United States v. Sean Sowards

TRAXLER, Chief Judge,

dissenting:

Today, we establish that a police officer cannot legally stop a speeding vehicle based only upon his visual estimate unless the vehicle is traveling in “significant excess” of the speed limit or the officer has the time and practical ability to confirm his belief that the vehicle is speeding through radar, pacing, or some other corroborating evidence. No longer will the officer’s professional judgment alone be adequate, and the prohibition applies regardless of the *598extent of the officer’s training, experience, or certified ability to accurately estimate vehicle speeds within a very narrow margin of error. Ironically, while a lay person can estimate the speed of such a vehicle based upon his or her personal observations, an experienced and trained police officer no longer can.

Deputy Elliott has more than eight years of experience in the daily enforcement of North Carolina’s traffic laws, and has three times demonstrated through North Carolina’s radar certification procedures an ability to accurately estimate the speed of moving vehicles within an average 3.5-mph margin of error. These facts are uncontradicted. Yet the majority holds that Deputy Elliott’s visual estimate that Sowards’s vehicle was traveling 5 mph over the posted speed limit is inherently unreliable and, without corroborating evidence, insufficient as a matter of law — not to sustain a conviction, but rather to provide probable cause to stop the vehicle. In doing so, the majority also effectively holds that North Carolina’s certification test required to demonstrate an officer’s ability to estimate the speed of vehicles, like similar programs employed by a number of states across our country, is invalid as a matter of law. Even though a motorist is speeding, knows he is speeding, and may well admit that he is speeding if stopped, an officer working alone and without radar cannot even pull the car over for a warning as long as the driver is reasonably believed to be only breaking the law slightly as opposed to significantly — a distinction this circuit has never made for unlawful behavior.

In adopting its inflexible corroboration requirement for “slight” speeding violations, I believe the majority has unnecessarily distorted the well-established “totality of the circumstances” test normally applicable to all probable-cause determinations, and has effectively required that an officer have evidence sufficient for a jury to convict beyond a reasonable doubt before he may stop the vehicle. And it does so based upon its belief, mistaken in my view, that speeding violations present some kind of “unique circumstance” requiring this heightened evidentiary burden. Because I cannot agree with this unwarranted limitation on probable-cause jurisprudence, I respectfully dissent.

I.

Deputy Elliott is a ten-year veteran in law enforcement in North Carolina, the last eight of which included as a regular part of his duties “the enforcement of traffic laws, including speeding.” J.A. 20. After working three years in the patrol division, he was selected for placement on the Governor’s Highway Safety Program (“GHSP”) highway interdiction team.1 Deputy Elliott’s qualifications to visually estimate the speed of vehicles, however, is not limited to his years of experience in speed enforcement. He also “received specialized training as it relates to the enforcement of the traffic laws in the state of North Carolina,” J.A. 20, including “specialized training in radar certification,” J.A. 21.

Deputy Elliott described in some detail the training and testing involved in North Carolina’s certification process. As a pre*599requisite, the certification candidate -first trains with a certified operator who “show[s] you how to work the radar” and “show[s] you how to estimate the speed[ ]” of vehicles. J.A. 23. Although there is “not really a technique [to] measuring speeds,” J.A. 24, it is nonetheless a skill developed by practice and experience. Indeed, in North Carolina, it is a critical one, as radar alone cannot support a conviction for speeding as a matter of law. See State v. Jenkins, 80 N.C.App. 491, 342 S.E.2d 550, 552 (1986) (“By the express provisions of [North Carolina’s] statute, ... the speed of a vehicle may not be proved by the results of radar measurement alone and ... such evidence may be used only to corroborate the opinion of a witness as to speed, which opinion is based upon actual observation.”); see also J.A. 78-79 (testimony of Deputy Elliott confirming that “you cannot run a radar unit at all unless you have a visual estimation” and that “[t]he radar unit is only there to corroborate what you already know”).

After training, candidates must pass a written test and a road-course test. To pass the road-course test, candidates observe twelve vehicles, “estimate their speed, and then corroborate [the] visual calculations with the use of [the] radar,” all under the supervision of a certified instructor. J.A. 25.2 The margin of error is a combined 42 mph, or an average of 3.5 mph per vehicle. However, the candidate will automatically fail if he varies more than 12 mph on any single vehicle. Deputy Elliott successfully passed the tests and received certification in May 1998, April 2000, and February 2004. His certification was current when he stopped Sowards.

On April 11, 2006, Deputy Elliott positioned his vehicle in the median of Interstate 77, pointed south so as to provide him with an unobstructed view of approaching northbound traffic and a tracking history of approximately 100 yards, or the length of a football field. Deputy Elliott had been working this specific stretch of 1-77 on a daily basis for more than 4 years. The posted speed limit was 70 mph.3

While so positioned, Deputy Elliott continuously observed Sowards’s vehicle as it approached and passed him. He estimated the vehicle’s speed to be 75 mph. After stopping the vehicle, Deputy Elliott advised Sowards that he had been stopped for traveling 75 mph and that he should be driving the speed limit to be safe. So-wards does not contest that he was speeding.

Sowards presented Deputy Elliott with an Ohio driver’s license and said that he had traveled from Ohio to Atlanta by bus to pick up the car and return to Ohio. He claimed that his girlfriend “Deanna” owned the vehicle, but that he “didn’t really know her last name.” J.A. 38. According to the registration, however, the vehicle belonged to “Retcha Daily” from Georgia. J.A. 36. Due to the discrepancies, Deputy Elliott contacted the Blue *600Lighting Operations Center (“BLOC”) to obtain additional information on the license and vehicle, and advised Sowards that he would issue a warning ticket for speeding if everything checked out. While waiting for the response, Deputy Elliott observed that Sowards was sweating profusely and had a pre-paid cellular phone, and Deputy Elliott decided to have his drug dog perform an open-air sniff of the exterior of the vehicle. The dog alerted at the trunk. During the ensuing search, the officers detected the smell of laundry detergent, which is often used to mask the scent of narcotics, and discovered a compartment containing 10 kilograms of cocaine.

Prior to pleading guilty to the resulting drug charges, Sowards filed a motion to suppress the evidence obtained during the search. Sowards argued that Deputy Elliott’s visual estimate of his speed was insufficient to establish probable cause to stop his vehicle and that the dog sniff occurred during a period of unlawful detention.

Deputy Elliott was the only witness who testified at the suppression hearing and his testimony is uncontradicted. Although lay witnesses may offer opinions as to speed estimates, the district court found that Deputy Elliott’s “considerable training in estimating speeds” and the “foundation [laid] to testify as a law enforcement officer trained in estimating speeds” also qualified him as an expert. J.A. 31. At the conclusion of the hearing, the court denied the motion to suppress. With regard to the question of whether there was probable cause to initially stop Sowards’s vehicle, the court held as follows:

[Deputy] Elliott had probable cause to believe a traffic violation had occurred based on speed. He’s trained to estimate speeds. His difficulty with measurements is immaterial to his estimate of speed as that did not depend on time or distance. And the certification that he received ... three times, depended on accuracy in estimating speeds. So he had a particularized and objective basis for suspecting that a traffic violation had occurred.

J.A. 121.4 I would affirm.

II.

A.

The Fourth Amendment guarantees the right of persons to be free from “unreasonable searches and seizures.” U.S. Const, amend. IV. A law enforcement officer’s decision to stop a motorist constitutes a seizure within the Fourth Amendment, see Whren v. United States, 517 U.S. 806, 809-10, 116 S.Ct. 1769, 135 L.Ed.2d 89 (1996), and will be reasonable so long as the officer has “probable cause to believe that a *601traffic violation has occurred,” id. at 810, 116 S.Ct. 1769.

Whether an officer has probable cause to believe that a traffic offense has occurred is determined by the “totality of the circumstances.” Maryland v. Pringle, 540 U.S. 366, 371, 124 S.Ct. 795, 157 L.Ed.2d 769 (2003). We examine all events leading up to the stop and decide “whether these historical facts, viewed from the standpoint of an objectively reasonable police officer, amount to probable cause.” Id. (internal quotation marks omitted). In doing so, we must also consider the officer’s practical experience and specialized training. Police may “draw on their own experience and specialized training to make inferences from and deductions about the cumulative information available to them that might well elude an untrained person.” United States v. Johnson, 599 F.3d 339, 343 (4th Cir.2010) (internal quotation marks omitted); see also United States v. Humphries, 372 F.3d 653, 657 (4th Cir.2004).

Probable cause to stop a vehicle based upon a suspected traffic violation exists when the facts and circumstances within the officer’s knowledge are sufficient to warrant a prudent person in believing that the suspect has committed a violation of a traffic law. See Michigan v. DeFillippo, 443 U.S. 31, 37, 99 S.Ct. 2627, 61 L.Ed.2d 343 (1979). “The substance of all the definitions of probable cause is a reasonable ground for belief of guilt,” but this means far “less than evidence which would justify condemnation or conviction,” Brinegar v. United States, 338 U.S. 160, 175, 69 S.Ct. 1302, 93 L.Ed. 1879 (1949) (internal quotation marks omitted), and even less than that required by the preponderance-of-the-evidence standard, see Illinois v. Gates, 462 U.S. 213, 235, 103 S.Ct. 2317, 76 L.Ed.2d 527 (1983); Humphries, 372 F.3d at 660. “[T]he probable-cause standard does not [even] require that the officer’s belief be more likely true than false.” Humphries, 372 F.3d at 660; see also United States v. Ortiz, 669 F.3d 439, 446 (4th Cir.2012) (“A ‘reasonable ground’ for belief [of guilt] is less demanding than a standard requiring a preponderance of the evidence for the belief.”); United States v. Jones, 31 F.3d 1304, 1313 (4th Cir.1994) (“The probable cause standard does not demand any showing that such a belief be correct or more likely true than false.” (internal quotation marks omitted)).5

B.

As the Tenth Circuit recently recognized, “[fit’s long been the case that an officer’s visual estimation can supply probable cause to support a traffic stop for speeding in appropriate circumstances,” and a radar reading or other such objective or mechanical corroboration is not re*602quired. United States v. Ludwig, 641 F.3d 1243, 1247 (10th Cir.2011) (affirming district court’s denial of motion to suppress illegal drugs found in search of vehicle stopped for traveling 10 mph over the speed limit); see also United States v. Pierce, 23 F.3d 404, 1994 WL 159767, at *2 (4th Cir. Apr. 28, 1994) (per curiam) (affirming denial of motion to suppress where officer visually estimated that vehicle was traveling 75 mph on interstate, but was blocked from obtaining a radar reading by a tractor-trailer); State v. Barnhill, 166 N.C.App. 228, 601 S.E.2d 215, 218 (2004) (rejecting trial court’s requirement that an officer’s visual estimate of speed must be corroborated by “objective facts” or “objective criteria” as contrary to North Carolina law, and noting that such a ruling “would have [had] the effect of preventing an officer from stopping a vehicle based solely upon the officer’s observations”).

In my opinion, the facts and circumstances known to Deputy Elliott, coupled with his practical experience, training, and the reasonable inferences drawn therefrom, were more than sufficient to warrant an objectively reasonable belief on his part that Sowards was speeding. Deputy Elliott is an experienced officer who has been engaged in the enforcement of traffic laws as a regular part of his duties for eight years. As a part of his training, certification, and recertification, he has demonstrated an uncontradicted ability to visually estimate the speed of vehicles within an average margin of error of 3.5 mph per vehicle. He had worked this particular stretch of 1-77 for over four years, at times in a position to corroborate his visual estimates with radar; had become familiar with the speed at which cars approached him; and had a clear, unobstructed view of Sowards’s vehicle as it approached and passed him.

In a number of places, the majority takes issue with my representation that North Carolina’s certification and training procedures for police officers encompass both the officer’s use of radar equipment and the officer’s ability to estimate vehicle speeds within narrow margins of error. The majority instead appears to believe that North Carolina’s procedures are somehow limited to the use of radar equipment, and goes so far as to state that the officer was merely “given the opportunity to ‘guess’ the speed of twelve vehicles” during his training. Majority Op. at 588. However, the majority’s view that the visual-estimate test is little more than an “opportunity to guess,” rather than a requirement to pass, is contradicted by the record. The officers are instructed how to operate the particular radar equipment, but there is no basis upon which we could conclude that the visual-estimate portion of the test is somehow a less-crucial component or, worse, some form of a guessing game. Deputy Elliott’s description of the training and the requirements to pass the course is fully consistent with what North Carolina law requires in order to obtain speeding convictions. An officer’s visual estimate of the speed of a vehicle is required to convict as a matter of law; radar, assuming it is able to be used and is not excluded for some reason, can only corroborate the officer’s observations. Thus, to the extent we should pass judgment upon North Carolina’s training and certification program, I am confident that North Carolina law enforcement would have the most interest in ensuring that their officers are trained in and develop the actual skill of accurately estimating the speed of vehicles in case radar equipment is unavailable, blocked by other vehicles or obstructions, or is otherwise excluded for lack of calibration or other deficiencies.

Under these uncontradicted facts, which in any event we must construe in the light *603most favorable to the government, see United States v. Seidman, 156 F.3d 542, 547 (4th Cir.1998), I believe the government has easily established that Deputy Elliott had an objectively reasonable belief that Sowards was speeding and, therefore, that he had probable cause to stop So-wards’s vehicle for the suspected violation.

III. The Corroboration Requirement

The majority opines that “the Fourth Amendment does not allow, and the case law does not support, blanket approval for the proposition that an officer’s visual speed estimate, in and of itself, will always suffice as a basis for an officer’s probable cause to initiate the traffic stop.” Majority Op. at 590 (emphasis added). I agree. Such an inflexible rule would ignore the totality-of-the-circumstances test and, in particular, the mandate that we evaluate reasonableness based upon all of the facts which led to the stop, including the officer’s training and experience. I disagree, however, with the majority’s equally inflexible rule that an officer’s visual speed estimate can never alone suffice as probable cause to stop a vehicle that the officer estimates to be traveling only in “slight excess” of the speed limit. Majority Op. at 591.

The majority’s holding is clear: “[T]he reasonableness of an officer’s visual speed estimate depends, in the first instance, on whether a vehicle’s speed is estimated to be in significant excess or slight excess of the legal speed limit.” Majority Óp. at 591 (emphasis added). Where the officer visually estimates that a vehicle is traveling “in significant excess of the legal speed limit,” a visual estimate may “provide sufficient ‘indicia of reliability’ to support an officer’s probable cause.” Majority Op. at 591 (emphasis added). But if the officer estimates that a vehicle is traveling only in “slight excess” of the speed limit, that estimate alone can never be enough; the “officer’s visual speed estimate requires additional indicia of reliability to support probable cause.” Majority Op. at 592 (emphasis added); see also Majority Op. at 591 (“If slight, then additional indicia of reliability are necessary to support the reasonableness of the officer’s visual estimate.” (emphasis added)).

A “slight excess” of the speed limit is defined only as a “speed differential difficult for the naked eye to discern,” Majority Op. at 592, or perhaps one that is otherwise believed to be beyond the abilities of humans to accurately determine, see Majority Op. at 594 (agreeing that “ ‘the accuracy of human estimation of speed cannot easily, readily and accurately discriminate between such small variations in speed’”) (quoting State v. Rimes, 234 S.W.3d 584, 588 (Mo.Ct.App.2007)). Accordingly, the corroboration rule fashioned by the majority applies regardless of the extent of any officer’s experience, specialized training, or demonstrated ability to accurately estimate vehicle speeds within a narrow margin of error. For the reasons set forth below, I do not think we should impose this threshold inquiry upon the normal probable-cause determination and hinge upon it an inflexible requirement of corroboration every time a police officer stops a vehicle for speeding in “slight excess” of the speed limit.

A.

At the outset, I find the threshold inquiry, and the basis for the significant/slight distinction used to trigger the corroborative-evidence requirement, to be unsupported by the record in this case and unworkable for police officers.

Courts routinely allowed lay witnesses, with no training or experience, to offer opinion testimony of their estimate of speed provided they have had a sufficient *604opportunity to observe the moving vehicle. And so long as that ability to observe is established, I have found no opinion restricting the reliability or admissibility of the opinion testimony based upon the time/distance mathematical formula for speed.6 Yet, under the majority’s decision, a police officer’s identical opinion will be deemed unreliable as a matter of law and, therefore, insufficient to establish probable cause to believe that a violation has occurred, unless the officer also estimates the speed differential to have been a “significant” one.7

Additionally, the majority provides no clear numerical or percentage division between driving in “slight excess” or “significant excess” of the speed limit. The majority explains that a “slight excess” is “a speed differential difficult for the naked eye to discern,” Majority Op. at 592, and ultimately declares that no human could accurately estimate a 5-mph variation of speed at 70-75 mph, see Majority Op. at 593-94 (citing Rimes, 234 S.W.3d at 589). However, there are no studies, expert testimony, or other evidence to support this conclusion or to enlighten law enforcement officers as to what speeding violation should be considered “slight” because it falls within that which is difficult for the naked eye to discern or otherwise beyond the capabilities of human estimation. Nor is there evidence to support the majority’s finding that uncorroborated, visual estimates of speeds in the “slight” category must be presumed to be unreliable for purposes of the probable-cause determination.

As support for its holding that the officer’s visual estimate of speed is inherently unreliable, the majority instead relies upon “common sense,” along with two cases involving anonymous tips, which we have held require additional corroboration or indicia of reliability. See United States v. Massenburg, 654 F.3d 480, 486 (4th Cir. *6052011); United States v. Reaves, 512 F.3d 123, 126 (4th Cir.2008). In this case, however, the evidence consists not of hearsay statements of anonymous witnesses regarding unlawful behavior, but rather the personal observation and resulting opinion of a trained and certified police officer, which the majority concludes is “inherently unreliable” based upon nothing more than its own view, unsupported by any evidence and contradicted by that which was presented, that such estimates are beyond the capabilities of any human being.

Deputy Elliott demonstrated an actual ability to visually estimate the speed of vehicles within an average, 3.5-mph margin of error, in accordance with the requirements for radar certification in North Carolina. And however remarkable one might believe this skill to be at first blush, it appears to be a common one among experienced traffic officers across the United States. See e.g., State v. Carter, No. 2CA-CR2008-0013, 2009 WL 1717812, at *3 (Ariz.Ct.App. June 18, 2009) (noting testimony that officer had completed a radar certification class which required officers to accurately estimate the speed of moving vehicles within 5 mph); State v. Estes, 148 Idaho 345, 223 P.3d 287, 288 (App.2009) (noting that the officer “had been trained in visually estimating the speed of vehicles and had received certification of the ability to make estimates within 5 miles per hour of the actual speed”); State v. McPartland, 36 A.3d 881, 887 n. 5 (Me.2012) (Jabar, J., dissenting) (noting officer’s testimony that, as part of her radar training course, “she was trained and.certified to make visual estimates of speed ‘within five miles per hour’ ”); State v. Ali, 679 N.W.2d 359, 368 (Minn.Ct.App. 2004) (noting that officer had been “trained ... to accurately estimate the speed of a moving vehicle within five mph”); Barberton v. Jenney, 126 Ohio St.3d 5, 929 N.E.2d 1047, 1049 (2010) (noting that, in order to be certified under Ohio’s requirements, the officer “was required to show that he could visually estimate a vehicle’s speed to within three to four miles per hour of the vehicle’s actual speed”); State v. Singh, No. F-98-022, 1999 WL 355270, at *1 (Ohio Ct.App.1999) (noting trooper’s testimony that he had been trained to visually estimate the speed of vehicles and was generally accurate within one or two miles per hour); Columbia County v. Kassens, 331 Wis.2d 729, 795 N.W.2d 492, 2011 WL 102598, at *1 (Wisc.Ct.App. Jan. 13, 2011) (noting officer’s testimony that he “ha[d] been trained to visually estimate a vehicle’s speed within three miles per hour”).8

In sum, I do not find the majority’s observations regarding the capabilities of law enforcement officers to be supported by the record. Moreover, I question how *606law enforcement officers, particularly those who have met their state certification requirements for visually estimating the speeds of vehicles within narrow margins of error, will know when they must forego stopping a speeding vehicle unless and until they observe the vehicle cross into the “significantly speeding” category or they are able to obtain other, corroborating evidence.

B.

Beyond these problems, I believe that the majority’s adoption of the corroboration requirement for slight speeding violations has no place in the probable-cause context. The origin of this new requirement is a handful of state court conviction cases which have held or implied that a police officer’s visual estimate of speed, standing alone, may not constitute sufficient evidence to prove a defendant guilty beyond a reasonable doubt if the variance between the visually estimated speed and the speed limit is determined to have been “slight” as opposed to “wide.” State v. Kimes, 234 S.W.3d 584, 588 (Mo.Ct.App. 2007) (internal quotation marks omitted); City of Kansas City v. Oxley, 579 S.W.2d 113, 116 (Mo.1979); see also State v. Estes, 148 Idaho 345, 223 P.3d 287, 289-91 (App. 2009); People v. Olsen, 22 N.Y.2d 230, 292 N.Y.S.2d 420, 239 N.E.2d 354, 355 (1968).9

To date, our circuit has not adopted a corroboration rule in conviction cases. See e.g., United States v. Dams, 164 F.3d 626, 1998 WL 726748, at *2 (4th Cir. Oct. 16, 1998) (per curiam) (noting that “the Government correctly points out that the officer’s visual estimate is also sufficient, by itself, to support a conviction”); see also United States v. Wornom, 754 F.Supp. 517, 519 (W.D.Va.1991) (affirming conviction based upon an officer’s visual estimate where radar evidence was suppressed). In my view, such matters should ordinarily be left for the jury or other factfinder to consider and weigh.10

*607But even if we had seen fit to require corroboration in conviction eases involving a slight speed differential, I do not think it prudent to import this reasoning into the probable-cause context and superimpose an inflexible corroboration requirement upon the totality-of-the-circumstances test. “The Supreme Court has repeatedly admonished that the standard for probable cause is not ‘finely, tuned’ or capable of ‘precise definition or quantification into percentages.’ ” Humphries, 372 F.3d at 660. Rather, the officer need only have an objectively reasonable belief that the defendant is speeding, and that belief need not even “be more likely true than false.” Id.

Thus, even in those states that have required more than just the officer’s visual observation to sustain a conviction for speeding where there is a “slight variance,” the courts have pointed out the important distinction between the evidence needed to establish probable cause and that needed to sustain a conviction. Of particular note, in State v. Ostdiek, 351 S.W.3d 758, 768-69 & 769 n. 10 (Mo.Ct. App.2011), the Missouri Court of Appeals, citing its earlier opinion in Kimes, recently reversed a defendant’s speeding conviction based solely upon the officer’s testimony that it “just appeared” that the vehicle was going faster than her vehicle and the others on the road. However, the court took care to point out that “[t]he reversal of the speeding conviction does not affect the legitimacy of the- initial traffic stop or any evidence which resulted from that stop,” the latter of which implicates a much different standard of review than the “beyond a reasonable doubt” determination necessary for a conviction. Id. at 769 n. 10; see also Estes, 223 P.3d at 289 n. 1 (noting that the issue of whether an officer’s visual estimate of a vehicle’s speed constitutes sufficient proof of speed beyond a reasonable doubt to sustain a conviction “should not be confused with the admissibility of an officer’s estimate of speed nor with the sufficiency of an estimate to provide reasonable suspicion to stop a vehicle, reasonable suspicion being a much less exacting standard than proof beyond a reasonable doubt” (emphasis omitted)). Thus, it appears that even the state courts in Missouri and Idaho would not import their corroboration requirement for “slight variance” speeding convictions into the probable-cause context.

C.

Finally, I turn to the two unpublished “probable cause” cases cited by the majori*608ty in support of its holding: United, States v. Moore, No. 10cr971(RJH), 2011 WL 6325973 (S.D.N.Y. Dec. 19, 2011), and State v. Petzoldt, 803 N.W.2d 128, 2011 WL 2556961 (Iowa Ct.App. June 29, 2011). Both involve a court’s determination that an officer’s visual speed estimate was insufficient to establish probable cause to stop a vehicle under the totality of the circumstances, but neither counsels our adoption of an absolute rule requiring corroborating evidence to establish probable cause in every case where the officer observes a slight violation of the legal speed limit.

In Moore, the two officers involved were anti-crime officers whose primary duties were “to respond to violent felonies rather than to enforce traffic laws.” Moore, 2011 WL 6325973, at *1. The first officer testified that the vehicle in question was “traveling in excess of the speed limit,” but he “did not describe how much experience he had had conducting traffic stops or estimating the speeds of traveling cars.” Id. at *2 (internal quotation marks omitted). The second officer likewise “did not describe his experience in enforcing speeding violations or in identifying the speed of vehicles by sight.” Id. at *3. Although noting that courts “will credit the observations of officers that a car was speeding when the officer has had special training [and experience] in detecting the speeds of vehicles,” id. at *5, the district court found no such foundation for the officer’s opinion in its case:

By contrast, here, there is no suggestion that the officers have received any such training. Moreover, nothing about the officers’ experience — nor their assigned duties on the night in question— suggests that they would have had the opportunity to become adept at estimating the speeds of vehicles. It seems likely that some police officers — such as state troopers who, as part of their regular duties, routinely estimate how fast particular cars are driving and test such an estimate by using a radar gun— might become very adept at judging a car’s speed. Here, however, the Court has no such reason for confidence, based on the officers’ training or experience ..., in the officer’s estimation — an “estimation,” the Court notes, that hardly qualifies as such, given the vagueness of the officers’ responses and their inability to give even a range of speeds at which the cab might have been traveling. The officers did not state that they had received specialized training in estimating the speed of moving vehicles, nor did they state that they had previously ... worked an assignment where their main responsibility was to apprehend violators of traffic laws. The officers both stated that on the night in issue, they were assigned to the Bronx Anti-Crime Unit, whose mission is not to enforce the Vehicle and Traffic Laws, but rather to focus on violent felonies. The officers were not even carrying books of summonses to issue in the event that they observed a traffic infraction.

Id. at *6. Thus, the Moore officers had no training or demonstrated ability to estimate the speed of vehicles, had no experience doing so, and offered no opinion as to the speed differential at all.

The majority’s reliance upon Petzoldt is similarly misplaced. In Petzoldt, the officer had “resorted to playing Solitaire on his computer to break the monotony of a very slow night” when “the stillness was broken [by] a pickup truck” passing by his location. Petzoldt, 2011 WL 2556961, at *1. “Believing the truck was speeding,” the officer pursued the vehicle. Id. As in Moore, the court held that “with proper foundation, an officer’s visual estimation of speed may be sufficient to supply probable cause to stop a vehicle for speeding.” Id. *609at *3. It did not, however, adopt a corroboration requirement based upon the degree of the estimated speed differential. Rather, it reversed the trial court’s denial of defendant’s motion to suppress because the officer did not have the requisite foundation for his opinion:

Here, Officer King testified he. was playing Solitaire when he observed Petzoldt’s pickup truck briefly as it passed in front of his patrol car. Although he testified he believed the truck was [speeding], Officer King made no estimate as to how fast the truck was travelling or how much over the posted limit he thought [it] was travelling. The posted speed limit is not even in the record before us. Officer King’s visual estimate of speed was not confirmed by any other means of corroboration of the speed, such as radar or pacing. Officer King observed no other traffic infractions or driving anomalies by the pickup.

Id. (footnotes omitted). Thus, the Petzoldt officer had no tracking history, had no training or demonstrated ability to visually estimate the speed of vehicles, and made no estimate of the speed differential.

Neither Moore nor Petzoldt support adoption of an absolute requirement for corroborating evidence in the probable-cause context where an experienced traffic officer has visually estimated the speed of a vehicle to be in only “slight excess” of the speed limit. The contrast between Deputy Elliott and the officers involved in Moore and Petzoldt could not be more stark. Deputy Elliott was trained and certified in radar enforcement, experienced in traffic enforcement, and had demonstrated through certification and testing procedures his adeptness at judging the speed of vehicles. His testimony was specific, both as to his estimate and as to the speed limit, and he had a continuous and unobstructed view of Sowards’s vehicle as it approached him. The facts of this case are simply not analogous.

D.

To summarize, neither the state-court conviction cases nor the unpublished probable-cause cases relied upon by the majority support its broad holding that an officer’s visual speed estimate can never suffice as a basis for an officer’s probable cause to initiate a traffic stop unless the suspect is estimated to be traveling in significant excess of the speed limit or the officer has the time and ability to corroborate his visual estimate through some other objective technique or circumstance. And the majority’s apparent basis for the rule — that “ ‘the accuracy of human estimation of speed cannot easily, readily, and accurately , discriminate between such small variations in speed,” Majority Op. at 594 (internal quotation marks omitted) — is not supported by the evidence in this case, which we must view in the light most favorable to the government.

In my opinion, an experienced officer such as Deputy Elliott, who has demonstrated an ability to estimate speed within a 3.5-mph margin of error, has all of the qualifications needed to form a reasonable belief that a speeding violation of 5 mph or more has occurred. Under today’s holding, however, such an experienced, trained and certified police officer cannot legally stop a vehicle that he legitimately and reasonably believes is traveling in “slight excess” of the speed limit and posing a potential threat to others based solely upon his visual estimate and professional opinion. I do not believe our probable-cause jurisprudence, which requires no more than that a law enforcement officer have such a reasonable belief, counsels or even permits that result.

*610IV. Factual Findings

A.

I turn now to the majority’s conclusion that the district court clearly erred in finding that Deputy Elliott “is trained to estimate speeds” and that “[h]is difficulty with measurements is immaterial to his estimate of speed as they did not depend on time or distance.” J.A. 121. As noted earlier, these factual findings, whether “clearly erroneous” or not, do not serve as the basis for the majority’s decision. No explanation is ever given by the majority to show how the rejection of the district court’s factual findings factors into the ultimate holding in the case. See Majority Op. at 593-94. But no explanation probably needs to be given, as the corroboration requirement alone is all that the majority needs to reverse the district court as a matter of law. Although I strive to find otherwise, the majority’s holding seems clear to me: as a matter of law, a police officer can never premise probable cause solely on his or her visual estimate of speed if the speed differential is “slight” as opposed to “significant” — no matter the officer’s training, knowledge, experience, certification, or demonstrated ability to estimate speeds or recite measurements. And this is because the majority “agree[s] that ‘the accuracy of human estimation of speed cannot easily, readily, and accurately discriminate between such small variations in speed.’ ” Majority Op. at 594 (quoting Rimes, 234 S.W.3d at 589). Thus, it matters not at all to the result whether Deputy Elliott was “trained” to visually estimate speed, as opposed to having developed and demonstrated the skill to do so during his training and certification, nor is there any indication that the result in this case would have been any different if Deputy Elliott had correctly recited the lengths of rulers and yardsticks.

In sum, Sowards’s difficulties with measurements — no matter how silly they seem or what fodder they would have made for cross-examination at trial — are irrelevant to the rule adopted today, and the district court’s factual findings, whether clearly erroneous or not, are not relevant to the legal determination that troubles me most.

B.

Had the majority relied upon Deputy Elliott’s lack of sufficient training or his inability to accurately recite measurements as a basis for determining that his visual estimate was unreliable, and therefore in need of corroboration, my dissent would remain. But it would rest on a much narrower basis: I disagree with the majority’s conclusion that the district court clearly erred in finding that Deputy Elliott was “trained to estimate speeds” and that his “difficulty with measurements [was] immaterial to his estimate of speed [because] they did not depend on time or distance.” J.A. 121.

Under the clear-error standard, “[a] factual finding by the district court may be reversed only if, although there is evidence to support it, the reviewing court on the entire evidence is left with the definite and firm conviction that a mistake has been committed.” Walton v. Johnson, 440 F.3d 160, 173-74 (4th Cir.2006) (en banc) (internal quotation marks omitted). “Where there are two permissible views of the evidence, the factfinder’s choice between them cannot be clearly erroneous.” Walker v. Kelly, 589 F.3d 127, 141 (4th Cir. 2009) (internal quotation marks omitted).

First, Deputy Elliott’s uncontradicted testimony is that he was required, as a part of his preparation for radar certification in North Carolina, to work with a certified instructor who “showfed] [him] how to estimate the speeds” of vehicles. *611J.A. 23. That sounds like training to me. I do not know what else it could mean. At the conclusion of the course, Deputy Elliott was also required to demonstrate to his instructor the ability to accurately estimate speeds within an average 3.5-mph margin of error. In my view, the district court’s choice of words — that Deputy Elliott was “trained to estimate speeds” — is accurate.

Second, Deputy Elliott inexplicably became confused when he was questioned about the measurements of feet and yards. He may well not know the correct answers to those questions, or he may just have had difficulty quickly recalling or converting such measurements on the witness stand. Unlike VASCAR, however, which involves a distance determination and timing mechanism to approximate speed according to the time/distance formula, there is no evidence that the reliability of an officer’s visual estimation of speed is or should be tied to a specific or minimum distance or time. See McBee v. State, 296 Ga.App. 42, 673 S.E.2d 569, 571 (2009) (rejecting defendant’s argument that his motion to suppress should have been granted based upon the officer’s “failure to use mathematical calculations or radar to estimate [the vehicle’s] speed” because “an officer’s visual estimate may be used to establish speed”); Barberton, 929 N.E.2d at 1051 (“Visual observation has long been held a valid means of determining the speed of a moving vehicle as long as the witness has a reliable opportunity to view the vehicle.” (internal quotation marks omitted)). There was likewise no evidence that North Carolina’s certification procedure for estimating speeds is dependent upon distance or time.

In my view, conditioning the materiality of a visual estimate of speed on the existence of a mathematical calculation ignores the realities of traffic enforcement and unduly ties the hands of officers who must have the freedom to exercise their judgment. Indeed, as a practical matter, it seems likely that visual estimates would perhaps most often come into play where police officers observe a speeding vehicle while engaged in their routine patrol duties and not while positioned where they can check speeds utilizing a time/distance mathematical calculation.11

*612Finally, and to the extent the majority’s determination on this factual finding or its holding at all rests upon Deputy Elliott’s personal competency, I note that whatever difficulties Deputy Elliott might have had in recalling the measurements of feet and yards while being questioned in a courtroom, these difficulties have not prevented him from passing the certification tests nor prevented him from accurately determining the speed of moving vehicles while on the street and engaged in his normal law enforcement duties.

The uncontradicted evidence in this case demonstrates that Deputy Elliott’s ability to estimate vehicle speeds was based upon his experience, training, and opportunity to observe the vehicle as it approached and passed him, rather than upon a known distance or minimum time. In my view, the evidence was more than sufficient to provide a foundation for his opinion in this case and to support his objectively reasonable belief that Sowards was speeding. For the same reasons, I am not left with a definite and firm conviction that the district judge, who had the opportunity to observe Deputy Elliott, made a mistake in finding that Deputy Elliott’s testimony regarding small measurements was immaterial to his estimate of the speed of Sowards’s vehicle, or in concluding that Deputy Elliott had probable cause to stop him.

V.

In my opinion, the majority’s decision today strikes a blow to the professional judgment of police officers, substitutes our opinion regarding the ability of officers to accurately assess the speed of vehicles for the facts presented, and severely ties the hands of trained and experienced police officers to enforce traffic safety laws. Whether, and to what extent, we might require corroborating evidence for slight speed differentials for purposes of sustaining a conviction should remain for another day. Whether visual estimates of speed within a slight speed differential are sufficiently reliable to prove that a vehicle was speeding “by a preponderance of the evidence” in the civil context is likewise not before us today. For purposes of the probable-cause determination, the officer need only have a “reasonable ground for belief of guilt.” Brinegar, 338 U.S. at 175, 69 S.Ct. 1302 (internal quotation marks omitted), which need not even “be more likely true than false,” Humphries, 372 F.3d at 660.

Additionally, the majority completely invalidates the road test North Carolina has employed for its traffic officers to demonstrate their ability to estimate the speed of cars. From pages 594 to 96 of the majority opinion, my colleagues give their reasons for finding the test inadequate to demonstrate an officer’s expertise to judge speed without the use of radar. Although my colleagues summarily deny that they have taken this step, they nonetheless set forth in some detail the perceived “shortcomings” of North Carolina’s test for the purpose of discrediting Deputy Elliott’s demonstrated abilities in this area and rendering his opinion, and all others like it, unreliable as a matter of law. Majority Op. at 595-96. Consequently, it seems clear to me that trained and certified police officers can no longer stop a speeding vehicle based only on their visual speed estimates unless satisfied that the vehicle falls within the as yet undefined “significant speeding” category. Given that it is only such “significant” speed differentials that do not require corroboration or other indicia of reliability under the majority’s holding, I can only view this as a wholesale rejection of North Carolina’s testing and visual-estimate certification procedures as they would apply to “slight” speeding violations. Yet the certification procedure was never challenged below or on appeal, *613and there is no evidence to support its invalidation. There is no telling what the ramifications will be in North Carolina and the other states of our circuit now that this certification procedure has been found insufficient.12

Based upon his substantial training, experience, multiple certifications, and personal observations, Deputy Elliott had an objectively reasonable belief that Sowards was speeding, and thus had probable cause to stop him. Accordingly, I would affirm the district court’s denial of his motion to suppress and respectfully dissent from the majority’s decision.13

. The Governor’s Highway Safety Program stems from a state-wide grant designed to facilitate enforcement of North Carolina’s traffic laws on the interstate systems and reduce the number of traffic crashes and fatalities. In the course of traffic enforcement, the officers also enforce narcotics and money laundering laws. See J.A. 20; North Carolina Department of Transportation, Governor's Highway Safety Program, http://www.ncdot. org/programs/ghsp (last visited Mar. 31, 2012).

. Deputy Elliott testified that the instructor "cover[s] up the ... radar unit, the target speed. And then at that point in time you’re formulating your opinion, you’re making a visual estimation. Once you make that visual estimation, then you are allowed to run the radar. At that point in time you make a clock with that radar. And once you confirm that that’s what ... has happened, that's when the certified instructor takes ... the piece of paper off your radar unit and advises you what the actual clock was.” J.A. 25.

. To obtain a reliable reading, the radar unit can be at no more than a 20-degree angle. Deputy Elliott parked his car at a 25- to 30-degree angle because on a previous occasion his car had almost been hit by a tire that had come off a passing trailer. After that incident, Deputy Elliott positioned his car farther from the road and at the broader angle.

. During the hearing, Deputy Elliott had difficulty answering questions regarding the measurements in feet and yards. The majority concludes that the district court clearly erred in finding that this difficulty with small measurements was immaterial to Deputy Elliott’s ability to estimate the speed of So-wards’s vehicle, and in finding that Deputy Elliott was trained to estimate speeds. For reasons discussed infra, I disagree. At the outset, however, I am compelled to point out that the majority does not ultimately reverse the district court based upon any perceived lack of confidence in Deputy Elliott's personal ability to offer a reliable opinion as to the speed of Sowards’s vehicle. Rather, the majority reverses because Deputy Elliott’s visual estimate fell within the slight-excess-of-the-speed-limit category and was not corroborated by other evidence. At no point does the majority explicitly indicate that the corroboration requirement was triggered by a lack of confidence in Deputy Elliott personally, nor does it expressly reverse the district court's denial of the motion to suppress based upon the district court's findings of fact. See Majority Op. at 593-94.

. In its response to the dissent, the majority seizes upon my citation of this latter portion of our established precedent comparing the probable-cause standard to the preponderance standard to charge that I advance a new "analytical approach” to probable-cause jurisprudence. Of course, this is no analytical approach at all; it is a well-established holding designed to ensure that police officers and courts do not conflate the probable-cause standard with those requiring much more in the way of evidence — specifically, the beyond-a-reasonable-doubt standard and the preponderance-of-the-evidence standard. To the extent I emphasize this portion of our precedent, it is because it is particularly apt in the case at hand, in which the majority imports into our probable-cause inquiry a corroboration requirement that originated in a handful of state court speeding conviction cases, discussed infra, based upon its view that "a speeding violation presents a unique circumstance” necessitating evidence for probable cause equivalent to that sufficient to sustain a conviction. Majority Op. at 590 n. 7. As I have clearly set forth above, probable cause does mean something, but, even in traffic-stop cases, it surely does not mean "beyond a reasonable doubt” or by "a preponderance of the evidence.”

. See State v. Barnhill, 166 N.C.App. 228, 601 S.E.2d 215, 217, 218 (2004) (noting "well established [rule], that any person of ordinary intelligence, who had a reasonable opportunity to observe a vehicle in motion and judge its speed may testify as to his estimation of the speed of that vehicle,” and that "it is not necessary that an officer have specialized training to be able to visually estimate the speed of a vehicle”); Walker v. State, 163 Ga.App. 638, 295 S.E.2d 574, 575 (1982) (allowing lay witness’s testimony as to speed of a vehicle to support vehicular homicide conviction); see also Asplundh Mfg. Div. v. Benton Harbor Eng’g, 57 F.3d 1190, 1197 (3d Cir. 1995) (noting that one example "of quintessential Rule 701 opinion testimony [is] ... the speed of a vehicle (footnote omitted)); Fed. R.Evid. 701 advisory committee's note (citing Asplundh for examples of opinion testimony by lay witnesses); United States v. Conn, 297 F.3d 548, 554 n. 2 (7th Cir.2002) (same); State v. McLean, 205 N.J. 438, 16 A.3d 332, 343 (2011) ("Traditional examples of permissible lay opinions include the speed at which a vehicle was traveling.”); Pierson v. Frederickson, 102 N.J.Super. 156, 245 A.2d 524, 527 (App.Div.1968) ("It is clear that based on adequate visual observation an ordinary witness can state his conclusion of whether a car was moving fast or slow or give an estimate of its speed.”); State v. Clayton, 272 N.C. 377, 158 S.E.2d 557, 561 (1968) ("Absolute accuracy ... is not required to make a witness competent to testify as to speed.”).

. Although the majority denies that it has created a heightened evidentiary burden for probable-cause in speeding cases, it sees no irony in the fact that there will now be two different standards governing the use of opinion testimony as to the speed of a vehicle— one for lay witnesses and another for law enforcement officers. My colleagues justify this difference by pointing out that a police officer uses his opinion as probable cause to stop and detain motorists. To this I would simply point out that the opinion of a lay person as to speed, admissible to support a criminal conviction, should surely be sufficient to establish probable cause. Cf. Barn-hill, 601 S.E.2d at 218 ("[I]f an ordinary citizen can estimate the speed of a vehicle, so can Officer Malone.”)

. The majority agrees that no evidence exists to support its view of the capabilities of human beings, but asserts that an evidentiary foundation is unnecessary because it is "common sense” that no human can accurately discriminate between such small variations of speed. In my view, the majority's conclusion concerning the limits of the abilities of trained police officers falls well outside the realm of "common sense” and clearly within the category of a factual finding necessitating evidentiary support. Further, I do not see how we can make such a factual finding in the absence of supporting evidence. Regardless of one’s view as to when logic and common sense may be relied upon in the absence of actual evidence of a particular fact, there is no absence of evidence in this case. Instead, the only evidence in the record contradicts the majority's finding. The evidence of So-wards's actual, demonstrated ability to accurately discriminate between speeds within a 3.5-mph margin of error during three separate tests was unchallenged by Sowards below, and cases from across the country show that many police officers are similarly trained in visually estimating speeds within such narrow margins of error.

. In Kimes, the Missouri Court of Appeals affirmed a speeding conviction based upon a 15-mph differential (a 75% variance) because it was deemed "not slight.” In doing so, it was compelled to distinguish the earlier Missouri Supreme Court opinion in Oxley, which had reversed a speeding conviction based upon a 10-mph differential (a 29% variance). By way of example, the Kimes court made a number of unsupported assumptions:

Where an officer's estimation of speed is 60 m.p.h., a fact-finder cannot conclude with any degree of certainty that a defendant was exceeding a 55 m.p.h. speed limit, because the accuracy of human estimation of speed cannot easily, readily, and accurately discriminate between such small variations in speed. Yet the same fact-finder, based upon that same 60 m.p.h. estimation of speed, could conclude beyond a reasonable doubt that a defendant was exceeding a 20 m.p.h. limit. This is so because the variance between the estimated speed and the speed limit falls within the margin of etror of accuracy within which an experienced person can discriminate between the two speeds.

Kimes, 234 S.W.3d at 589 (emphasis added). In support of Kimes’ blanket statement regarding human capabilities (with which the majority here agrees), Kimes also cites no studies, expert testimony, or other evidence. However, there is also no indication that the officers in Oxley or Kimes had developed or demonstrated any specialized ability to estimate the speed of vehicles. In my view, the finding regarding human abilities adopted by the majority is directly contradicted by the only evidence that was presented on the point in this case, as well as by observations to the contrary made by many other courts throughout the country. It appears that police officers work and train to develop this ability, and become very adept at estimating speeds within quite narrow margins of error. See also City of Rockford v. Custer, 404 Ill.App.3d 197, 344 Ill.Dec. 244, 936 N.E.2d 773, 776-77 (2010) (discussing this line of speeding-conviction cases but indicating that a conviction might still be affirmed where the officer gives a visual estimate of the defendant’s actual speed that falls within the appropriate margin of error).

. Other state courts have also expressed similar views in the conviction context. See e.g., Ferguson v. State, 263 Ga.App. 40, 587 S.E.2d 195, 196 (2003) (holding that an officer’s visual estimation of a vehicle’s speed is sufficient to support a conviction for speeding); Jackson v. State, 223 Ga.App. 27, 477 S.E.2d 28, 29 (1996) (affirming conviction for speeding based solely upon the officer’s visual speed estimate, and noting that “opinion testimony of an eyewitness may be used to establish speed, its credibility being for the jury to determine” (citation and alterations omitted)); State v. Ali, 679 N.W.2d 359, 368 (Minn.Ct.App.2004) (holding that trained officer’s visual estimate that defendant was traveling 11 mph above the speed limit was sufficient by itself to support a. conviction for speeding); Barberton v. Jenney, 126 Ohio St.3d 5, 929 N.E.2d 1047, 1049, 1051 (2010) (rejecting "a bright line rule that an officer’s visual estimation of speed, without other evidence to support it, is insufficient to sustain a conviction for speeding” and ”hold[ing] that a police officer’s unaided visual estimation of a vehicle's speed is sufficient evidence to support a conviction for speeding ... without independent verification of the vehicle's speed if the officer is trained [and] certified by [an] organization that develops and implements training programs to meet the needs of law-enforcement professionals and the communities they serve, and is experienced in visually estimating vehicle speed”).

. The majority additionally charges that the district court’s finding "rings in the absurd,” because no one can “discern the speed of a vehicle ... without discerning both the increment of distance travelled and the increment of time passed.” Majority Op. at 589. To answer the majority's question, I do not quibble with the mathematical formula for speed, and express no opinion as to whether we should take judicial notice of it. However, I do disagree with the majority's holding that witnesses cannot offer an estimate of speed without knowing precisely the distance traveled and time elapsed. This is not a classroom or a laboratory, nor are the roads upon which police officers patrol. District courts are often called upon to make determinations as to a witness's ability to offer opinion testimony, based upon the witness's personal observations. Estimates of speed fall squarely and historically within the realm of opinion testimony that may be received even from lay witnesses, and which fall within the province of the factfinder (be that the district court or a jury) to evaluate and weigh. Credibility determinations must be made, which likewise rest with the factfinder and not with us.

In sum, a lay witness’s estimation of speed, as opposed to an expert’s mathematical calculation of speed, is not tied to distance or time, or contingent upon a mathematical formula, when offered in a court of law, and the majority points to no case in which it has been. It is instead tied to the "witness's perception,” and, in particular, his fair opportunity to personally observe the moving object. See e.g. Fed.R.Evid. 701. It is an estimate, not "a guess,” based upon the witness’s personal observation, and the type of opinion evidence routinely received by courts. When evaluated in the context of probable cause, it is further evaluated with reference to the training and experience of the police officer, which, as it happens in this case, was quite substantial.

. The majority asserts that my concern regarding North Carolina’s certification test is nothing more than a policy consideration. I express no policy view as to whether North Carolina, Ohio, or any other state should certify officers in this manner, or' what their legislatures should require in the way of traffic stops and convictions. My concern is that the majority has invalidated what appears to be a well-accepted test for police officers to demonstrate their proficiency in this area, without any evidentiary support and based instead upon their common sense and judicial notice of the mathematical formula for speed. If the reliability of North Carolina’s certification procedure is to be challenged, I believe we should leave it to a defendant who chooses to challenge it and who introduces evidence in support of that challenge, and that the government should be given the opportunity to defend it.

. I would also affirm the district court’s decision that Deputy Elliott had reasonable suspicion sufficient to prolong the stop and perform a canine sniff while waiting for the BLOC information, and that he had probable cause to conduct a search of the vehicle once the canine alerted to illegal drugs.