
. See 4/19 Tr. at 4181 (Ourlicht) ("Q: Why didn’t you just walk away when [the] officer started asking you questions? A: He was a police officer. He had a gun. I don’t know anybody that would in that situation, walk away.”).
. See id. at 4180-4181. Ourlicht disputes that he was yelling or being hostile, see id., while Officer Moran testified that Ourlicht was irate, screaming and using obscene language. See 4/18 Tr. at 4059 (Moran). Although I believe that Ourlicht may have used obscene language in response to receiving the summons, I credit Ourlicht's testimony that he was not disorderly during the initial stop. At most he became irate after being frisked. See id. at 4069 (Moran) ("A: In the beginning I started to question — asked him a question. Began a conversation. I frisked him. He became very irate. He was yelling obscenities ... He stated he wanted to fight me.”).
. See 4/19 Tr. at 4183 (Ourlicht).
. See id.
. See id. at 4184.
. See 4/18 Tr. at 4081 (Moran).
. 4/19 Tr. at 4148 (Ourlicht).
. See id., at 4185. Moran testified that based on threats Ourlicht was making, "after [Officer Moran] had frisked him and found nothing, [he] reasonably suspected at that time that [Ourlicht] was hiding something else”). See 4/18 Tr. at 4070.
. See 4/19 Tr. at 4186 (Ourlicht).
. See id. at 4187-4188.
. See id. at 4189-4190.
. See id. at 4191; PX 248 (criminal court summons). Moran never handcuffed Ourlicht.
. See 4/19 Tr. at 4192 (Ourlicht).
. See 4/18 Tr. at 4079 (Moran).
. See id. at 4051-4052.
. See id. at 4056.
. See id.; PX 250 (UF-250 form).
. See 4/19 Tr. at 4068 (Moran); PX 249 (Memo Book Entry).
. Wardlow, 528 U.S. at 124, 120 S.Ct. 673 (citing Texas, 443 U.S. at 47, 99 S.Ct. 2637).
. I note that because Officer Moran’s memo book and the UF-250 did not state what the suspicious bulge actually was, I conclude that it was either entirely fabricated, or was, in fact, the five subject notebook.
. Terry, 392 U.S. at 25-26, 88 S.Ct. 1868 (contrasting a search based on probable cause, which is justified on grounds other than protecting the arresting officer, including obtaining evidence of the crime, with "[a] search for weapons [which must] be strictly circumscribed....”). Nor does Terry authorize stopping an individual, antagonizing him, and then conducting a more intrusive search based on post-stop allegations of disorderly conduct. Even Moran doesn’t contend that Ourlicht was being disorderly before he frisked him. Ourlicht admits that he questioned the basis for the stop and the need for ID but not that he was being disorderly. Because Officer Moran's memo book had no details and his memory is limited, I credit Ourlicht, who acknowledged that he used obscenity after receiving .the summons and is generally more credible.
. See 4/1 Tr. at 1728-1729, 1762 (Lino).
. See 4/16 Tr. at 3468-3471, 3485 (Arias). See also id. at 3479 (there are higher crime rates near subways stations).
. I note that this description is so generic as to cover the vast majority of young black males.
. See 4/10 Tr. at 3045, 3063-3064 (Kovall). Officer Kovall testified that the man running "looked like' — -I don’t want to say a normal person, but of a medium stature, as [did] Mr. Lino and the other gentleman stopped.” Id. at 3077.
. See 4/1 Tr. at 1729-1730, 1752, 1765.
. See id. at 1731-1732.
. See 4/10 Tr. at 3050, 3069.
. See 4/1 Tr. at 1732-1733, 1768 (Lino).
. See 4/10 Tr. at 3048 (Kovall); 4/16 Tr. at 3473 (Arias).
. See 4/1 Tr. at 1734 (Lino). I credit this testimony both because Lino’s recollection of the stop was generally credible and because Officer Arias testified that the particular corner of 103rd and Lexington was a high crime, high traffic area. See 4/16 Tr. at 3468-3471, 3485, 3479 (Arias).
. Lino conceded that neither officer checked his pockets. See 4/1 Tr. at 1734-1735. The officers also frisked James. See id.
. See 4/10 Tr. at 3071 (Kovall).
. Id.
. 4/1 Tr. at 1735-1736. Accord 4/16 Tr. at 3491-3492 (Arias).
. See 4/10 Tr. at 3072 (Kovall).
. See 4/1 Tr. at 1735-1736 (Lino). But see 4/16 Tr. 3492 (Arias); 4/10 Tr. at 3072 (Kovall) (testifying that they did not see the other officers frisk either man). I credit Lino’s testimony that Kovall and Arias were sent to run the names because he said they returned and reported that he had an outstanding summons.
. See 4/1 Tr. at 1736-1737 (Lino).
. See id. at 1738; 4/16 Tr. at 3492 (Arias).
. See 4/1 Tr. at 1774 (Lino).
. See 4/16 Tr. at 3486-3489 (Arias).
. See id. at 3490. Officer Arias did not fill out a UF-250 and did not follow the requirements for filling out his memo book in connection with this stop. See id. at 3475 (Arias); PX 214.
. See 4/10 Tr. at 3053-3054 (Kovall); PX 211.
. Contrast 4 LaFave § 9.5(h) (5th ed.) (discussing Hampleton v. United States, 10 A.3d 137 (D.C.App.2010) (finding reasonable suspicion where robbers were described only as "black males in dark clothing,” but direction of flight and distance from the crime location fit, and defendant was stopped in close time/ space relationship to that event, and "was the only one in the immediate area who fit the lookout description”)) and United States v. Turner, 699 A.2d 1125 (D.C.App.1997) (reasonable suspicion where officers arrived at scene of crime within a minute and encountered only two persons fitting fairly general description).
. While observing the men standing facing the street across from the check-cashing location with no apparent agenda for an extended period of time might give rise to the suspicion *645that they were casing a location for a robbery, several minutes is insufficient.
. Lino's rights were further violated when he was detained for additional time by the plainclothes officers. By the time these officers arrived, Lino had dispelled any suspicion that he and James were engaging in criminal activity by showing that they were in fact waiting for Chinese food. See Royer, 460 U.S. at 500, 103 S.Ct. 1319 (Terry stop must "last no longer than is necessary to effectuate the purpose of the stop”).
. The Supreme Court held that the Fourth Amendment does not require police to use the least intrusive means possible to verify their suspicions, reasoning that such a rule would "unduly hamper the police's ability to make swift, on-the-spot decisions.” Sokolow, 490 U.S. at 10-11, 109 S.Ct. 1581. However, this deference only applies when the suspicion to be verified is sufficient to justify a Terry stop. See id. at 10, 109 S.Ct. 1581 (holding that the circumstances gave rise to reasonable suspicion that the suspect was smuggling drugs, and only then turning to whether police were required to use least intrusive means possible to investigate). See also El-Ghazzawy, 636 F.3d at 457-58.
. I note that because the suspected crime was robbery at gunpoint, if the stop had been reasonable at its inception, the frisk would also have been justified, particularly in light of the men’s hesitancy to remove their hands from their pockets, which occurred before the men were frisked.
. See 4/8 Tr. at 2634-2635 (Clarkson).
. See id. Clarkson was not sure of the exact day he was stopped. No CCRB complaint was filed and no OF-250 or memo book entry was identified in connection with this stop. The officers who conducted the stop were never identified. Therefore, my findings regarding this stop are based entirely on Clarkson's testimony, which I find credible.
. See id. at 2636-2637.
. See id. at 2638-2641.
. See id. at 2641-2642.
. See id. at 2643.
. Pinto-Montoya v. Mukasey, 540 F.3d 126, 132 (2d Cir.2008) (discussing Delgado, 466 U.S. at 218, 104 S.Ct. 1758).
. Id. (citing Brendlin v. California, 551 U.S. 249, 254, 127 S.Ct. 2400, 168 L.Ed.2d 132 (2007) ("A person is seized ... when the officer, by means of physical force or show of authority, terminates or restrains his freedom of movement through means intentionally applied.”) (quotation marks and citations omitted)).
. See Wardlow, 528 U.S. at 124, 120 S.Ct. 673 (citing Texas, 443 U.S. at 47, 99 S.Ct. 2637) (presence in high crime area insufficient).
. See 4/8 Tr. at 2586-2587 (Sindayiganza).
. See id. at 2588-2592, 2598, 2606. Sindayiganza testified that the officers looked like they were in their mid-twenties and questioned him like rookie cops. See id. at 2597. See also 4/10 Tr. at 3087-3088 (White); 4/15 Tr. at 3420 (Gillespie) (both testifying that the majority of the officers in the impact squad to which they were assigned at the time of the stop were recent graduates of the police academy).
. See 4/8 Tr. at 2592-2595 (Sindayiganza). Officer White spoke first with the Petco employee and then directly with the woman. The woman did not say that she was verbally threatened or touched and she did not identify herself when she initially spoke with Officer White. See 4/10 Tr. at 3093-3094 (White).
. See 4/8 Tr. at 2596-2597 (Sindayiganza).
. See 4/10 Tr. at 3115-3116 (White) (testifying that although he may have frisked Sindayiganza before arresting him, it was all part of a continuous event and acknowledging that he testified in his CCRB report that he patted down Sindayiganza before the woman informed him that she wanted to press charges).
. See 4/8 Tr. at 2597-2599 (Sindayiganza); 4/10 Tr. at 3102 (White). Officer White clarified that he did not specifically point out Sindayiganza to the woman, but rather asked her if she recognized anyone. See 4/10 Tr. at 3105 (White).
. See 4/10 Tr. at 3106-3107 (White).
. See 4/8 Tr. at 2599-2600 (Sindayiganza). See id. at 2616. Accord 4/10 Tr. at 3108 (White).
. See 4/10 Tr. at 3146 (White).
. See 4/8 Tr. at 2600 (Sindayiganza). Just before handcuffing Sindayiganza, Officer White spoke to the woman who identified Sindayiganza and told her that Sindayiganza was refusing to walk north, at which point she told Officer White that she wanted to press charges. Until she said this, Officer White had no intention of arresting Sindayiganza, although he believed he had probable cause to do so based on the witness’s identification. See 4/10 Tr. at 3112 (White).
. See 4/8 Tr. at 2600-2602 (Sindayiganza).
. See id. at 2623-2624. Sindayiganza explained that he did not want to provide a sworn statement to the CCRB until his summons was resolved because he was unsure how legal proceedings worked. See id. at 2629.
. See 4/10 Tr. at 3104, 3143 (White).
. See id. at 3117.
. See id. at 3112-3113, 3120. One of the officers obtained the woman's personal identification prior to handcuffing Sindayiganza but Officer White threw it out once it was determined that Sindayiganza would not be arrested. See id. at 3123-3124.
. See id. at 3121. No UF-250 was filled out. See id. at 3125. Officer White recorded the arrest in his memo book and noted that the arrest occurred about fifteen minutes after he stopped Sindayiganza. See id. at 3131.
. See Williams, 407 U.S. at 147, 92 S.Ct. 1921 (holding that "when the victim of a street crime seeks immediate police aid and gives a description of [her] assailant,” that may form the basis for reasonable suspicion and a forcible stop).
. See Def. Findings ¶ 2 (describing Sindayiganza stop).
. See PL § 240.26 (defining harassment in the second degree as "following] a person in or about a public place or places; or ... engaging] in a course of conduct ... which alarm[s] or seriously annoy[s] such other person and which serve[s] no legitimate purpose”). Although Officer White stated that he had probable cause to arrest Sindayiganza for aggravated harassment, nothing about the *649woman's statement suggested any fear of physical injury, much less any of the elements of aggravated harassment. See id. § 240.30 (aggravated harassment, even in the second degree, involves use of the mails or telephone, or use of physical violence).
. See CPL § 140.10 (commentary) ("Where [violations] are concerned ... while an arrest, even for an offense punishable only by a fine, does not violate federal constitutional Fourth Amendment rights ..., [New York law] bars an arrest based upon a report made to the officer about something that occurred outside the officer's presence.”). Accord People v. Solomon, 31 A.D.3d 1178, 817 N.Y.S.2d 819 (4th Dep't 2006) ("The warrantless arrest of defendant for a violation, i.e., harassment, that did not occur in the presence of the arresting officers was illegal.”). Although arrest for the violation would not have violated Sindayiganza’s Fourth Amendment rights per se, the fact that the officers lacked authority to arrest Sindayiganza under New York law forecloses using search incident to arrest as a basis for the stop.
. The Second Circuit has upheld the constitutionality of the search incident to arrest conducted prior to an arrest where probable cause to arrest existed and an arrest was later effectuated. See United States v. Wilson, 94 Fed.Appx. 14, 17 (2d Cir.2004) ("Once probable cause was established, it is irrelevant whether the officers’ searches of Wilson occurred prior or subsequent to his arrest.”); United States v. Jenkins, 496 F.2d 57, 72-73 (2d Cir.1974). However, these cases left open the question whether a search could be justified as a search incident to arrest where no arrest was ever made. See Evans v. Solomon, 681 F.Supp.2d 233, 249 (E.D.N.Y.2010). Although the court in Evans held that the search was justified even if no arrest was ultimately made, I decline to go a step further and hold that a search can be justified as incident to arrest where the officer has no intention of arresting the individual at the time the search is conducted.
. No officers were ever identified in connection with this stop. I am satisfied that the defendant performed a thorough investigation to identify the officers who conducted this stop. See 4/30 Tr. at 5480 (Stipulation). Floyd was uncertain about the exact day that the stop occurred. See 3/18 Tr. at 196-197 (Floyd). However, I find Floyd's testimony generally credible and sufficiently detailed, and base my findings on that testimony.
. 3/18 Tr. at 160-161 (Floyd).
. See id. at 161-162.
. See id. at 164-166, 213.
. See id. at 166.
. See id. at 166-168. See also 3/19 Tr. at 223 ("Q: [Y]ou knew that you were free not to give the ID to the police officer, correct? A; I did not feel like it would have been smart for me not to have given it to them. So I did give it to them.”).
. See 3/18 Tr. at 170-172; 3/19 Tr. at 229, 253-254.
. See 3/18 Tr. at 172-173. The badge numbers that Floyd identified did not belong to officers named Rodriguez or Goodman. See id. at 204-205.
. See 3/19 Tr. at 232.
. See 3/18 Tr. at 174-177, 180.
. See id. at 177-178; 3/19 Tr. at 237 (Floyd).
. See 3/19 Tr. at 239-241 (Floyd); 3/28 Tr. at 1312 (Joyce).
. See 3/28 Tr. at 1322 (Joyce) (testifying that the officers would not have permitted the men to leave).
. See 3/18 Tr. at 179-180 (Floyd). Joyce admitted frisking Floyd but disputes that he put his hands in Floyd's pockets. See 3/28 Tr. at 1329, 1364 (Joyce).
. See 3/19 Tr. at 244 (Floyd).
. See id. at 247.
. See id. at 244-246.
. See 3/18 Tr. at 181; 3/19 Tr. at 245.
. See 3/28 Tr. at 1360, 1362. Accord id. at 1404-1406 (Hernandez) (testifying that many robberies and burglaries happened near the Cross-Bronx and that there was a specific burglary pattern in private homes in the northern part of the 43rd Precinct that day). See also DX K13 (mapping out burglary pattern).
. See 3/28 Tr. at 1411 (Hernandez); DX L4. Officer Hernandez could not confirm whether he saw the specific burglary pattern identified in DX L4, but testified that he was aware of patterns of burglaries and robberies in that area on the day he stopped Floyd.
. See 3/28 Tr. at 1319 (Joyce) (discussing UF-250 form).
. See id. at 1450 (Kelly).
. See 3/29 Tr. at 1480.
. See id. at 1363 (Joyce). See also id. at 1509 (Kelly).
. See 3/28 Tr. at 1327 (Joyce); DXX4 (UF-250 form). Officer Joyce also checked the box for "evasive, false or inconsistent re*652sponse to officer’s question,” which he explained was based on the out-of-state ID, which was inconsistent with Floyd living on Beach Avenue. See 3/28 Tr. at 1330-1331 (Joyce). This is irrelevant, however, because it relates to events that occurred after the stop and frisk had already taken place.
. See 3/28 Tr. at 1333-1334 (Joyce).
. See Terry, 392 U.S. at 28, 88 S.Ct. 1868 (frisk justified where ”[t]he [suspects'] actions ... were consistent with ... a daylight robbery — which, it is reasonable to assume, would be likely to involve the use of weapons”).
. See 4/1 Tr. at 1738-1739. Lino's jacket was red leather, said ‘‘Pelle Pelle” on the front and had no white stripes or numbers on it. Lino testified that it is a popular brand.
. See id. at 1739 (Lino). On February 24, Officers Leek and Figueroa were conducting a directed patrol in the subway station at 103rd Street and Lexington Avenue because it is a high pedestrian traffic, high crime area. See 4/8 Tr. at 2695, 2725 (Leek).
. See 4/1 Tr. at 1739 (Lino).
. See id. at 1740.
. See id. at 1741-1742. See also 4/8 Tr. at 2737, 2773 (Leek).
. See 4/1 Tr. at 1742 (Lino).
. 4/8 Tr. at 2754 (Leek).
. Id. at 2759 (Figueroa). See also PX 216 (Officer Leek memo book) (stating that Lino gave permission for frisk when Leek asked him).
. 4/1 Tr. at 1743 (Lino). Accord 4/8 Tr. at 2760 (Figueroa) (testifying that Lino jerked his body around when he felt Officer Figueroa’s hands on his back pockets, at which point Officer Figueroa asked if there was a problem).
. See 4/1 Tr. at 1743 (Lino).
. See id. at 1745-1746; 4/8 Tr. at 2734 (Leek) (investigative stop was not completed when the officers took Lino upstairs); id. at 2774 (Figueroa) (same).
. See 4/8 Tr. at 2704-2706, 2735-2736 (Leek). Sergeant Shirvis never exited the van, and did not confirm whether Lino was the suspect.
. See 4/1 Tr. at 1746-1747 (Lino). Lino then pulled out his cell phone and Officer Leek said "Oh, no, you’re not taking no fucking pictures of me.” Id. at 1748. Lino asked Officer Figueroa for his information, and Officer Figueroa mumbled it. When Lino asked for his information again, Officer Figueroa said, "I already gave it to you, if you didn’t get it the first time too bad.” Id.
. See id. Accord 4/8 Tr. at 2706 (Leek).
. See PX 194. Lino received no additional follow up. See 4/1 Tr. at 1748-1749 (Lino). Officer Leek received a letter notifying him that the charges against him had been substantiated but received no reprimand or further training. See 4/8 Tr. at 2708 (Leek). Officer Figueroa was put on monitoring because it was the third CCRB complaint filed against him and the CCRB recommended that charges be brought. See id. at 2761 (Figueroa).
. See PX 187; 4/8 Tr. at 2698-2699, 2722 (Leek).
. See 4/8 Tr. at 2723 (Leek).
. See id. at 2726-2727. Accord id. at 2756 (Figueroa).
. See id. at 2731-2732 (Leek).
. 4 LaFave § 9.5(h) (quotations and citations omitted). See also U.S. v. Maneen, 410 F.3d 326, 330 (6th Cir.2005) (“The mere passage of time [ — eleven days — ], however, does not negate the lawfulness of the stop under the circumstances presented.”).
. See Terry, 392 U.S. at 27, 88 S.Ct. 1868 ("The officer need not be absolutely certain that the individual is armed; the issue is whether a reasonably prudent man in the circumstances would be warranted in the belief that his safety or that of others was in danger.”).
. Id. at 29, 88 S.Ct. 1868 (citations omitted).
. See 3/19 Tr. at 263 (D. Dennis).
. See id. at 278-279.
. See id. at 266.
. See id. at 268, 286.
. The officers were patrolling Seventh Avenue that night on assignment to enforce "quality of life” offenses. See 3/22 Tr. at 852 (Salmerón).
. See 4/27 Tr. at 1270 (Pichardo).
. See 3/19 Tr. at 269-272 (D. Dennis).
. See 3/22 Tr. at 853 (Salmerón). The officers also claim that when they asked what he was doing Dennis replied that he was drinking some Hennessy, that it was a Saturday night and his girlfriend’s birthday and that they were going out later. The officers claim that there was a bottle of Hennessy on the sidewalk next to Dennis. See id.
. See 3/19 Tr. at 272-274 (D. Dennis).
. See 3/22 Tr. at 854-855, 872 (Salmerón); 3/27 Tr. at 1272 (Pichardo) ("So instead of giving him a summons you arrested him, correct? A. Yes[.]”).
. See 3/19 Tr. at 274-275 (D. Dennis).
. See 3/27 Tr. at 1272-1273 (Pichardo); 3/19 Tr. at 275-276 (D. Dennis).
. See 3/28 Tr. at 1286 (Pichardo).
. See 3/19 Tr. at 276-277 (D. Dennis)
. Id. at 295.
. I do not, however, find that the officers had probable cause to arrest Dennis when they approached him.
. See People v. Mack, 49 A.D.3d 1291, 853 N.Y.S.2d 764, 765 (4th Dep’t 2008) (holding that the viewing of three men, one of whom was drinking from a large bottle wrapped in a brown paper bag, established reasonable suspicion of violation of the open container ordinance).
. See People v. Muhammad, 120 A.D.2d 937, 502 N.Y.S.2d 859, 860 (4th Dep't 1986) ("Even assuming, however, that defendant carried an open beer bottle or was drinking beer, the officer would not have been authorized to frisk or search defendant for a suspected violation of an ordinance.”) (citations omitted).
. See, e.g., People v. King, 102 A.D.2d 710, 476 N.Y.S.2d 847, 849 (1st Dep't 1984) (officer has authority to conduct a search for safety purposes in circumstances that reveal only the commission of a violation where it was reasonable for an arresting officer to believe that the suspect might resist the officers efforts to arrest him or issue a summons to him).
. Compare Mack, 853 N.Y.S.2d at 765 (holding, in the context of a stop for an open container violation, that, "[t]he disproportionately frightened reaction of defendant upon seeing the [officers], his refusal to remove his hand from his pocket despite the repeated demands of [the officer], his conduct in walking toward that officer with his hand in his pocket, and the fact that the area in which the incident occurred was one in which violent crimes and shootings were common provided the officers with reasonable suspicion [] that defendant posed a threat to their safety”).
. I find that the NYPD’a attempts to identify the officers involved in the stops of Ourlicht and Peart were sufficient. See 4/30 Tr. at 5471-5476 (Stipulation) (discussing Ourlicht stops); id. at 5477-5480 (discussing Peart *657stops). Therefore, no adverse inference against the defendant is warranted.
. See 3/19 Tr. at 327-336 (Peart).
. See id. at 337-342.
. See 4/19 Tr. at 4193-4203 (Ourlicht).
. See id. 4204-4209, 4265.
. See 4/1 Tr. at 1693-1694 (Acevedo).
. See id. at 1695-1696, 1698.
. See 4/8 Tr. at 2659, 2664, 2684 (DeMarco).
. See 4/29 Tr. at 5198 (Vizcarrondo).
. See id. at 5199.
. 4/1 Tr. at 1697-1698 (Acevedo); 4/30 Tr. at 5458 (Hawkins). All three detectives were in plainclothes. See 4/1 Tr. at 1698-1699 (Acevedo).
. See 4/8 Tr. at 2685 (DeMarco); 4/29 Tr. at 5199 (Vizcarrondo); 4/30 Tr. at 5456 (Hawkins).
. See, e.g., 4/8 Tr. at 2659 (DeMarco); 4/30 Tr. at 5468 (Hawkins) (explaining that she did not record the stop in her memo book because she did not believe that the interaction ever rose to the level of a stop — rather, she perceived it to be a simple request for information).
. See 4/1 Tr. at 1704-1705 (Acevedo).
. See PX 5.
. In fact, Acevedo did not believe the officers were police at all, so she cannot have been stopped based on their show of authority, and the officers did not use force.
. I also do not find that race played any role in this event. I note, moreover, that this stop is anomalous insofar as the officers do not assert that they possessed reasonable suspicion of any crime and never intended to conduct a stop at all. Thus, it is of limited value in assessing the central claims at issue in this case.
. See 4/1 Tr. at 1749-1750 (Lino).
. Lino testified that Officer Hassan received a phone call during the stop, and suggested that his ringtone, which was a rap song, would calm Lino down. See 4/1 Tr. at 1749-1750 (Lino). Officers Hassan and Colon deny that this ever occurred. See 4/18 Tr. at 4018 (Hassan); see also id. at 4024-4025 (Colon).
. These policies and practices are summarized supra Part IV.C.
. See Def. Mem. at 2-4.
. See Def. Findings ¶¶ 50-52.
. See DX L14.
. Def. Mem. at 11. Accord Def. Findings ¶ 60 ("The recordings made by P.O.s Polanco and Serrano support” the conclusion that “[fjrom the Chief of Patrol down to officers straight out of the Police Academy, the message is clear: address conditions, not 'numbers for numbers^] sake.’ ”).
. See Def. Mem. at 7.
. See id. at 12 (presenting plaintiffs' criticisms of the NYPD’s training, supervision, monitoring, and discipline as baseless); id. at 24 ("Plaintiffs failed to prove that the NYPD systems already in place ... might require more than minor adjustments....").
. See id. at 24-25.
. Id. at 8.
. Cash, 654 F.3d at 334 (quoting Reynolds, 506 F.3d at 192).
. Connick, 131 S.Ct. at 1359.
. I was not persuaded, however, by plaintiffs' argument that "stopping people at random would more frequently uncover criminal activity,” based on the 9% arrest rate for the suspicionless roadblocks in Edmond, 531 U.S. at 32, 121 S.Ct. 447. PL Mem. at 2. There may be a higher likelihood of finding contraband in a randomly stopped vehicle than on the person of a randomly stopped pedestrian. In addition, the police in Edmond, unlike the NYPD in this case, used narcotics-detection dogs. See Edmond, 531 U.S. at 35, 121 S.Ct. 447.
. Wolf v. Colorado, 338 U.S. 25, 27, 69 S.Ct. 1359, 93 L.Ed. 1782 (1949).
. Whren, 517 U.S. at 813, 116 S.Ct. 1769.
. United States v. Davis, 11 Fed.Appx. 16, 18 (2d Cir.2001) (citing United States v. Fares, 978 F.2d 52, 59 (2d Cir.1992)). Accord, e.g., Chavez, 251 F.3d at 635 (noting that the use of "impermissible racial classifications in de*661termining whom to stop, detain, and search” violates the Equal Protection Clause).
. See Melendres v. Arpaio, No. 07 Civ. 02513, 2013 WL 2297173, at *67-68 (D.Ariz. May 24, 2013).
. Brown, 221 F.3d at 337 (citing Yick Wo, 118 U.S. at 373-74, 6 S.Ct. 1064). See also Whren, 517 U.S. at 813, 116 S.Ct. 1769 ("[T]he constitutional basis for objecting to intentionally discriminatory application of laws is the Equal Protection Clause.”).
. See Brown, 221 F.3d at 337.
. Chavez, 251 F.3d at 635-36. In some contexts, discriminatory effect may be presumed based on proof of discriminatory intent. See Doe v. Village of Mamaronedc, 462 F.Supp.2d 520, 543 (S.D.N.Y.2006) (“Once racially discriminatory intent infects the application of a neutral law or policy, the group that is singled out for discriminatory treatment is no longer similarly situated to any other in the eyes of the law, so adverse effects can be presumed.”). The terms “discriminatory purpose” and "discriminatory intent” are interchangeable.
. See supra Parts IV.A and IV.B. See also Chavez, 251 F.3d at 638 (noting that the Supreme Court has "repeatedly relied on statistics” to prove discriminatory effect, citing, for example, Hunter v. Underwood, 471 U.S. 222, 227, 105 S.Ct. 1916, 85 L.Ed.2d 222 (1985), and concluding that the Supreme Court in Armstrong rejected statistical evidence of discriminatory effect "not because plaintiffs can never use statistics to prove discriminatory effect, but because the particular statistics presented to the Court did not address the relevant issue,” id. at 639). As the discussion of benchmarking in Part IV.B makes clear, this Opinion assumes — unlike the presumption criticized in Armstrong — that there are racial disparities in crime participation rates.
. Evidence of a racial disparity in stop justifications included the following: (a) the odds of a stop resulting in any further enforcement action are lower if the person stopped is black than if the person stopped is white; (b) stops of blacks and Hispanics are less likely to result in the seizure of a weapon than stops of whites; (c) officers are more likely to check Furtive Movements as the basis for stopping blacks and Hispanics than for whites; and (d) the greater the black population in a precinct, the less likely that a stop will result in a sanction. See generally supra Part IV.A-B.
. Paterson, 594 F.3d at 163 (quoting Feeney, 442 U.S. at 279, 99 S.Ct. 2282 (citation and footnote omitted)) (some quotation marks omitted).
. Judge G. Murray Snow of the District of Arizona provides a useful summary:
A government policy is presumed to be racially discriminatory when it is "based in part on reports that referred to explicit racial characteristics.” Flores v. Pierce, 617 F.2d 1386, 1389 (9th Cir.1980) (emphasis added) (Kennedy, J.). In Grutter [v. Bollinger], the Supreme Court applied strict scrutiny to a policy which involved race as one factor among many even though plaintiffs expert conceded that "race is not the predominant factor” in the policy. 539 U.S. [306, 320, 123 S.Ct. 2325, 156 L.Ed.2d 304 (2003) ]; see also Arlington Heights, 429 U.S. at 263, 97 S.Ct. 555 (subjecting government action to equal protection review on "proof that a discriminatory purpose has been a motivating factor in the decision”).
Melendres, 2013 WL 2297173, at *69 (holding that the Maricopa County Sheriffs Office violated the Equal Protection Clause through its use of Hispanic ancestry or race as a factor in forming reasonable suspicion of violating immigration laws, despite express policy against racial profiling).
. Ferrill v. Parker Grp., Inc., 168 F.3d 468, 473 & n. 7 (11th Cir.1999).
. The NYPD’s use of local crime suspect data to target racially defined groups for stops is not only a form of racial profiling, it is also a deeply flawed way of identifying the criminal population. See supra Part IV.B.3.b.
. Plaintiffs have presented other circumstantial evidence of discriminatory intent, such as the NYPD’s longstanding indifference to evidence of racial discrimination in stops, discussed at greater length below. See infra Part V.B.2 (deliberate indifference); PL Mem. at 20-23.
. "It is well established that when the government distributes burdens or benefits on the basis of individual racial classifications, that action is reviewed under strict scrutiny.” Parents Involved in Cmty. Sch. v. Seattle Sch. Dist. No. I, 551 U.S. 701, 720, 127 S.Ct. 2738, 168 L.Ed.2d 508 (2007). In order to satisfy strict scrutiny, a racial classification must be " 'narrowly tailored' to achieve a 'compelling' government interest.” Id. (quoting Adarand, 515 U.S. at 227, 115 S.Ct. 2097).
. See 4/10 Tr. at 3034 (Chief Esposito acknowledging that "the right people” are young black and Hispanic youths 14 to 20 for whom there is reasonable suspicion); id. at 3027-3029 (Chief Esposito acknowledging that stops are based on "who is doing th[e] shootings,” specifically "young men of color in their late teens, early 20s”). Again, I note that this targeting is based on the inaccurate assumption that the characteristics of the criminal population can be applied to the non-criminal population. The policy assumes that all members of a racially defined group are “the right people” to target for stops because some members of that group committed crimes.
. To be sure, the policy’s use of racial classifications only becomes express at the managerial level, when commanders and supervisors acting in accordance with the policy instruct officers to target racially defined groups for stops. I note, however, that four judges of the Second Circuit, including now-Associate Supreme Court Justice Sonia Soto-mayor, concluded that even the use of race in Brown — where the police attempted to stop every black person in Oneonta as part of a search for a suspect described as a young black male with a cut on his hand— constituted an express racial classification requiring strict scrutiny. See Brown v. City of Oneonta, New York, 235 F.3d 769, 779-83 (2d Cir.2000) (Calabresi, J., dissenting from denial of rehearing in banc). Accord In re Cincinnati Policing, 209 F.R.D. 395, 401 (S.D.Ohio 2002) (noting that "[i]t is difficult to comprehend how the use of race in routine policing — for example, in the description of a suspect when the police are not in hot pursuit — could satisfy” strict scrutiny).
. Brown, 221 F.3d at 338. The Second Circuit emphasized the narrowness of its holding:
While we ... believe that the conduct of the police in the circumstances presented here did not constitute a violation of the equal protection rights of the plaintiffs, we do not establish any rule that would govern circumstances giving rise to liability that are not present in this case. Any such rule will have to wait for the appropriate case.
*664Id. at 339. As already noted, the Second Circuit has subsequently reiterated that it is concerned about allegations of racial profiling. See Davis, 11 Fed.Appx. at 18.
. Brown, 221 F.3d at 337.
. The Second Circuit specifically distinguished the facts of Brown from the type of policy at issue here, stating: "Plaintiffs do not allege that ... the police used an established profile of violent criminals to determine that the suspect must have been black. Nor do they allege that the defendant law enforcement agencies have a regular policy based upon racial stereotypes that all black Oneonta residents be questioned whenever a violent crime is reported.” Id. Indeed, Brown may be strictly limited to its facts — the use of race in the search for a particular suspect in a particular crime who was identified in part by his race.
. Finally, the City argues, erroneously, that "plaintiffs failed to prove individual equal protection claims against known or unknown NYPD officers by proving discriminatory purpose against and discriminatory effect upon at least one of the named plaintiffs.” Def. Mem. at 18 (emphasis added). Such proof is unnecessary. Even if named plaintiffs are unable to *665prove their claims at trial, the claims of an unnamed member of the class is sufficient for the purposes of Article III standing and may serve as a basis for liability. See Sosna v. Iowa, 419 U.S. 393, 402, 95 S.Ct. 553, 42 L.Ed.2d 532 (1975) ("The controversy [required by Article III] may exist ... between a named defendant and a member of the class represented by the named plaintiff, even though the claim of the named plaintiff has become moot.’’); Whitlock v. Johnson, 153 F.3d 380, 384 & n. 1 (7th Cir.1998) (noting that “[t]he reasoning of Sosna ... applies with equal force” whether named plaintiffs’ claims are terminated based on mootness or a failure on the merits); Melendres, 2013 WL 2297173, at *59 ("[W]hen the claims of named plaintiffs are not proven at trial, unnamed class members may be awarded relief so long as a 'controversy' still exists between the unnamed class members and the defendants.”). As discussed above, Officer French treated McDonald, an unnamed class member, differently than the similarly situated whites or Asians leaving the bowling alley nearby, and Officer French stopped McDonald in part because of his race, as a result of the NYPD’s policy of indirect racial profiling. See supra Part IV.C.3 (McDonald’s stop as illustration of indirect racial profiling policy).
. See Cash, 654 F.3d at 334. See also DiStiso v. Cook, 691 F.3d 226, 241 (2d Cir.2012) (noting that an equal protection claim against school officials can be based on their deliberate indifference to student-on-student racial harassment, even in the absence of direct proof of the officials’ discriminatory intent, provided that the evidence of their "clearly unreasonable” response to "actually known” circumstances "give[s] rise to a reasonable inference that [the officials] intended for the harassment to occur”).
. Although the following sources lie outside the trial record and therefore play no role in my decision, I observe for the public record that notice of racial bias in NYPD stops has continued since the close of discovery. For example, the Human Rights Committee of the United Nations recently asked the United States to "provide information on steps taken to address discriminatory and unlawful use of ‘stop and frisk' practices by officers of the New York Police Department,” as part of the Committee's preparation of the fourth periodic report on U.S. compliance with the International Covenant on Civil and Political Rights, ratified by the United States in 1992. UN Human Rights Committee, List of issues in relation to the fourth periodic report of the United States of America (CCPR/C/USA/4 and Corr. 1), adopted by the Committee at its 107th session (11-28 March 2013) (4/29/13). No other police department is mentioned. See id.
In addition, as has been widely reported, a teenager named Alvin Cruz has made public an audio recording of his stop in Harlem on June 3, 2011. It appears to be the only known recording of a stop by a civilian. In the recording, the officers verbally abuse Cruz, threaten to break his arm, and appear to physically abuse him. After an officer asks Cruz if he wants to go to jail, Cruz asks why the officers are threatening to arrest him, and one replies: "For being a fucking mutt! You know that?” Ross Tuttle & Rein Schneider, Stopped-and-Frisked: ‘For Being a F* *king Mutt’ [VIDEO], The Nation, Oct. 8, 2012; *6661/11/13 Letter from Darius Charney, Counsel for Plaintiffs, to the Court ("1/11/13 Charney Letter”) at 2. An amicus brief from a caucus of the New York City Council states that during three heavily attended public hearings in October 2012, several constituents who had been stopped “compared their own experiences of needlessly aggressive policing” to the Cruz recording. 3/4/13 Brief of Amicus Curi[a]e the Black, Latino, and Asian Caucus of the Council of the City of New York at 4. Plaintiffs requested on January 11, 2013 that they be allowed to add Cruz as a trial witness, after first contacting Cruz in December 2012. See 1/11/13 Charney Letter at 2. I denied the request on the basis of prior scheduling orders and the imminent trial date. See 1/17 Tr. at 4-5.
. See supra Part IV.C; PI. Findings ¶¶ 186— 190 (collecting sources regarding lack of internal attention to racial profiling).
. See supra Parts IV.C.l, 4, 7. I note that the City’s highest elected official. Mayor Bloomberg, has also defended the racial disparities in the NYPD's stops by invoking the crime suspect data benchmark. See April 30, 2013 Bloomberg Address at 3. On that basis, he has argued that the police “ 'disproportionately stop whites too much and minorities too little.' ” Associated Press, Bloomberg: Police Stop Minorities “Too Little," June 28, 2013. Echoing Chief Esposito’s testimony, Mayor Bloomberg stated that " '[t]he cops' job is to stop (people in) the groups fitting the description. It’s society’s job to make sure that no one group is disproportionately represented as potential perpetrators.’ ” Id. (emphasis added).
Similarly, Commissioner Kelly has embraced the use of violent crime suspect data as a benchmark, and on this basis has argued that " ‘really, African-Americans are being under-stopped.’ ” Lawrence Downes, What Ray Kelly Said About Stop-and-Frisk, Taking Note: The Editorial Page Editor’s Blog, N.Y. Times, May 3, 2013 (quoting May 1, 2013 interview with ABC News). Of course, using violent crime suspect data as a benchmark for measuring racial bias in the NYPD’s stops is even less valid than using general crime suspect data as a benchmark. See supra Part IV.C.7 (discussing RAND Report's use of violent crime suspect data benchmark). I note that Commissioner Kelly has also promoted the use of a violent crime suspect benchmark for stops through yearly public reports released by the NYPD. See, e.g., Kelly, Crime and Enforcement Activity, at 15 (presenting, without explanation, racial composition of stop subjects alongside racial composition of violent crime suspects).
. See supra Part IV.C.7 (noting skepticism of NYPD officials toward reports of racial profiling, which they claim to have not personally heard).
. See supra Part IV.C.3 (targeting “the right people”). The City finds nothing problematic in Inspector McCormack's instruction to Officer Serrano to target "male blacks 14 to 20, 21” for stops based on local crime suspect data. See Def. Findings ¶ 65; PX 332T at 24.
. See, e.g., 4/9 Tr. at 2824 (Esposito); Def. Mem. at 19 ("Plaintiffs ... have failed to prove a widespread pattern of suspicionless stops, and consequently cannot prove a wide*667spread pattern of race-based stops.” (emphasis added)).
. Whren, 517 U.S. at 813, 116 S.Ct. 1769 (emphasis added).
. I note, however, that plaintiffs have offered evidence of a widespread pattern of suspicionless stops. See supra Part V.A.
. The direction not to stop "a 48-year-old lady [who] was walking through St. Mary’s Park when it was closed,” is just one example of instructions not to stop all individuals for whom a justification for a stop exists, but only to stop the right people. PX 332T at 21. While this particular instruction seems benign, to the extent that the NYPD focuses its resources on blacks and Hispanics to the exclusion of whites generally, the result is deeply troubling. White people also carry guns and contraband, but if the NYPD declines to stop them, they will go undetected and unrepresented in crime statistics.
In addition, applying law enforcement tactics unequally between various racial groups is a recipe for abuse. “[N]othing opens the door to arbitrary action so effectively as to allow ... officials to pick and choose only a few to whom they will apply [the law] and thus to escape the political retribution that might be visited upon them if larger numbers were affected.” Railway Express Agency, 336 U.S. at 112, 69 S.Ct. 463 (Jackson, J., concurring). I note one poll shows that 76% of black voters disapprove of stop and frisk. See Quinnipiac University, New Yorkers Back Ban on Take-Out Foam More Than 2-1, at 8 (Feb. 28, 2013).
. Charles Blow, The Whole System Failed Trayvon Martin, N.Y. Times, July 15, 2013.