United States v. Ernesto Godinez

In the

United States Court of Appeals
For the Seventh Circuit
____________________
No. 19‐3425
UNITED STATES OF AMERICA,
Plaintiff‐Appellee,
v.

ERNESTO GODINEZ,
Defendant‐Appellant.
____________________

Appeal from the United States District Court for the
Northern District of Illinois, Eastern Division.
No. 18 CR 278 — Harry D. Leinenweber, Judge.
____________________

SUBMITTED JANUARY 20, 2021 — DECIDED AUGUST 4, 2021
____________________

Before EASTERBROOK, WOOD, and BRENNAN, Circuit Judges.
BRENNAN, Circuit Judge. Law enforcement officers entered
a southwest Chicago neighborhood one night to replace
tracking devices on the cars of several Latin Saints gang mem‐
bers. Shortly after the officers arrived, they came under gun‐
fire and a federal agent was shot and seriously injured.
A federal grand jury indicted Ernesto Godinez, a member
of the gang, for the shooting. In the government’s view,
2 No. 19‐3425

Godinez, tasked with guarding the neighborhood, mistook
federal agents from the Bureau of Alcohol, Tobacco, and Fire‐
arms (“ATF”) for rival gang members and shot Special Agent
Kevin Crump. After a six‐day trial, a jury found Godinez
guilty.
Godinez now appeals, arguing that the district court
wrongly admitted certain evidence and that the jury did not
receive sufficient evidence to convict him of shooting Crump.
We conclude that the district court properly admitted ballis‐
tics evidence concerning the shots fired, although evidence
from and testimony about a gunshot detection system—
ShotSpotter—should have been handled differently. Because
a rational jury, even without the improperly admitted evi‐
dence, could have found beyond a reasonable doubt that
Godinez shot Crump, we affirm.
I
Given the jury’s verdict, we view the trial evidence in the
light most favorable to the government. United States v.
Wallace, 991 F.3d 810, 812 (7th Cir. 2021).1 A portion of the gov‐
ernment’s case consisted of video evidence compiled and
synchronized from various police and private surveillance
cameras. Some videos have multiple panes, allowing the
viewer to track the movement of individuals or cars. The
ATF’s technical specialist also inserted circles around cars or
persons of interest for the compiled videos. At trial, the de‐
fense used these videos as well, which without dispute depict

1The facts relayed below are taken from the trial transcript, videos,
and exhibits in the district court.
No. 19‐3425 3

Godinez. No video showed Godinez shooting a gun, how‐
ever.
A
On May 4, 2018, ATF agents and Chicago police officers
went to the “Back of the Yards” neighborhood2 to replace
court‐approved global positioning system trackers on cars be‐
longing to members of the Latin Saints gang. The neighbor‐
hood has a typical municipal grid pattern with streets at right
angles. These events took place between the 4300 and 4400
block of parallel north‐south streets, South Wood Street and
South Hermitage Avenue. An alley runs between and parallel
to those streets, and various gangways between houses allow
east‐west access mid‐block.
Latin Saints gang member Ernesto Godinez lived in the
Back of the Yards. Another gang member, Hector Ruiz, who
is also a paid government informant, confirmed that Godinez
was in the gang and thus would have been expected to patrol
the neighborhood. For the jury, Ruiz outlined the gang’s ter‐
ritory and described the expectations of Latin Saints to protect
that area from rival gang members by “[p]osting up,” “24/7.”
Ruiz affirmed that meant Godinez had to shoot any rival gang
members he spotted while on patrol, testifying that if rival
gang members enter the neighborhood, Latin Saints gang
members are to “[c]hase them or shoot at them.”

At 2:56 a.m. on May 4, video shows Godinez wearing dark
clothing leaving the area of his house on South Wood Street.
He drove around the neighborhood at a relatively slow speed,

2 So named for its proximity to the former Union Stock Yards. See Up‐

ton Sinclair, The Jungle, 2 (1906).
4 No. 19‐3425

as if patrolling, and he arrived home and parked near the in‐
tersection of 44th and Wood.
At around 3:15 a.m., ATF agents in street clothes and an
unmarked car drove slowly around this same neighborhood
to assess how safe it was and to locate the target cars. They
parked on the 4400 block of South Hermitage. When this first
undercover car passed, video shows Godinez running into his
house, leaving shortly after, and then running north. He
crossed the alley and then entered a gangway between the
houses at 4332 and 4336 South Hermitage Avenue.
At about 3:17 a.m., a second group of agents, using a rental
car to avoid detection, drove down Hermitage. These plain
clothes agents—Crump, Daniel Winter, and Thomas
Spratte—exited the car near the intersection of 44th and Her‐
mitage to approach the target vehicles on foot. They wore
sweatshirts with hoods up.
At 3:18 a.m., those same agents began to cross the street
and heard five gunshots. Winter and Spratte both recalled
that the shots came from the northwest. Spratte sought cover
behind a parked car and looked over his shoulder. He testi‐
fied that he saw the shots come from about halfway up the
4300 block of South Hermitage Avenue on the west side of the
sidewalk. Spratte saw two muzzle flashes, immediately fired
his gun twice toward those flashes, and yelled, “shots fired.”
When Winter asked Spratte where the shots came from,
Spratte pointed to the location of the two muzzle flashes.
Spratte did observe an individual with a white t‐shirt on Her‐
mitage, although he does not recall exactly when. But Spratte
did not see the person next to the muzzle flashes, and Spratte
did not fire his gun at that person.
No. 19‐3425 5

As the agents started to head north towards the shots, they
realized that Crump had been hit and was on the ground
bleeding. The bullet entered Crump’s neck just below his left
ear, exiting his face through the bridge of his nose between
his eyes. Since being shot, Crump has undergone two surger‐
ies, and the wound affects his vision. Crump did not see who
shot him or where the shots came from.
At 3:18:26 a.m., immediately after the shooting, video
shows Godinez running across the alley and heading toward
his house. Godinez moves with his right hand at his side, but
the video does not reveal whether he is holding something.
Before the shooting, Godinez had been exchanging mes‐
sages via Snapchat with his then‐girlfriend, Valerie Jean‐Bap‐
tiste, who lived in the neighborhood. Godinez told Jean‐Bap‐
tiste to pick him up by 44th and Wood. Before she left her
house, Jean‐Baptiste heard the shooting outside her open win‐
dows. Then she quickly sent a message to Godinez asking his
location. He responded, “[b]y his house,” so Jean‐Baptiste
drove her car there. Video showed her arriving in a sedan and
picking Godinez up within five minutes of the shooting.
Godinez entered her car wearing dark clothing and a baseball
cap that was later found in the car and containing his DNA.
Jean‐Baptiste testified that Godinez told her: “I feel good. F***
that flake.” Jean‐Baptiste confirmed that Godinez was the per‐
son in the videos.
Together, Jean‐Baptiste and Godinez drove to a nearby gas
station. On the way, they downloaded a cellphone application
for a police scanner to monitor law enforcement activity in the
neighborhood. They next drove to a second gas station where
Godinez bought a white t‐shirt that he put on over his black
shirt.
6 No. 19‐3425

Later that morning, Godinez went to his cousin’s house
and dropped off the car he had driven earlier that day to pa‐
trol the neighborhood. In the evening, Godinez returned to
that house with his son and son’s mother, Destiny Rodriguez.
She showed Godinez a purported “wanted” picture with his
face on it and started to cry. Hugging Rodriguez and their
son, Godinez said he loved Rodriguez and that he was sorry
for everything he ever did that hurt her. At that house,
Godinez also stopped using one of his cellphones, which he
left there.
The same day of the shooting, Chicago police went to the
Back of the Yards neighborhood to investigate. At trial, Foren‐
sic Investigator Paul Presnell testified about his work at the
scene, which included videotaping, photographing, drawing
a plat map, and searching for and marking evidence. Presnell
described the ballistics evidence recovered at the scene. Five
casings were found in a gangway between 4332 and 4336
South Hermitage Avenue. Those casings were clean and
shiny, not weathered. Presnell came upon four of these cas‐
ings after other officers had placed yellow crime scene mark‐
ers noting their location. He photographed and collected
those four casings, and after he left the gangway to search fur‐
ther, he found a fifth casing nearby under a leaf. Presnell also
found a bullet lodged in the north side of a tree in front of
4343 South Hermitage Avenue and a bullet in the grass be‐
tween the street and the sidewalk in front of 4416 South Her‐
mitage Avenue.
The jury heard expert testimony from ATF Firearms Ex‐
aminer Arnold Esposito as well. He testified that the five cas‐
ings recovered from the gangway were fired from the same
9mm caliber firearm; that the bullet from the tree and the
No. 19‐3425 7

bullet found in the grass were both fired from the same gun;
and that most of the possible firearms from which the bullets
could have been fired were 9mm caliber.
The government also presented evidence from ShotSpot‐
ter, an acoustic gunshot detection and location system. The
district court qualified ShotSpotter employee Paul Greene as
an expert witness. Greene described ShotSpotter as a series of
sensors installed in a geographic area that listen specifically
for the sound of gunfire. When shots are detected, the time
and certain measurements are sent to a central location where
software classifies the sound as gunfire or not. If it is gunfire,
ShotSpotter calculates the point of latitude and longitude
where the gun was fired. That information is then reported to
law enforcement.
Greene testified about his analysis of the May 4 audio clips
from the sensors near 44th and Hermitage. Initially, ShotSpot‐
ter identified two gunshots as having been fired near the
southeast corner of 44th and Hermitage between 3:18:14 a.m.
and 3:18:15 a.m. Greene testified that ShotSpotter typically
has an accuracy of 10 to 12 feet. An hour or so later, Chicago
police contacted ShotSpotter and asked them to search for ad‐
ditional audio clips of five shots, which had not been auto‐
matically identified as gunshots. Greene said this happens
“on a semiregular basis with all of our customers.” ShotSpot‐
ter located those audio clips and identified five gunshots lo‐
cated at 4338 South Hermitage Avenue, between 3:18:10 a.m.
8 No. 19‐3425

and 3:18:12 a.m. The district court admitted the audio record‐
ings of the shots into evidence.3
The Snapchat text messages exchanged between Godinez
and Jean‐Baptiste before and after the shooting were also re‐
ceived into evidence, as were location records from their
Snapchat accounts, which are accurate within 100 meters. The
government created animations to show the location of their
phones on May 4, 2018. Those animations indicated that
Godinez’s phone was near the location of the shooting when
it occurred, as well as near the home of his cousin later the
same day.
In addition, the jury heard from Chicago Police Sergeant
Neil Evans, who interviewed Spratte the morning of the
shooting. Evans was questioned about Spratte’s statements
that he had seen muzzle flashes halfway up the 4300 block of
South Hermitage Avenue as well as an unknown person in a
white t‐shirt. ATF Special Agent Beau Jacobsen also testified
about video showing an unknown person wearing a white t‐
shirt at the south end of the alley between South Wood Street
and South Hermitage Avenue at 2:52 a.m., 26 minutes before
the shooting.

3 Before the jury, Greene described how ShotSpotter works. He testi‐
fied that ShotSpotter employs “multilateration”—its term for using sev‐
eral known geographic points to identify an unknown one. It calculates a
sound’s different arrival times at different sensors to plot a hyperbolic
curve. Repeating this process over multiple pairs of sensors generates a
series of curves. Plotting the curves on a map results in the curves inter‐
secting at a geographic coordinate with a specific latitude and longitude.
ShotSpotter then uses software to match that geographic coordinate with
a street address along with a time of the shooting and the number of
rounds fired.
No. 19‐3425 9

B
A federal grand jury indicted Godinez for forcibly assault‐
ing a federal officer while using a deadly weapon in violation
of 18 U.S.C. § 111(a) and (b), along with discharging a firearm
during a crime of violence in violation of 18 U.S.C.
§ 924(c)(1)(A)(iii). For both crimes, the only element in dispute
was whether Godinez was the person who shot Crump.
Godinez moved in limine to bar the ballistics evidence and
firearms expert Esposito’s testimony, as well as the ShotSpot‐
ter evidence and ShotSpotter employee Greene’s testimony.
Godinez argued the government had not matched any spent
casings to bullets and that no evidence had been presented as
to the make and model of the gun used or as to when the bul‐
lets were fired or casings released. On the admissibility of the
ShotSpotter evidence under Federal Rule of Evidence 702,
Godinez also requested that the district court conduct a hear‐
ing under Daubert v. Merrell Dow Pharmaceuticals, Inc., 509 U.S.
579 (1993).
In his motion in limine, Godinez argued that the ShotSpot‐
ter methodology and its underlying data were not the product
of reliable principles and methods. ShotSpotter’s program‐
ming and algorithms have never been peer reviewed,
Godinez contended, so the district court should hold a
Daubert hearing and require the government to establish
ShotSpotter’s reliability. If the district court did find the tech‐
nology reliable under Daubert, Godinez argued that ShotSpot‐
ter had deliberately placed the approximate locations of shots
at the crime scene, so that evidence should be excluded as un‐
duly prejudicial under Federal Rule of Evidence 403.
10 No. 19‐3425

In pretrial rulings, the district court orally denied
Godinez’s evidentiary motions. First, the district court con‐
ducted a Rule 403 analysis and concluded that the ballistics
evidence was relevant and not overly prejudicial to Godinez.
To the district court, that evidence could be used to show how
the bullets were matched to each other and how the casings
were matched to each other. The government planned to offer
expert testimony to shed light on what type of gun was
used—a matter, the district court concluded, central to the
jury’s decision.
The district court also ruled that the ShotSpotter evidence
was admissible under Rule 702 and Daubert. The district court
relied upon State v. Hill, in which the Supreme Court of Ne‐
braska affirmed a trial court’s qualification of Greene as an
expert and its determination that the ShotSpotter system was
sufficiently reliable. 851 N.W.2d 670, 689–90 (Neb. 2014).
Godinez had cited no authorities to demonstrate otherwise,
the district court noted, and cross‐examination could test the
accuracy of Greene’s methods and conclusions. Ultimately,
the district court remained unconvinced that the ShotSpotter
evidence was unduly prejudicial.
Later in that hearing, Godinez argued that no party in the
jurisdiction had challenged ShotSpotter’s methodology under
Daubert. So Godinez again asked for a Daubert hearing and
that Greene testify outside the jury’s presence. According to
Godinez, this would allow for vetting of ShotSpotter’s tech‐
nology and methodology.
The district court rejected as unnecessary Godinez’s re‐
quest for a Daubert hearing, noting that this type of evidence
had been tested in 80 cases involving Greene. Without
Godinez calling an expert witness on this topic, the district
No. 19‐3425 11

court concluded that the government had satisfied the re‐
quirements of Daubert. Still, the district court devised a com‐
promise—the government was to make Greene available for
questioning by the defense before cross‐examination on the
stand, so the defense would have an advantage. For the dis‐
trict court:
[T]he principles that [Greene]’s espousing, it’s not a
question where he’s pulling this out of the air. … the
principles, I think, are valid. Whether or not they’ve
been properly utilized, of course, is subject to cross ex‐
amination. … That’s all that Daubert requires.
The district court reasoned those witnesses would also cor‐
roborate where the shots came from on Hermitage. So it stood
by its initial ruling denying Godinez’s motion to bar the
ShotSpotter evidence and Greene’s testimony.
After a six‐day trial, a jury found Godinez guilty on both
counts. He then moved for a judgment of acquittal under Fed‐
eral Rule of Criminal Procedure 29 and for a new trial under
Federal Rule of Criminal Procedure 33. The district court de‐
nied both motions. Godinez received a total sentence of 200
months’ imprisonment and timely appealed his convictions.
II
First, we address the district court’s decisions to admit: (A)
the ballistics evidence, and (B) the ShotSpotter evidence, in‐
cluding to qualify Paul Greene as an expert witness.
A
The district court’s admission of the ballistics‐related evi‐
dence—the testimony from forensic investigator Presnell and
firearms examiner Esposito, as well as the physical evidence
12 No. 19‐3425

of the bullets and casings—is reviewed for abuse of discre‐
tion. United States v. Johnson, 916 F.3d 579, 588 (7th Cir. 2019).
Godinez argues that Presnell should not have been per‐
mitted to testify that the casings were located at the mouth of
the gangway between the houses at 4332 and 4336 South Her‐
mitage Avenue. Recall that Presnell did not know who found
four of the casings, only that he came upon them after yellow
crime scene markers already noted their location. To Godinez,
this meant that reasonable precautions required to show a
chain of custody had not been taken. For example, after
Presnell left the gangway to continue his search, it was left
unsecured, and only after Presnell returned did he find a fifth
casing under a leaf.
Godinez’s contention falls short, though, because a chain
of custody need not be perfect. “[T]he government need only
show that it took reasonable precautions to preserve the orig‐
inal condition of the evidence.” United States v. Lee, 502 F.3d
691, 697 (7th Cir. 2007) (internal quotation marks omitted).
“When the evidence is in police custody, a presumption of
regularity applies; in the absence of evidence to the contrary,
we assume the police did not tamper with the evidence. And
any gaps in the chain of custody go to the weight given the
evidence, not its admissibility.” United States v. Vitrano, 747
F.3d 922, 925 (7th Cir. 2014) (citations omitted). At trial,
Presnell described the reasonable precautions undertaken to
handle this ballistics evidence. These included wearing latex
gloves, placing it into envelopes or bags, sealing those in a
clear plastic bag with the date and his identifying infor‐
mation, placing the evidence in inventory using the Chicago
Police Department “eTrack” system, storing the bag in his
locked office, and recovering the evidence using barcode
No. 19‐3425 13

scanning. Godinez was free to, and did, raise and explore dur‐
ing cross‐examination any alleged gaps in the custody chain.
But because the evidence was in police custody with no signs
of tampering, the presumption of regularity applies. See id.
Therefore, the district court did not abuse its discretion in al‐
lowing Presnell to testify to the location of the casings.
According to Godinez, no connection exists between the
casings found in the gangway between 4332 and 4336 South
Hermitage Avenue and the bullet that struck Crump. Bullets
and casings are strewn throughout a violent, high‐crime
neighborhood like Back of the Yards, he claims. For Godinez,
the presence of the casings near the gangway—and of bullets
in a straight line from where the casings were found—does
not make it more likely that the officers were shot at from the
gangway, so this evidence is not relevant and should not have
been admitted.
We disagree. Although the gun used to shoot Crump was
not recovered, that does not render inadmissible the physical
evidence that was recovered. Esposito testified that the five
casings located in the gangway were all fired from the same
gun. And according to Presnell, those casings were shiny, not
weathered; so the jury could infer they resulted from a recent
shooting. Esposito also testified that the two bullets found
were fired from the same gun, and that the gun was most
likely a 9 mm firearm. This type of gun usually ejects casings
to the right. As the government argued at closing, if the
shooter stood at the mouth of the gangway and fired south
down Hermitage toward its intersection with 44th where
Crump was hit, the casings would have ejected from the gun’s
right, landing where they were found—five to ten feet into the
gangway. Esposito’s testimony at trial thus fits, rather than
14 No. 19‐3425

conflicts, with what had been recovered on the ground after
the shooting. Because Godinez has offered only speculation
as to how five fresh casings ejected from the same gun were
found in this same location, the district court did not abuse its
discretion in admitting the ballistics evidence.
B
Next, we consider the admission of the ShotSpotter evi‐
dence and the expert testimony of Paul Greene, a ShotSpotter
employee.
Federal Rule of Evidence 702 and Daubert govern the ad‐
missibility of expert testimony. We review de novo whether a
district court properly applied the framework in Rule 702 and
Daubert. United States v. Jett, 908 F.3d 252, 265 (7th Cir. 2018).
If the district court did so, the decision to admit or exclude
expert testimony is reviewed for abuse of discretion. Id. But if
the district court failed to perform a Daubert analysis, the ad‐
missibility of the expert testimony must be reviewed de novo.
Kirk v. Clark Equip. Co., 991 F.3d 865, 872–73 (7th Cir. 2021)
(collecting cases); Metavante Corp. v. Emigrant Sav. Bank, 619
F.3d 748, 760 (7th Cir. 2010).
Here, the district court sufficiently grounded its rulings on
the ShotSpotter evidence within the Daubert framework to en‐
sure deferential standard of review. To be sure, a failure to
perform a Daubert analysis entirely, or a conclusory, one‐sen‐
tence ruling, can trigger de novo review. See Baugh v. Cuprum
S.A. de C.V., 845 F.3d 838, 844 (7th Cir. 2017); Metavante, 619
F.3d at 760. But the “district court’s analysis here, although
concise, is much more thorough than these conclusory deter‐
minations.” Kirk, 991 F.3d at 873. So we review the district
court’s analysis for an abuse of discretion.
No. 19‐3425 15

Among Rule 702’s requirements are that a witness must
be “qualified as an expert by knowledge, skill, experience,
training, or education” and that the expert’s testimony must
be “the product of reliable principles and methods.” FED. R.
EVID. 702. In Daubert, the Supreme Court explained that Rule
702 grants to the district court the gatekeeping responsibility
to ensure the reliability of proposed expert testimony. 509
U.S. at 589. In serving as gatekeeper, the district court is to
evaluate: “(1) the proffered expert’s qualifications; (2) the re‐
liability of the expert’s methodology; and (3) the relevance of
the expert’s testimony.” Kirk, 991 F.3d at 872 (internal quota‐
tion marks omitted).
Godinez asked for a Daubert hearing three times—in his
motion in limine, during a pretrial hearing, and just before
Greene testified in front of the jury. Although Godinez pri‐
marily challenged the reliability of the ShotSpotter methodol‐
ogy, at times he also contested Greene’s qualifications.
Whether a Daubert hearing is necessary before trial lies within
the district court’s discretion, and here the district court de‐
clined Godinez’s requests for a hearing as unnecessary. In‐
stead, the district court performed its Daubert analysis based
on the parties’ written submissions and their oral argument.4
The district court initially strayed in its Daubert analysis by
relying on the Supreme Court of Nebraska’s decision in Hill.
Unlike Godinez, the criminal defendant in Hill did not “chal‐
lenge the underlying GPS triangulation methodology upon
which the ShotSpotter location is based.” 851 N.W.2d at 690.

4 We note that the compromise offered by the district court—allowing
the defense to question Greene before he took the stand—does not appear
to have occurred.
16 No. 19‐3425

Additionally, the trial court in Hill conducted a Daubert hear‐
ing, id. at 679–81, unlike the district court in this case. That
Greene had been qualified as an expert in several other cases,
as the district court noted, is material to any Daubert analysis.
But his qualification in Hill does not ensure the reliability of
ShotSpotter’s methodology here.
In concluding that Daubert had been satisfied, the district
court also stated that the principles underlying ShotSpotter
are not pulled “out of the air” and pointed out Godinez’s
opportunity to cross‐examine Greene on whether those prin‐
ciples were properly utilized. Godinez’s objections and
arguments in this case, however, merited an opportunity to
question and dispute those principles and their reliability be‐
fore trial.
Most important for us in this review is that the ShotSpotter
evidence admitted at trial implicated that system’s methodol‐
ogy. Recall that at first, the system identified two gunshots—
Spratte’s return fire. Then, after Chicago police contacted
ShotSpotter and asked them to search for additional audio
clips, the system identified five more preceding shots. This
identification and location process goes to how ShotSpotter
collects, analyzes, and reports its data—its methodology. In
this case, the district court never sufficiently explored this is‐
sue. Criticism of that methodology formed part of the defense
here, therefore deserving a more thorough exploration by the
district court.
To be sure, we have no reason to conclude that the
ShotSpotter reports or analysis were manipulated to support
a certain outcome. Greene’s testimony may well be correct
that the request to review audio for further shots occurs “on
a semiregular basis with all of our customers.” And the
No. 19‐3425 17

ShotSpotter system may adhere to the requirements of the
best available science. But without a more searching examina‐
tion of ShotSpotter’s methods under Daubert, we cannot con‐
clude that this evidence was properly admitted against
Godinez. After giving due deference to the district court’s
consideration of the record, we conclude that it abused its dis‐
cretion in admitting Greene’s testimony and the ShotSpotter
evidence.
III
In addition to his evidentiary arguments, Godinez con‐
tends that insufficient evidence was presented to convict him
of forcibly assaulting a federal officer while using a deadly
weapon and of discharging a firearm during a crime of vio‐
lence. We review de novo a district court’s denial of a motion
for a judgment of acquittal for insufficient evidence. Wallace,
991 F.3d at 812. In doing so, we construe the evidence in the
light most favorable to the government, and we affirm a jury’s
verdict if any rational trier of fact could have found the of‐
fense’s elements satisfied beyond a reasonable doubt. Id.; see
Jackson v. Virginia, 443 U.S. 307, 319 (1979).
To succeed on a claim of insufficient evidence at trial, a
convicted defendant must clear what this court has consist‐
ently characterized as a high bar. See United States v. Torres‐
Chavez, 744 F.3d 988, 993 (7th Cir. 2014) (“nearly insurmount‐
able hurdle”); United States v. Griffin, 684 F.3d 691, 694 (7th
Cir. 2012) (“a heavy burden” (internal quotation marks omit‐
ted)); United States v. McCaffrey, 181 F.3d 854, 856 (7th Cir.
1999) (“a daunting task”). “We will reverse a conviction only
where the record is devoid of evidence from which a reason‐
able jury could find guilt beyond a reasonable doubt.” United
States v. Khan, 937 F.3d 1042, 1055 (7th Cir. 2019) (citing United
18 No. 19‐3425

States v. Heon Seok Lee, 937 F.3d 797, 807–08 (7th Cir. 2019)).
And because we afford great deference to a jury’s verdict of
conviction, Khan, 937 F.3d at 1055, that deference to the jury’s
deliberations prevents us from assessing the quality of the ev‐
idence. See, e.g., United States v. Smallwood, 188 F.3d 905, 914
(7th Cir. 1999). In other words, we respect “the exclusive func‐
tion of the jury to determine the credibility of witnesses, re‐
solve evidentiary conflicts, and draw reasonable inferences.”
United States v. Reed, 875 F.2d 107, 111 (7th Cir. 1989) (internal
quotation marks omitted).
On this argument, we review the entirety of the trial rec‐
ord without the ShotSpotter evidence and Greene’s testi‐
mony. As for the remaining evidence, some of it is circum‐
stantial. But an implication that circumstantial evidence is
weaker than direct evidence is incorrect. See Murrell v. Frank,
332 F.3d 1102, 1117 (7th Cir. 2003) (“Circumstantial evidence
is of equal probative value to direct evidence and in some
cases is even more reliable.” (internal quotation marks omit‐
ted)); United States v. Moore, 115 F.3d 1348, 1364 (7th Cir. 1997)
(noting much the same). We note that the district court here
correctly instructed the jury, under the Pattern Criminal Jury
Instructions of the Seventh Circuit No. 2.03 (2012 ed.), that
“[t]he law does not say that one is better than the other.”
For the crimes he faced, only one element was in dispute
at trial: was Godinez the person who shot Crump? Based on
the location of the shots, Godinez’s whereabouts during the
shooting, his motive, and his actions after the shooting, suffi‐
cient evidence supported the jury’s decision that Godinez
shot Crump.
The location from where the shots were fired. From the gang‐
way between 4332 and 4336 South Hermitage Avenue,
No. 19‐3425 19

Presnell recovered five shiny casings. ATF expert Esposito
testified that all five casings had been ejected from the same
9mm caliber gun, and that a recovered bullet was fired from
what was most likely a 9mm caliber gun. The parties stipu‐
lated that this recovered bullet contained Crump’s DNA.
Godinez objects to this ballistics evidence, as he did at trial.
But as discussed, the district court did not err in admitting
that evidence. The jury could properly draw a straight line
from the gangway—which the videos show Godinez entering
before and exiting after the shooting, and in which the casings
were recovered—to the tree where one of the bullets was re‐
covered and the grass where the second bullet was found.
From this evidence, a rational jury could find that the casings
and bullets were fired from the same gun and by a person
shooting from the gangway. Godinez’s speculation that the
casings and bullets came from other shootings cannot over‐
come the jury’s permissible inference on this point.
At trial, Winter and Spratte—who were near Crump when
he was shot—also corroborated the location from where the
shots were fired. Winter testified the shots were fired from
northwest of the agents’ location, which was near the gang‐
way. And Spratte, who returned fire, testified to the location
of the muzzle flares (halfway up the 4300 block of South Her‐
mitage Avenue on the west side of the sidewalk) as by the
gangway. Nevertheless, no video directly captures the shots
being fired. The government maintained at trial that an awn‐
ing blocked the camera from capturing the firing of the gun.
Godinez responds that the muzzle flashes still should have
been visible, especially in the dark after 3 a.m. But even if the
flashes were not seen on video, Spratte testified he saw them
come from the same area of Hermitage as the gangway.
20 No. 19‐3425

Whether to favor Godinez’s theory or Spratte’s testimony is a
decision for the jury.
Godinez’s whereabouts during the shooting. Godinez’s then‐
girlfriend, Jean‐Baptiste, identified him in the video presented
to the jury, and Godinez does not contest that he is depicted.
Video shows him dressed in dark clothing and moving from
where he lived on South Wood Street across the alley into the
gangway. After the shooting, Godinez is shown running back
across the alley with his hand stationary at his right side mov‐
ing at a faster pace, for example, than he is depicted later in
video taken at the gas station he and Jean‐Baptiste visit. View‐
ing the evidence in the light most favorable to the govern‐
ment, a rational jury could have found that video places
Godinez where the shots were fired at the time they were
fired. Our dissenting colleague writes that we “oversimplify
the facts when [we] assert that the jury could have drawn a
straight line ‘from the gangway’ point where Godinez entered
and exited, and the spot where the bullet casings were recov‐
ered, to the area on S. Hermitage where the bullets themselves
were found.” Post at 28–29. Rather, the evidence, taken as a
whole, permits the jury to reach the conclusion that Godinez
was at that location when the shooting occurred.
Against all this, Godinez offers an alternative explanation
of who shot Crump—the same one he offered at trial. Spratte
told Evans that he saw muzzle flashes and someone wearing
a white t‐shirt down the street. On appeal, Godinez again as‐
serts the person wearing a white t‐shirt was the shooter.
Godinez argues he could not have been the shooter because
at the time of the shooting—as the video shows—he was
wearing a black shirt, and he did not wear a white t‐shirt until
No. 19‐3425 21

he bought one at the second gas station that he and Jean‐Bap‐
tiste visited sometime after the shooting.
But at trial, Spratte did not identify the person in the white
t‐shirt as the shooter. Instead, Spratte testified he did not re‐
call when he saw the person in the white t‐shirt in relation to
the shots being fired, and he did not see the person in white
behind the muzzle flashes. Moreover, video shows a person
in a white t‐shirt appearing in a different location than the
gangway—at the south end of the alley between Wood and
Hermitage—26 minutes before the shooting, as Jacobsen testi‐
fied. So a rational jury could reject Godinez’s theory of the
mystery shooter in a white t‐shirt. By this argument, Godinez
asks us to assess the evidence and to resolve this possible ev‐
identiary conflict, a task beyond our appellate role. See Small‐
wood, 188 F.3d at 914.
Motive. Ruiz testified that Godinez, as a Latin Saint, was
required to patrol the gang’s territory and to shoot at rival
gang members who entered that territory. At trial, Godinez
took the position that he was not part of protecting the gang’s
turf. But video from the minutes leading up to the shooting
shows Godinez slowly driving around the neighborhood, as
if patrolling. Noticing a slow‐moving, unfamiliar car (which
contained the agents), Godinez next entered and quickly ex‐
ited his house (presumably to get a gun). He then ran to the
gangway opening onto Hermitage, fired shots, and ran back
home. Mistaking the agents for rival gang members fits with
Godinez’s actions before, during, and after the shooting.
In addition, Jean‐Baptiste testified that Godinez sent her a
Snapchat message shortly after the shooting asking her to
pick him up. And after getting in Jean‐Baptiste’s car, Godinez
said, “I feel good. F*** that flake.” According to Ruiz’s
22 No. 19‐3425

testimony, a “flake” means a rival gang member. The jury
could reasonably interpret this statement as Godinez thinking
he had shot at a rival gang member. Still, Godinez argues that
Jean‐Baptiste testified at trial only that he said, “’I feel good,
f***’ and something about ‘flake.’” But when pressed, Jean‐
Baptiste testified that Godinez said something like “f*** that
flake” or “f*** those flakes.” To the extent Jean‐Baptiste back‐
tracked at trial, the government impeached her using her
grand jury testimony. Again, what account to credit on this
score is a determination for the fact finder, here the jury.
Godinez’s actions after the shooting. Godinez and Jean‐Bap‐
tiste drove to a gas station after the shooting, on the way
downloading a police scanner cell phone application to mon‐
itor law enforcement in the neighborhood. True, they may
have wanted to know what was going on there. But a rational
jury could also infer that Godinez—aware that law enforce‐
ment officers had flooded the Back of the Yards and estab‐
lished two perimeters around the crime scene—wanted to
avoid questioning or capture. From Godinez later leaving his
car and phone at a cousin’s house, a rational jury similarly
could infer that he wanted to keep a low profile. At a second
gas station, Godinez bought a white t‐shirt, which he put on
over the black shirt he wore during the shooting, allowing a
rational jury to infer that he wished to change his appearance.
And from Godinez’s reaction to Rodriguez’s expression of
emotion, a rational jury could also conclude Godinez regret‐
ted his actions, including the shooting. Viewed in the light
most favorable to the government, this evidence of what
Godinez did after the shooting could allow a rational jury to
find that Godinez shot Crump.
No. 19‐3425 23

Based on the ample and intersecting evidence presented
during the six‐day trial, a rational jury could conclude beyond
a reasonable doubt that Godinez shot Crump. Our dissenting
colleague disagrees, offering competing inferences from the
evidence at each step. The jury could have reached those in‐
ferences, but it did not. And even if no single inculpatory fact
proves guilt, the complementary nature of the evidence as a
whole permitted a reasonable jury to find Godinez guilty. Be‐
cause Godinez has not shown that the government submitted
no evidence for a rational trier of fact to find him guilty be‐
yond a reasonable doubt, we must defer to the jury’s verdict
on both counts. See Jackson, 443 U.S. at 319; Khan, 937 F.3d at
1055.
On a final note, Godinez and the government disagree as
to whether the admission of the ShotSpotter evidence and
Greene’s related expert testimony was harmless error. Federal
Rule of Criminal Procedure 52(a) states “[a]ny error … that
does not affect substantial rights must be disregarded.” We
have concluded that it was erroneous to admit that evidence
and related testimony without a Daubert hearing, so the bur‐
den rests on the government to show that the error was harm‐
less. See United States v. Olano, 507 U.S. 725, 741 (1993). “The
test for harmless error is whether, in the mind of the average
juror, the prosecution’s case would have been significantly
less persuasive had the improper evidence been excluded.”
United States v. Curtis, 781 F.3d 904, 911 (7th Cir. 2015). “We
will affirm if ‘the error had no substantial influence on the
verdict’ because ‘other untainted incriminating evidence is
overwhelming.’” United States v. Chaparro, 956 F.3d 462, 482
(7th Cir. 2020) (quoting United States v. Zuniga, 767 F.3d 712,
717 (7th Cir. 2014)). So “[e]ssentially, an evidentiary error is
harmless if it did not have a substantial influence on the
24 No. 19‐3425

verdict.” United States v. Pulliam, 973 F.3d 775, 782 (7th Cir.
2020), as amended (Sept. 8, 2020).
The government has met its burden here. The ShotSpotter
evidence was cumulative of other evidence presented at trial
as to the location of the shots:
 Spratte observed where the shots came from and the
muzzle flashes;
 Winter saw the shots came from the northwest;
 the casings were found in the gangway between 4332
and 4336 South Hermitage Avenue;
 the plat map showed that the gangway where the cas‐
ings were recovered was north and west of the agents;
and
 a straight line could be drawn from the gangway to
where the bullets were found and Crump was shot.
So ShotSpotter was not the only evidence pointing to the lo‐
cation from where the shots were fired. Godinez argues that
the ShotSpotter evidence—mentioned in the government’s
closing argument—“filled [in the] holes” of the circumstantial
case presented against him. But that contention misunder‐
stands that circumstantial evidence is evidence all the same.
See Murrell, 332 F.3d at 1117; Moore, 115 F.3d at 1364. Parsing
circumstantial evidence from direct evidence reweighs, rather
than reviews, that evidence. As the Pattern Criminal Jury In‐
structions of the Seventh Circuit No. 2.03 (2012 ed.) states, “[i]t
is up to [the jury] to decide how much weight to give to any
evidence, whether direct or circumstantial.”
Given the aggregate evidence about the location of the
shots, the government’s case against Godinez was not
No. 19‐3425 25

significantly less persuasive without the ShotSpotter evi‐
dence. And looking beyond the shooting location evidence
only confirms our conclusion. The rest of the evidence against
Godinez—his whereabouts at the time of shooting confirmed
by video and Snapchat records, his motive, and his actions
before and after the shooting—was overwhelming. Based on
all of this, the ShotSpotter evidence did not have a substantial
influence on the verdicts, rendering its admission harmless.
IV
For these reasons, Godinez’s convictions are AFFIRMED.
26 No. 19‐3425

WOOD, Circuit Judge, dissenting. Kevin Crump, a federal
agent with the Bureau of Alcohol, Tobacco, Firearms and Ex‐
plosives (“ATF”), was shot in the head while conducting a
late‐night field operation in Chicago at around 3:18 am on
May 4, 2018. Fortunately, despite his catastrophic injury
Agent Crump survived. He later testified in the trial against
his alleged assailant, Ernesto Godinez, a longtime resident of
the neighborhood where the shooting took place.
At first glance, it may seem that there was no shortage of
evidence against Godinez. But, as I detail below, a closer look
at the record shows that the government failed to connect the
dots between isolated pieces of information, and in the end, it
failed to present sufficient evidence to permit a jury to decide,
beyond a reasonable doubt, that Godinez was the shooter.
The government’s theory was that Godinez stood at the
entrance of a gangway between 4332 and 4336 S. Hermitage
Avenue and fired several rounds southward, toward the ATF
agents who stood at the intersection of W. 44th St. and S. Her‐
mitage. (A “gangway” is a narrow alley between two build‐
ings with an entrance on the front street and an exit onto a
rear alley). After a six‐day trial, a jury found Godinez guilty
of assaulting a federal officer with a deadly weapon and using
a gun to commit a violent felony. In a post‐trial motion for
acquittal, FED. R. CRIM. P. 29, Godinez insisted that the evi‐
dence did not prove his guilt beyond a reasonable doubt. My
colleagues, however, are satisfied that the evidence squeaked
by the deferential standard that applies to review of jury ver‐
dicts, and so they affirm the district court’s denial of
Godinez’s motion for acquittal and hold that he was properly
convicted.
No. 19‐3425 27

Although I recognize that it is rarely error to affirm a jury’s
verdict, I believe that this case is the exception that proves the
rule. Even viewing the evidence in the light most favorable to
the government as required by Jackson v. Virginia, I would
hold that no rational trier of fact could convict Godinez be‐
yond a reasonable doubt. 443 U.S. 307, 319 (1979). See also
Musacchio v. United States, 577 U.S. 237, 243 (2016) (applying
Jackson standard to a direct criminal appeal). At best, the gov‐
ernment presented a case consistent with Godinez’s being the
gunman, but no evidence (if believed) was enough to push
the ball over the mid‐field mark, much less to prove guilt be‐
yond a reasonable doubt. Accordingly, I respectfully dissent.
I
The rule in Jackson v. Virginia guides the inquiry for Rule
29 motions for acquittal. 443 U.S. 307; United States v. Torres‐
Chavez, 744 F.3d 988, 993 (7th Cir. 2014). Under this rule, we
view the evidence in the light most favorable to the govern‐
ment and “defer to the credibility determination[s] of the
jury.” United States v. Blassingame, 197 F.3d 271, 284 (7th Cir.
1999). We may “overturn a verdict only when the record con‐
tains no evidence, regardless of how it is weighed, from
which the jury could find guilt beyond a reasonable doubt.”
Id. While few criminal defendants surmount the hurdle pre‐
sented by Jackson, relief is not unheard of. See, e.g., Smith v.
Brookhart, 996 F.3d 402 (7th Cir. 2021) (granting relief in the
habeas corpus context, which if anything is even more defer‐
ential to the trial outcome).
A. The Evidence Against Godinez
The government presented ample evidence to establish
that someone located in the entrance of the gangway between
28 No. 19‐3425

4331 and 4336 S. Hermitage Ave fired the gunshots. Aside
from the ShotSpotter testimony which the majority thor‐
oughly discusses, there is the following evidence in the rec‐
ord: (1) five bullet casings recovered at the mouth of the gang‐
way, (2) eyewitness testimony from ATF agents describing
muzzle flashes in the direction of the gangway, and (3) bullets
lodged in various locations, whose reconstructed trajectories
suggest that they were fired from the gangway. This is
enough for any rational trier of fact to conclude that whoever
shot Agent Crump did so from the entrance of the gangway.
But nothing in that evidence was enough to identify the
shooter.
The evidence did show that Godinez was in the general
area at the time the shooting took place. But there was nothing
necessarily incriminating about that fact. He lived just north
of the intersection of W. 44th Street and S. Wood Street. His
home was just one block west of the intersection where Agent
Crump was shot. Tr. 629–30. Real‐time geolocation data from
his Snapchat account (a social media application) confirm that
at 3:18 am he was within 100 meters (about 110 yards—more
than a football field) of the shooting at 3:18 am. Security foot‐
age visually confirms his presence near 44th and Wood before
the shooting. And a security camera in the alley between
S. Wood Street and S. Hermitage Avenue also captures
Godinez crossing the north‐south alley before the shooting
(heading westward towards Hermitage) and crossing back af‐
ter the shooting (heading eastward towards Wood). On his
return trip, he seems to run across the alley with his right
hand stationary at his hip.
My colleagues oversimply the facts when they assert that
the jury could have drawn a straight line “from the gangway”
No. 19‐3425 29

point where Godinez entered and exited, and the spot where
the bullet casings were recovered, to the area on S. Hermitage
where the bullets themselves were found. Ante at 19. As they
acknowledge, the place where Godinez entered and exited
the gangway is not the same as the place where the bullet cas‐
ings were recovered. In fact, these two locations are separated
by a 100‐foot distance along an east‐west axis, and the spent
bullets were found directly south of where the casings were
recovered. How can one draw a “straight line” among these
three points? By referring to “the gangway” in generalized
terms, my colleagues gloss over the spatial details of the evi‐
dence to derive a single, fixed point in space at which (1)
Godinez is observed, and (2) bullet casings are found. But it is
precisely the separation of these two points that is at issue.
In fact, nothing directly places Godinez at the place where
the casings were found—not film, not physical evidence, not
testimony. Throughout this time the gangway was shrouded
in darkness, outside the range of the many security cameras
in the area. Nothing, in short, placed Godinez at the shooter’s
location at the critical time. Something more was needed be‐
fore the jury could infer that he was that shooter.
To close the evidentiary gap, the government presented
what my colleagues refer to as Godinez’s “motive” and his
“actions after the shooting.” Hector Ruiz, a gang member
turned informant, testified at trial that the Latin Saints street
gang controlled the Back of the Yards neighborhood where
the shooting took place. Members of the gang, according to
Ruiz, are responsible for keeping rival gang members out of
the neighborhood by “[c]has[ing] them or shoot[ing] at them”
when spotted. Tr. 729. Ruiz described Godinez as a “chief” in
the street gang, subject to the duties of all other members. Tr.
30 No. 19‐3425

723. (Although Ruiz declined to name Godinez a “chief” in
his grand jury testimony and in discussions with federal in‐
vestigators, Tr. 748, 753, I assume that the jury considered
Ruiz’s trial testimony as the truthful account.)
Before the shooting, security footage captured Godinez
slowly driving around the Back of the Yards in his white Kia
SUV. (The majority describes this as “patrolling.”) At one
point before the shooting, Godinez entered and left his home.
(“[P]resumably,” the majority guesses, “to get a gun,” though
there is no evidence either that this was his motive or that he
retrieved any gun while he was there.) Five minutes or so af‐
ter the shooting, Godinez was picked up in front of his house
on S. Wood Street by his then‐girlfriend, Valerie Jean‐Bap‐
tiste. He commented to her that he “[felt] good.” He also said
“fuck” that (or those) “flake[s]”—a term used to refer to rival
gang members, though nothing indicates whether this was
simply a general sentiment or if something particular
prompted it.
Jean‐Baptiste and Godinez then proceeded to a gas station.
En route, Godinez downloaded a smartphone application
that picks up police chatter. At the gas station, Godinez pur‐
chased a white t‐shirt and layered it on top of his black t‐shirt.
Security footage captured Godinez in the gas station around
3:33 am.
Shortly after 4:40 am, Godinez returned to the neighbor‐
hood and got back into his white Kia SUV. Jean‐Baptiste and
he drove to O’Hare International Airport to drop off a friend.
Instead of driving straight back home, however, Godinez
stopped by his cousins’ home (in a different neighborhood)
and traded his white Kia for his cousin’s gray SUV. He re‐
turned to the Back of the Yards that morning, driving the gray
No. 19‐3425 31

SUV and wearing the same clothes he had worn for hours.
Dkt. 99 at 10. Around noon, Godinez picked up the mother of
his child, Destiny Rodriguez, and their child, near W. 43rd St.
and S. Honore St., two blocks away from the shooting.
In the evening of May 4, Godinez returned to his cousin’s
house, this time with Rodriguez and his baby. There, they all
learned that he was wanted by police. Rodriguez began cry‐
ing. Godinez hugged her and apologized for anything he had
done to hurt her. Tr. 685. Godinez then left “one of his
phones” and his white Kia SUV at his cousin’s home, and later
arranged for his surrender. Dkt. 99 at 11; Tr. 708.
In my colleagues’ view, these facts gave the jury enough
to infer that Godinez was the shooter. If he was duty‐bound
to shoot rival gang members, they suggest, then it is conceiv‐
able that he mistook the ATF agents for a rival gang and took
the steps needed to drive them from the neighborhood. When
entering Jean‐Baptiste’s car, his comment “I feel good, fuck
that flake” might have been a way of saying that he was sat‐
isfied with what had just happened nearby (whatever that
was—nothing indicates that the shooting was the only con‐
ceivable event). My colleagues believe this is enough to per‐
mit the jury to infer that Godinez pulled the trigger. Further,
they believe that Godinez’s act of downloading a police scan‐
ner, layering up with a different colored t‐shirt, and trading
his white Kia for a gray SUV permitted the jury to infer that
he was trying to avoid capture. Finally, they reason that
Godinez’s apology to Rodriguez would permit a jury to infer
that he was expressing remorse for shooting a federal agent.
But each of these leaps goes too far.
32 No. 19‐3425

B. Gaps in Evidence
If the only evidence against Godinez were his supposed
motive or actions after the shooting, Godinez would likely
have walked away a free man. Let’s start with motive. Our
court recognizes that when “evidence of gang membership
[is] necessary to explain the motive behind the crime
charged,” United States v. Montgomery, 390 F.3d 1013, 1018 (7th
Cir. 2004), the jury may rely on evidence of gang affiliation to
establish the requisite motive. But the jury still needs a basis
for concluding that the putative gang member actually com‐
mitted the crime. Mere gang membership does not suffice, lest
police receive carte blanche authority to round up every sus‐
pected gang member in hopes that the jury will conclude that
at least one of them must have been near the alleged crime site
and thus was guilty. At best, Ruiz’s testimony characterizing
Godinez as a “chief” in the Latin Saints gang permitted the
jury to draw the inference that Godinez had a reason to shoot
Agent Crump (mistaking him for a rival gang member). But
motive alone provides no basis for inferring that Godinez per‐
formed any particular act.
Another piece of evidence that my colleagues find damn‐
ing is Godinez’s act of driving slowly around the neighbor‐
hood before the shooting. They characterize this as “patrol‐
ling,” again, with no supportive evidence. Even more specu‐
lative is their view of Godinez’s actions after he spotted an
unknown vehicle. He stopped off briefly at his house. From
that simple act, they surmise that he retrieved a gun from his
home. But there was no gun: there was no proof that a gun
was in the house; there was no proof that anyone in the house
had a gun and gave it to Godinez; and there was no proof that
Godinez had a gun after he left the house. I cannot accept the
No. 19‐3425 33

proposition that proof of gun possession is complete simply
by showing that a person entered and left his own home, pe‐
riod. Too many other reasons easily explain that brief stop:
Godinez may have needed to use the restroom, grab a cell‐
phone, or drink some water; the possibilities are endless.
There is no basis for the inference that Godinez obtained a
gun during that stop. Inferences are based on fact, not specu‐
lation, and all we can do with the knowledge that Godinez
enters his home out of sight of security cameras is speculate
on what he did inside.
Five minutes after the shooting, Godinez said “I feel good,
fuck that flake.” Are these the words of a gunman? Perhaps.
But they could also be the words of a gang member who heard
gunfire nearby, walked away from the area unscathed, and
assumed that a rival gang member had been shot. The sup‐
posed “motive” evidence presents, at best, a case just as con‐
sistent with Godinez’s guilt as with his innocence. That’s not
enough.
What about Godinez’s actions after the 3:18 am shooting?
He downloaded a police scanner app, a decision that my col‐
leagues believe support the inference that he “wanted to
avoid questioning or capture.” But how is this any different
from flipping to the local news or turning to Twitter to learn
what had happened nearby? Again, an act that is equally con‐
sistent with criminal culpability (e.g., a desire to stay one step
ahead of the police by tuning into their public broadcasts) and
with a variety of innocent reasons cannot support a convic‐
tion. This is not a question of weight or credibility; it is a ques‐
tion about whether the factual prerequisites are in place to
support the inference that my colleagues believe the jury
drew.
34 No. 19‐3425

Godinez then “changes his appearance,” when he layered
the white t‐shirt on top of his black t‐shirt. They see this as a
sinister move designed to evade capture. In my view, it is too
weak a reed to support the jury’s verdict. What really
changed? Godinez still had the discernable long hair under a
cap, dark colored shorts, and the same black t‐shirt under‐
neath. Particularly in his own neighborhood, which he did
not try to avoid, slipping on a white t‐shirt means little. The
Old Farmer’s Almanac reports that the temperature that night
went down to 58 degrees Fahrenheit, that there was light pre‐
cipitation, and that the wind ranged from 17 mph to gusts of
41 mph. See https://www.almanac.com/weather/history/IL/
Chicago/2018‐05‐04. For all we know, Godinez thought he
would be more comfortable with another layer. While I am no
expert in the art of disguise, I would expect more from a
shooter attempting to avoid detection.
My colleagues then suggest that “[f]rom Godinez later
leaving his car and phone at a cousin’s house, a rational jury
similarly could infer that he wanted to stay away from the
neighborhood.” But that makes no sense. In the first place, af‐
ter trading his white Kia for his cousin’s gray SUV, he drove
right back into the neighborhood wearing the same clothes he
had worn hours before. No rational jury could conclude that
Godinez wanted to do the exact opposite (leave) of what he
actually did (return to the neighborhood several times in the
hours after the shooting). My colleagues then misread the
facts pertaining to Godinez’s phone. Godinez visited his
cousin’s house twice—once to drop off his Kia, and later to
return with Rodriguez. It was after the second visit, during
which he learned that he was wanted by police, that Godinez
“abandoned one of his cellphones” at his cousin’s house. Dkt.
No. 19‐3425 35

99 at 11 (emphasis added). Presumably he still had the other
phone.
My colleagues make another leap: that “from Godinez’s
reaction to Rodriguez’s expression of emotion [after finding
out that he was wanted by police], a rational jury could also
conclude Godinez regretted his actions, including the shoot‐
ing.” But Rodriquez’s testimony does not leave the door open
for this theory. When Godinez hugged Rodriguez, “he was
just saying [to her], ‘I love you’ and was saying that he was
sorry for everything that he’s ever done to [her] and ever hurt
[her].” Tr. 685. Maybe a generalized expression of remorse
(such as “I’m sorry for everything”) might permitted the in‐
ference suggested by my colleagues, but that is not what
Godinez said. Instead, he specifically included the entire
course of their relationship, and in so doing, left no room for
the inference my colleagues believe the jury could draw.
Last, I must address the evidence that might have swayed
the jury to convict Godinez. The government’s key exhibit, la‐
beled GX106, contains the footage from the security camera in
the alley between S. Wood Street and S. Hermitage Avenue.
In this video, Godinez is seen walking westbound from the
Wood Street side of the alley to the Hermitage Avenue side.
The time was 3:16:69 am, minutes before the shooting. (The
still image below is timestamped 3:19:47 am, which would be
after the shooting if the camera’s internal clock were accurate.
At trial, however, the government’s expert clarified that the
alley camera’s internal clock ran 2 minutes and 47 seconds
ahead). After the shooting, at 3:18:27 am, the alley camera cap‐
tures Godinez heading eastbound back towards Wood Street.
(That still image is timestamped at 3:21:15 am).
36 No. 19‐3425

As my colleagues describe it, the image on the right shows
Godinez “running back across the alley with his hand station‐
ary at his right side moving at a faster pace” than usual. Of
course, there is nothing inherently suspicious about a per‐
son’s running away from gunfire; failing to stand in place
when the bullets start flying is not a prerequisite for inno‐
cence. And while holding one’s hand stationary while all else
is in motion might indicate that the person was holding some‐
thing (in this case, the majority speculates, a gun), Godinez’s
arm positioning loses its significance when we consider his
depiction in GX104, which contains footage from a gas station
recorded approximately 15 minutes after Agent Crump was
shot. Godinez is seeing walking in with his right‐hand sta‐
tionary at his side, holding what seems to be a cell phone.
No. 19‐3425 37

Thus, neither Godinez’s speed nor his posture indicates
guilt. Nonetheless, my colleagues are willing to say that “a
rational jury could have found that video places Godinez
where the shots were fired and at the time they were fired.”
But it does no such thing; it does not even place him within a
short distance of the place where the shots were fired. All it
does is indicate that he may have been nearby. Anything fur‐
ther is nothing but speculation.
In GX106, Godinez is seen entering and exiting a gangway
from which, at the gangway’s entrance on Hermitage (100 feet
west from where Godinez is filmed), a person fired several
shots towards federal agents. But walking toward a destina‐
tion is not the same as arriving. Nothing in the government’s
videos depicts Godinez arriving at the gangway’s entrance on
S. Hermitage Avenue. At most, they show Godinez on a tra‐
jectory consistent with arriving at the destination from which
the gunshots were fired. But the remainder of the 100‐foot
journey remains unobserved:
38 No. 19‐3425

Above, I highlight Godinez’s observed crossing in the al‐
ley with a yellow triangle, and I denote the unobserved
mouth of the gangway with a red circle. The only way to close
the gap between these two points is through speculation that
Godinez continued on a straight‐line path. True, “a body in
motion will remain in motion unless it is acted upon by an
external force.” Isaac Newton, Philosophiæ Naturalis Principia
Mathematica (1687). But we can take judicial notice of the fact
that human beings have independent agency. People stop.
People hesitate. People turn around, or change their minds.
And sometimes, people run. There was not enough in the rec‐
ord in this case for the jury to place Godinez at the mouth of
the gangway at the critical moment beyond a reasonable
doubt.
I would accordingly reverse the denial of Godinez’s mo‐
tion for acquittal, and I respectfully dissent from the decision
to affirm.