*230 ¶ 1 Following a 2016 bench trial, defendant, Derrick Macklin, was convicted of *1249 *231 armed robbery involving the personal discharge of a firearm causing great bodily harm ( 720 ILCS 5/18-2(a)(4) (West 2010) ) and sentenced to 40 years' imprisonment. The only argument Macklin raises on appeal is that the State did not sustain its burden to prove him guilty beyond a reasonable doubt because the eyewitness testimony implicating him should not have been believed by the trial judge. Related to this claim, Macklin argues that his trial counsel rendered ineffective assistance by not presenting an expert witness to support his defense that the eyewitness identifications were unreliable. Finding no error, we affirm.

¶ 2 Macklin was charged with six counts of attempt first degree murder, five counts of armed robbery, one count of aggravated battery, one count of armed habitual criminal, two counts of aggravated discharge of a firearm, and four counts of aggravated unlawful restraint, all arising out of an incident occurring on October 2, 2011, and involving two victims, Jose Gomez and Wilfredo Garcia.

¶ 3 At about 10:30 p.m. on October 2, Gomez and his nephew, Garcia, were walking in the vicinity of 156th Street and Central in Harvey. As they were walking, a white car with four passengers approached from behind and stopped next to them. Three men got out of the car and walked toward Gomez and Garcia. The driver remained in the car and eventually drove off without the others.

¶ 4 The area was well lit with streetlamps, including one near Gomez and Garcia and another near the car. All three men wore dark sweatshirts and baseball caps with the hoods pulled up over the caps. Their faces were not covered. The man in the middle of the three men, later identified as Macklin, was taller than the others, and Gomez and Garcia were able to see his face. When he was about 12 feet away from Gomez and Garcia, Macklin pulled out a gun and said, "your money or you die" and fired a single shot toward them. The gunshot struck Garcia in the right hand and he fell to the ground facedown. The other two men took over $ 150 in cash and identification cards from Gomez's pockets, as Macklin pointed the gun at Gomez and Garcia. They also took Garcia's wallet. The three men fled on foot.

¶ 5 After the men left, Gomez found that he still had his cell phone and called 911. Police officers responded to the call. Gomez spoke English in the 911 call and with the officers. The record does not contain the police report or any description of the offenders Gomez gave to police. Garcia cannot speak or read English so he not did speak to the responding officers directly. It is unclear whether he communicated with the officers through Gomez.

¶ 6 Garcia's right hand was bleeding from a through and through gunshot wound, and an ambulance took him to the hospital. He remained in the hospital overnight and was transferred to another hospital the next day. He has permanent scarring from the wound and loss of feeling in one of his fingers.

¶ 7 Macklin was arrested on October 10, 2011, as a result of another incident involving shots fired at another location in Harvey. He was charged and convicted of being an armed habitual criminal in connection with that incident. See People v. Macklin , 2016 IL App (1st) 140697-U .

¶ 8 The day after Macklin's arrest, Gomez and Garcia went to the police station to view a lineup. Detective Andrew Wallace, who speaks a little Spanish but is not fluent, met them at the station. Before the lineup, Gomez signed a lineup advisory form and saw that Garcia signed one as well. Gomez was able to read the English-language advisory form. He explained the *1250 *232 form to Garcia. A Spanish-speaking officer, whose name Wallace could not recall, also explained the form to Garcia (although, according to Garcia, Gomez was not present at the time). Gomez and Garcia were notified that (i) the suspect may or may not be in the lineup, (ii) they were not required to make an identification, and (iii) they should not assume that the person administering the lineup knew which person was the suspect.

¶ 9 Garcia, and then Gomez, separately viewed the lineup and both identified Macklin as the person who had robbed them and shot Garcia. Macklin sat in a different position in each lineup. Garcia was "70 percent sure" of his identification. Garcia did not tell Gomez who to identify before Gomez viewed the lineup. When Gomez viewed the lineup, he identified Macklin. Gomez was "100 percent sure" of his identification. Nobody told Gomez to choose Macklin; instead, he recognized Macklin from the incident. While only Macklin wore braids in the lineup, that did not affect Gomez's identification because it was based on "[h]is eyes and mouth," which Gomez recognized.

¶ 10 Before trial, defense counsel filed a motion to suppress the lineup identifications on the ground that the composition of the lineup was suggestive. In particular, counsel contended that Macklin was the only person in the lineup wearing a white T-shirt and who had braids. No testimony was taken at the hearing and counsel relied only on the lineup photos to argue that they were suggestive. Defense counsel never argued that the offender was described as having braids so there was no issue that Macklin's hairstyle was not suggestive. Focusing on Macklin's claim that his lineup attire was suggestive, the court inquired what information regarding the offenders' attire had been described by the victims. Both defense counsel and the State agreed that the victims had described the men as wearing black hoodies and baseball caps. The court responded, "So in the lineup, there are no black hoodies and no baseball caps." Noting that Macklin's white T-shirt did not make it more likely that he would identified since he was not described as wearing a white T-shirt at the time of the incident, the court found that the lineup was not suggestive and denied the motion to suppress. At no point during the hearing did the State indicate that the only description given by the victims was that the offenders were wearing black hoodies and baseball caps.

¶ 11 The State presented the testimony of Gomez, Garcia, and Wallace at trial. Macklin elected not to testify and did not present any evidence.

¶ 12 Following closing arguments, the court found Macklin guilty of all charges but attempted first degree murder. Since it was undisputed that Garcia had been shot, the court found that the only issue was the reliability of the victims' identification of Macklin. Although both Gomez and Garcia had testified through an interpreter at trial, the court noted that Gomez understood and spoke English. The court found Gomez credible, consistent, and unimpeached. The court noted that Garcia expressly denied being told who to identify. The court stated that while there was evidence that Garcia had been only 70% certain in his identification, "Gomez is 100 percent sure. He never wavered." In other words, Garcia and Gomez "were both sure of one thing, that [Macklin] was the guy that shot" Garcia.

¶ 13 In his posttrial motion, Macklin challenged the sufficiency of the evidence and the denial of his motion to suppress identification testimony. At the motion hearing, defense counsel argued in relevant part that the testimony of all three witnesses was not credible, and in particular *233 *1251 Garcia "had difficulty actually identifying [Macklin] as the shooter" and "said in court in testimony that he was 70 percent sure at first that it was [Macklin] that indeed shot him." The court denied the posttrial motion.

¶ 14 Following a sentencing hearing, the court sentenced Macklin to 40 years' imprisonment for the armed robbery of Garcia involving the personal discharge of a firearm proximately causing great bodily harm. The convictions on all other counts merged into the armed robbery conviction.

¶ 15 As noted, Macklin raises no claim of trial error other than the trial court's decision to credit the eyewitness identifications of him as the assailant.

¶ 16 A person commits armed robbery when he (1) knowingly takes property from another by the use of force, or by threatening imminent use of force and (2) in committing that offense, he personally discharges a firearm proximately causing great bodily harm to another. 720 ILCS 5/18-1(a), 18-2(a)(4) (West 2010). Macklin does not challenge that an armed robbery occurred but argues that the State failed to prove he was the perpetrator.

¶ 17 On a claim of insufficient evidence, we must determine whether, taking the evidence in the light most favorable to the State, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt. People v. Gray , 2017 IL 120958 , ¶ 35, 418 Ill.Dec. 916 , 91 N.E.3d 876 . It is the responsibility of the trier of fact to weigh, resolve conflicts in, and draw reasonable inferences from the testimony and other evidence, and it is better equipped than this court to do so as it heard the evidence. Id. ; In re Jonathon C.B. , 2011 IL 107750 , ¶ 59, 354 Ill.Dec. 484 , 958 N.E.2d 227 . We do not retry a defendant; that is, we do not substitute our judgment for that of the trier of fact on witness credibility or the weight of evidence. Gray , 2017 IL 120958 , ¶ 35, 418 Ill.Dec. 916 , 91 N.E.3d 876 . Contradictory evidence or minor or collateral discrepancies in testimony do not automatically render a witness's testimony incredible, and it is the task of the trier of fact to determine if and when a witness testified truthfully. Id. ¶¶ 36, 47. A conviction will be reversed only where the evidence is so unreasonable, improbable, or unsatisfactory that a reasonable doubt of the defendant's guilt remains. Id. ¶ 35. When a finding of guilt depends on eyewitness testimony, we must decide whether the trier of fact could reasonably accept the testimony as true beyond a reasonable doubt. Id. ¶ 36.

¶ 18 We recite variations of the above well-established precepts in every case involving a claim that the evidence was insufficient to convict. As familiar as they are, they are vital rules of law that govern the respective roles of the trial and appellate courts in all cases. In cases where the State is able to present not only eyewitness testimony, but also physical evidence connecting defendant to a crime such as shell casings matching a weapon recovered in defendant's possession, DNA evidence, or an inculpatory statement, these principles are easy to apply. In contrast, in cases like this where the evidence is sufficient, but not overwhelming, those standards become more difficult to apply and it is tempting to second-guess a trial judge's determination of the sufficiency of the evidence or witness credibility. Under such circumstances, the true measure of a court's fidelity to the rule of law is its acknowledgement of the difficult decision the trial judge was called upon to make, but recognition of the duty the law imposes to afford that decision deference. We do not share our dissenting colleague's "grave *1252 *234 and serious doubt" about the eyewitness identifications in this case.

¶ 19 We also do not "rubber stamp" credibility determinations, and when identifiable factors undermining those determinations exist, it is appropriate to conclude that a court or jury acted unreasonably in accepting a witness's testimony. See People v. Coulson , 13 Ill. 2d 290 , 296, 149 N.E.2d 96 (1958) (rejecting trial court's credibility determination where victim's testimony that men who robbed him accompanied him to his home and permitted him to go inside alone with promises that he would not call the police "taxes the gullibility of the credulous"); People v. Wright , 147 Ill. App. 3d 302 , 321, 100 Ill.Dec. 899 , 497 N.E.2d 1261 (1986) (disbelieving alleged rape victim's testimony that defendant dragged her by her wrist for nearly four hours when there was no evidence of swelling, soreness, bruises, or tenderness to her wrist); People v. Hernandez , 312 Ill. App. 3d 1032 , 1037, 246 Ill.Dec. 65 , 729 N.E.2d 65 (2000) (reversing defendant's conviction based on eyewitness identification by witness who only saw back of shooter's head and a glimpse of the shooter's profile from 90 feet away). But given the foregoing framework that guides every case we consider, the circumstances that lead us to conclude that no reasonable person could have accepted a witness's testimony should naturally be few and far between. This is not one of those cases.

¶ 20 Context is critical when referencing authorities referring to the identification of strangers as " 'proverbially untrustworthy.' " United States v. Wade , 388 U.S. 218 , 228, 87 S.Ct. 1926 , 18 L.Ed.2d 1149 (1967) (quoting Felix Frankfurter, The Case of Sacco and Vanzetti: A Critical Analysis for Lawyers and Laymen 30 (1927) ). Many courts questioning the reliability of eyewitness identifications have done so in cases where the identifications were made under suggestive conditions. For example, State v. Lawson , 352 Or. 724 , 291 P.3d 673 (2012), and State v. Henderson , 208 N.J. 208 , 27 A.3d 872 (2011) are not wholesale indictments of the reliability of eyewitness identifications. Rather, both Lawson and Henderson dealt with suggestive identification and lineup procedures and the effect such suggestiveness has on a witness's identification of the defendant. See Lawson , 291 P.3d at 679-80 (victim was unable to pick defendant out of photo array at hospital shortly after she was shot and again one month later, but despite her lack of recollection of the hospital interview, five weeks after the incident she believed defendant was the assailant; victim later shown newspaper reports of the crime with photographs of the defendant as well as other photographs of the defendant and was taken to court before trial so she could observe him in person); Henderson , 27 A.3d at 879-81 (when interviewed by police immediately after murder, eyewitness told police a false story and later changed his story when confronted by police; in viewing photo array, witness did not initially pick out defendant's photograph, claiming to be unsure, but when shown the photo array again after police told him not to be afraid and that they would take care of him, selected defendant's photo). Here, in contrast and as we discuss below, Macklin's lineup was not suggestive, and he does not pursue his contention on appeal that it was. Leaving aside whether a Spanish-speaking officer translated the advisory form for Garcia (both Garcia and Wallace testified that she did) or whether that officer should have been present in the room when Garcia viewed the lineup, the record does not reveal that suggestive procedures contributed to Garcia's identification of Macklin. And there is absolutely no basis *1253 *235 to contend that there was anything suggestive about Gomez's identification.

¶ 21 As recognized in Lawson , "the scientific research is 'probabilistic'-meaning that it cannot demonstrate that any specific witness is right or wrong, reliable or unreliable, in his or her identification." 291 P.3d at 685 . While we respect our dissenting colleague's views, we do not share his approach to evaluating the eyewitness testimony in this case.

¶ 22 It is well-settled that a valid conviction may be based on a positive identification by a single eyewitness who had ample opportunity to observe. In re M.W. , 232 Ill. 2d 408 , 435, 328 Ill.Dec. 868 , 905 N.E.2d 757 (2009). A trier of fact assesses the reliability of identification testimony in light of all the facts and circumstances including (1) the witness's opportunity to view the offender at the time of the offense, (2) the witness's degree of attention at the time of the offense, (3) the accuracy of any previous description of the offender by the witness, (4) the degree of certainty shown by the witness in identifying the defendant, and (5) the length of time between the offense and the identification. Id. ; People v. Joiner , 2018 IL App (1st) 150343 , ¶ 47, 423 Ill.Dec. 162 , 104 N.E.3d 1251 . These are often referred to as the Biggers factors. Joiner , 2018 IL App (1st) 150343 , ¶ 47, 423 Ill.Dec. 162 , 104 N.E.3d 1251 (citing Neil v. Biggers , 409 U.S. 188 , 93 S.Ct. 375 , 34 L.Ed.2d 401 (1972) ). No single Biggers factor by itself conclusively establishes the reliability of identification testimony; instead, the trier of fact must consider all the factors. Id.

¶ 23 The Biggers factors are applied identically at the trial and appellate levels. Trial judges employ them all the time in cases where the identity of the defendant is at issue. Accordingly, when the trial judge here found the victims' and, in particular, Gomez's positive identification of Macklin sufficient to sustain the State's burden of proof, he was not merely finding that the victims were generally credible; rather, he was finding those identifications reliable under Biggers . We perform the same function on review without, of course, the advantage of being able to see and hear the witnesses.

¶ 24 Here, after "carefully examin[ing] the record evidence" ( People v. Cunningham , 212 Ill. 2d 274 , 280, 288 Ill.Dec. 616 , 818 N.E.2d 304 (2004) ) and taking the evidence in the light most favorable to the State, we conclude that a rational trial judge could convict Macklin of armed robbery based on the eyewitness identifications. In other words, this is simply not a case in which "only one conclusion may reasonably be drawn from the record" that is contrary to the conclusion drawn by the trier of fact. Id. The testimony of both Gomez and Garcia was largely consistent. Although Macklin points to certain discrepancies between their accounts of the robbery (Gomez testified that the two other men took his money and identification, while Garcia said Macklin did; Gomez estimated that he was able to see Macklin for "three to five minutes" while Garcia estimated he saw him for "three seconds" and while he was facedown on the ground), both consistently testified that the area was well lit, they were able to see Macklin, whose face was not covered, as he approached them, and that Macklin shot Garcia in the hand from about 12 feet away. We disagree with Macklin's contention that Gomez and Garcia gave inconsistent accounts of whether Macklin said anything as he approached them. Gomez testified that as he approached, Macklin said, "your money or your life." Garcia, on the other hand, was never asked whether Macklin said anything; he was instead asked on cross-examination whether, as the men approached, they said anything among themselves:

*236 *1254 (1) "And these individuals, the car that you saw, you didn't hear them make any conversation[ ] between them when they got out of the car, is that correct?" and (2)"And when that shot was fired, again there was no conversation between the person that fired that shot and the two individuals on the side?". And in any event, Garcia did not understand English, so his failure to recount anything Macklin said would be unsurprising. Ultimately, the effect of any discrepancies in the victims' accounts of the robbery, which defense counsel explored on cross-examination and stressed in closing, presented a credibility determination for the trial court, which it is inappropriate to second-guess.

¶ 25 Macklin's argument regarding the reliability of the victims' identification of him in the lineups rests primarily on the assumption that the only description of the offenders given to police was that they wore black hoodies and baseball caps. But, as noted, the record does not disclose what description Gomez gave police at the time of the robbery because Macklin has not included the police report in the record. It is axiomatic that it is improper to draw an inference in favor of a defendant based on material missing from the record. See People v. Urdiales , 225 Ill. 2d 354 , 419, 312 Ill.Dec. 876 , 871 N.E.2d 669 (2007) (any doubt that arises from an incomplete record on appeal will be resolved against the appellant). But even though we do not know what contemporaneous description of the offenders Gomez gave to police, Macklin argues and the dissent presumes it consisted only of "black hoodies and baseball caps," a description that "applies to tens of thousands of black men in Chicago on any given day" ( infra ¶ 69). 1 Again, indulging this presumption favoring Macklin runs counter to the principles that govern our review.

¶ 26 Macklin mischaracterizes the suppression hearing when he argues that the State conceded that a generic description of black hoodies and baseball caps was the only thing police had to go on. Rather, as to the clothing the offenders wore, the State merely confirmed the description of the offenders' clothing given by the victims, which, in turn, led to the trial court's finding that the white T-shirt Macklin wore during the lineup was not suggestive. Gomez was never asked at trial what description he gave to police, and the details of Macklin's appearance that Gomez testified he observed at the time of the robbery-his eyes, nose, mouth, and facial hair, consisting of a small beard and moustache-were all brought out on cross-examination.

¶ 27 Macklin contends that Gomez "embellished" his description of the shooter with details he learned by observing Macklin at trial. Macklin did not advance this argument in the trial court, and it is based both on a mischaracterization of Gomez's trial testimony and yet another assumption in Macklin's favor, not borne out by the record, about what Macklin looked like at the time of trial. Gomez testified that when he saw Macklin in the lineup , he recognized the face he saw when Macklin robbed him, including his eyes, nose, mouth, and facial hair. In fact, the lineup photos in the record show that at the time of the lineup, 10 days after the robbery, Macklin had a moustache and small beard. However, the record does not disclose *1255 *237 what Macklin looked like at the time of trial and, in particular, whether he still had facial hair, an identifying feature-along with hairstyle-that a defendant can change before trial in an effort to make an in-court identification less likely. See People v. Irwin , 2017 IL App (1st) 150054 , ¶ 14, 413 Ill.Dec. 574 , 78 N.E.3d 566 (based on defendant's photograph taken at the time of his arrest and the description of him at trial, court concluded that defendant had cut his long hair); People v. Clark , 335 Ill. App. 3d 758 , 767, 269 Ill.Dec. 936 , 781 N.E.2d 1126 (2002) (defendant grew beard before trial; not improper for State in closing to comment on defendant's attempt to change his appearance). So there is plainly no basis for assuming that Gomez "embellished" his description of Macklin by observing him at trial.

¶ 28 Further, even if we assume that the entirety of the description was of three African-American males wearing black hoodies and baseball caps, Gomez's certain identification of Macklin as the offender would be sufficient to sustain the conviction. See People v. Tomei , 2013 IL App (1st) 112632 , ¶¶ 51, 52, 369 Ill.Dec. 209 , 986 N.E.2d 158 (business owner's description of "two white males" wearing a "dark cap" and "heavy jackets" sufficient to sustain conviction given that owner made a positive identification of defendant and testified that he recognized defendant's face). A general or imprecise description by a witness does not necessarily render the witness's identification unreliable. People v. Williams , 2015 IL App (1st) 131103 , ¶ 75, 400 Ill.Dec. 763 , 48 N.E.3d 1203 ; People v. Miller , 254 Ill. App. 3d 997 , 1009, 193 Ill.Dec. 799 , 626 N.E.2d 1350 (1993). "It has consistently been held that a witness is not expected or required to distinguish individual and separate features of a suspect in making an identification. Instead, a witness' positive identification can be sufficient even though the witness gives only a general description based on the total impression the accused's appearance made." People v. Slim , 127 Ill. 2d 302 , 308-09, 130 Ill.Dec. 250 , 537 N.E.2d 317 (1989) ; see also Lawson , 291 P.3d at 687-88 (recognizing, based on eyewitness identification research, that "[c]ontrary to a common misconception, there is little correlation between a witness's ability to describe a person and the witness's ability to later identify that person").

¶ 29 Gomez's trial testimony that he identified Macklin in the lineup with "100 percent" certainty based on his eyes, mouth and facial hair, which he was able to observe during the robbery, is sufficient even if there was nothing particularly distinctive about those facial features that would have prompted him to separately describe them to police. Indeed, if all Gomez and Garcia "really" saw was three men in dark hoodies and baseball caps, it is a remarkable coincidence that they both separately picked Macklin out of lineups in which he sat in different positions. To paraphrase a theme prosecutors often invoke in closing arguments, Macklin must be the unluckiest man on the face of the earth to have not one, but two eyewitnesses mistakenly, separately, and independently identify him as the perpetrator.

¶ 30 Analysis of the Biggers factors does not undermine the reliability of the eyewitness identifications of Macklin. Gomez and Garcia certainly had an adequate opportunity to view Macklin as he approached them as there was sufficient artificial lighting and Macklin's face was not covered. Gomez also had the opportunity to view Macklin during the robbery, which, for purposes of this appeal, we will assume lasted seconds rather than minutes. The dissent dismisses the opportunity Gomez had to observe Macklin, characterizing it as "mere seconds." Infra ¶ 88. But if you *1256 *238 count it out (as in "one, one thousand, two, one thousand, three, one thousand"), it is apparent that the trial judge was entitled to credit that as a sufficient opportunity to observe Macklin. While Garcia testified he fell facedown after he was shot, he also testified he glanced up at Macklin while he was on the ground. Both victims were paying attention to the three men as they approached, and there is no suggestion that they focused only on the gun Macklin pulled out. Because we have no contemporaneous descriptions in the record, we cannot say, as Macklin urges, that a lack of detail undermines their reliability. Gomez was certain in his identification of Macklin; Garcia, less so. And the lineups were promptly conducted 10 days after the robbery. See Slim , 127 Ill. 2d at 313 , 130 Ill.Dec. 250 , 537 N.E.2d 317 (interval of 11 days between robbery and identification of defendant "not significant"); People v. Green , 2017 IL App (1st) 152513 , ¶ 113, 421 Ill.Dec. 474 , 100 N.E.3d 491 ("reviewing courts have found identifications reliable where nearly three months or more elapsed between the crime and the witness's identification").

¶ 31 The dissent dismisses Gomez's certainty in his identification of Macklin as this factor has been "roundly criticized," citing another opinion this writer authored. Infra ¶ 77; see People v. Starks , 2014 IL App (1st) 121169 , ¶ 72, 382 Ill.Dec. 588 , 13 N.E.3d 1 . But, of course, context is everything. Starks involved a jury trial and a defendant's claim that the trial court erroneously refused to permit him to present expert testimony on the issue of the reliability of eyewitness identification. Given the developing body of law regarding the fallibility of eyewitness testimony and the recognition that, in appropriate cases, expert testimony may assist a jury in evaluating such testimony, we found that the trial court abused its discretion in dismissing out of hand defendant's proffered eyewitness expert. Id. We neither directed the trial court to admit expert testimony nor did we reject a witness's expression of certainty as an appropriate factor in the reliability analysis. The cited comments in Starks have no application here in a case involving a bench trial and where no expert witness testimony was proffered.

¶ 32 Moreover, the dissent conflates an eyewitness's degree of certainty at the time of initial identification with certainty at the time of trial, disregarding any distinction. Here, Macklin's trial took place four years after the robbery, so it stands to reason that the victims' in-court identification of Macklin at trial is correspondingly less relevant. But recent research has recognized a distinction between the reliability of lineup and in-court identifications and concluded that expressions of certainty at the time of initial identification are a relevant indicator of accuracy. See John T. Wixted & Gary L. Wells, The Relationship Between Eyewitness Confidence and Identification Accuracy: A New Synthesis , 18 Psychological Sci. in the Pub. Interest 10, 55 (2017) (concluding that "According to the available data, the relationship between confidence and accuracy for an initial ID from an appropriately administered lineup is sufficiently impressive that it calls into question the very notion that eyewitness memory is generally unreliable. * * * [W]hen pristine testing procedures ^[ 2 ^] are *1257 *239 used, an initial ID made with high confidence is highly indicative of accuracy."). Although there were arguable irregularities in connection with Garcia's viewing of the lineup (relating to explanation of the advisory form and the presence of a Spanish-speaking officer in the viewing room), there were none with respect to Gomez. Accordingly, there is no basis to dismiss out of hand his 100% certainty in identifying Macklin. See id. at 13 (noting that most wrongfully convicted defendants exonerated by DNA who were misidentified by an eyewitness were, at the outset of the investigation, identified with low confidence (citing Brandon L. Garrett, Convicting the Innocent: Where Criminal Prosecutions Go Wrong (2011) ) ).

¶ 33 Additionally, as noted, the fact that Gomez and Garcia both identified Macklin as the armed robber, after separately viewing a lineup in which he sat in different positions, enhances and corroborates the accuracy of their respective identifications. Against this evidence, Macklin notes and the dissent finds "troubling" discrepancies in the testimony of Gomez, Garcia, and Wallace regarding the admonishments and advisory forms preceding the lineups. Again, these same arguments were made to Macklin's trial judge. And in light of the unequivocal testimony of Gomez and Garcia that nobody told them who to choose in the lineup, we consider (as did the trial court) the discrepancies to be merely collateral and not fatal to the reliability of their identifications of Macklin.

¶ 34 We acknowledge studies and decisions cited by Macklin that have called into question the reliability of eyewitness identifications. See, e.g. , People v. Lerma , 2016 IL 118496 , ¶ 24, 400 Ill.Dec. 20 , 47 N.E.3d 985 . That said, each case must be judged on its own facts, and nothing in this case compels us to reject the identifications that formed the basis of Macklin's conviction. Stated another way, we cannot view the evidence in the light most favorable to the State, as we must, and conclude that no reasonable trial judge could have believed the eyewitness identifications of Macklin.

¶ 35 Macklin also contends that trial counsel was ineffective for not presenting an expert witness to support his defense that the eyewitness identifications were unreliable.

¶ 36 A defendant's claim that trial counsel failed to render effective assistance is governed by a two-pronged test: the defendant must establish that (1) counsel's performance fell below an objective standard of reasonableness and (2) the defendant was prejudiced by that performance. People v. Brown , 2017 IL 121681 , ¶ 25, 421 Ill.Dec. 924 , 102 N.E.3d 205 . Prejudice is a reasonable probability that the result of the proceeding would have been different absent counsel's error, and a reasonable probability is in turn a probability sufficient to undermine confidence in the outcome of the proceeding. People v. Peterson , 2017 IL 120331 , ¶ 79, 423 Ill.Dec. 776 , 106 N.E.3d 944 . The decision whether to call a particular witness is a matter of trial strategy left to counsel's discretion, and thus generally not a proper basis for an ineffectiveness claim. Id. ¶ 80. A strategy is not unreasonable merely because it proved unsuccessful. Id. ¶ 88. Representation is not constitutionally defective unless the strategy was so unsound that counsel failed to conduct meaningful adversarial testing of the State's case, or so irrational that no reasonably effective counsel in similar circumstances would use that strategy. Id. ¶ 80.

¶ 37 Macklin places great weight on Lerma , in which our supreme court held that the trial court abused its discretion in excluding a defense expert witness on the reliability of eyewitness identification when the only evidence against the defendant *1258 *240 was identification by two eyewitnesses, one of whom did not testify and was not subject to cross-examination. 2016 IL 118496 , ¶ 26, 400 Ill.Dec. 20 , 47 N.E.3d 985 . The Lerma court found that "research concerning eyewitness identification[ ] * * * is well settled, well supported, and in appropriate cases a perfectly proper subject for expert testimony." Id. ¶ 24. It found an abuse of discretion in "the trial court denying defendant's request to present relevant and probative testimony from a qualified expert that speaks directly to the State's only evidence against him, and doing so for reasons that are both expressly contradicted by the expert's report and inconsistent with the actual facts of the case," which rose "to the level of both arbitrary and unreasonable to an unacceptable degree." Id. ¶ 32.

¶ 38 We note first that Lerma had not been decided at the time of Macklin's trial; it was decided shortly after the trial concluded. The Lerma court acknowledged that expert witnesses on the reliability of eyewitness testimony were being routinely excluded at the time, at least partly due to skepticism expressed by the supreme court and repudiated in Lerma itself. Id. ¶ 24 (citing People v. Enis , 139 Ill. 2d 264 , 286-87, 289, 151 Ill.Dec. 493 , 564 N.E.2d 1155 (1990) ). Lerma recognized "the dramatic shift in the legal landscape concerning the use of identification expert testimony." (Emphasis added.) People v. Ortiz , 2017 IL App (1st) 142559 , ¶ 42, 411 Ill.Dec. 542 , 73 N.E.3d 626 . Representation based on the law prevailing at the time of trial is adequate, and counsel is not incompetent for failing to correctly predict that the law will change. People v. English , 2013 IL 112890 , ¶ 34, 369 Ill.Dec. 744 , 987 N.E.2d 371 .

¶ 39 Further, the issue in Lerma -whether the trial court abused its discretion in rejecting proffered expert testimony-is manifestly different than the issue presented here, i.e. , whether defense counsel's performance fell below an objectively reasonable standard based on the failure to call an expert witness at trial. The finding that "research concerning eyewitness identification[ ] * * * is well settled, well supported, and in appropriate cases a perfectly proper subject for expert testimony" ( Lerma , 2016 IL 118496 , ¶ 24, 400 Ill.Dec. 20 , 47 N.E.3d 985 ) does not, standing alone, support the conclusion that trial counsel here was per se ineffective for not presenting such expert testimony or that expert testimony is required in every case. For example, counsel is entitled to consider as a matter of trial strategy that the designation of an eyewitness expert by the defense will likely be met with a counterdesignation by the State, which would highlight and bolster the accuracy of the eyewitness identification. In any event, the argument that trial counsel failed to conduct meaningful adversarial testing of the State's case is refuted by counsel's pretrial motion to suppress identification testimony and extensive cross-examination and argument at trial.

¶ 40 Finally, Lerma involved a jury trial while Macklin elected a bench trial. As the Lerma court stated, "expert testimony is only necessary when the subject is both particularly within the witness's experience and qualifications and beyond that of the average juror's , and when it will aid the jury in reaching its conclusion." (Emphases added.) Id. ¶ 23. We do not find a reasonable probability that the presentation of an expert witness on the reliability of eyewitness testimony in this bench trial would have had any impact on the outcome of the proceedings.

¶ 41 Accordingly, the judgment of the circuit court is affirmed.

¶ 42 Affirmed.

Justices Lavin concurred in the judgment and opinion.

Justice Hyman dissented, with opinion.

¶ 43 JUSTICE HYMAN, dissenting:

*1259 *241 ¶ 44 The question itself is troubling: Did the victims correctly identify Macklin as one of the offenders? The testimony of the victims and the case law raise grave and serious doubt.

¶ 45 " 'What is the worth of identification testimony even when uncontradicted? The identification of strangers is proverbially untrustworthy. The hazards of such testimony are established by a formidable number of instances in the records of English and American trials.' " United States v. Wade , 388 U.S. 218 , 228, 87 S.Ct. 1926 , 18 L.Ed.2d 1149 (1967) (quoting Felix Frankfurter, The Case of Sacco and Vanzetti: A Critical Analysis for Lawyers and Laymen 30 (1927) ). This is one of those instances, and so I respectfully dissent.

¶ 46 Before proceeding, I feel compelled to respond to the majority's characterizations of my views. The majority begins its analysis with a discussion of the standard of review in cases implicating reasonable doubt. The discussion focuses on the appellate court's deference to the trier of fact and its limited role on review of challenges to the sufficiency of the evidence. But, the majority's focus on what appellate justices must not do conveniently disregards discussion of what appellate justices must do.

¶ 47 Especially in criminal cases, each member of the appellate panel has been entrusted with a solemn responsibility-to ensure that justice has been done to the appellant . This includes "carefully examin[ing] the record evidence" to determine whether the State has proven its case. People v. Cunningham , 212 Ill. 2d 274 , 280, 288 Ill.Dec. 616 , 818 N.E.2d 304 (2004). Although we accord the trial court's conclusion about the sufficiency of the evidence great deference, it "is not conclusive and does not bind [us]." Id.

¶ 48 To that end, we cannot unmoor the standard of review from the underlying substantive question-whether the State proved every element of the offense beyond a reasonable doubt. The reasonable doubt standard provides a direct corollary to the presumption of innocence and "plays a vital role in the American scheme of criminal procedure." (Internal quotation marks omitted.) Jackson v. Virginia , 443 U.S. 307 , 315, 99 S.Ct. 2781 , 61 L.Ed.2d 560 (1979). It holds the State to its burden to muster sufficient evidence to allow the fact finder "to reach a subjective state of near certitude of the guilt of the accused." Id. The standard presupposes a rational trier of fact who uses reason to apply the standard to the evidence. Id. at 317 , 99 S.Ct. 2781 . We must similarly view the evidence rationally (see Cunningham , 212 Ill. 2d at 280 , 288 Ill.Dec. 616 , 818 N.E.2d 304 ), and the standard of review neither deprives an appellate court of its ability to reason nor divests it from the use of common sense.

¶ 49 Applying this framework to identification cases, we view the evidence to determine whether the five factors enunciated in Neil v. Biggers , 409 U.S. 188 , 93 S.Ct. 375 , 34 L.Ed.2d 401 (1972), have been met. People v. Slim , 127 Ill. 2d 302 , 307-08, 130 Ill.Dec. 250 , 537 N.E.2d 317 (1989). The primary goal when assessing the Biggers factors involves avoiding " 'a very substantial likelihood of irreparable misidentification.' " Biggers , 409 U.S. at 198 , 93 S.Ct. 375 (quoting Simmons v. United States , 390 U.S. 377 , 384, 88 S.Ct. 967 , 19 L.Ed.2d 1247 (1968) ). In service of that goal, we ultimately consider whether identification is reliable. See In re Christian W. , 2017 IL App (1st) 162897 , ¶ 83, 416 Ill.Dec. 794 , 86 N.E.3d 1079 ("The linchpin of this five-factor test is the reliability *242 *1260 of the eyewitness identification." (Internal quotation marks omitted.) ).

¶ 50 If, based on the record, I concluded that Gomez and Garcia were incredible enough to be unworthy of any belief whatsoever, I would advocate for reversal without analysis of the Biggers factors. See id. ¶ 85 (collecting cases reversing on sufficiency of identification without considering Biggers ). An analysis of the Biggers factors supposes the witness's testimony to be generally credible. See id. ¶¶ 83-85. In other words, a Biggers analysis does not depend on a subjective reevaluation of the trial court's credibility determination; rather, it requires an objective view of the evidence supporting or, here, not supporting, each factor.

¶ 51 Ultimately, the majority misapplies the reasonable doubt standard to identification testimony. It is unquestionably the job of the fact finder to determine the credibility of the witnesses, and we defer to those findings. See, e.g. , People v. Siguenza-Brito , 235 Ill. 2d 213 , 228, 336 Ill.Dec. 223 , 920 N.E.2d 233 (2009) (trier of fact determines credibility of witnesses). Our job entails careful review of the credible evidence to determine whether it establishes proof of every element beyond a reasonable doubt. See Cunningham , 212 Ill. 2d at 280 , 288 Ill.Dec. 616 , 818 N.E.2d 304 (noting, appellate court must "carefully examine the record evidence" and "if only one conclusion may reasonably be drawn from the record, a reviewing court must draw it even if it favors the defendant"). We need not question the veracity or good faith of Gomez and Garcia. The cases and the academic literature, as I will discuss, explain that eyewitnesses, even credible eyewitness, make mistakes. Accepting that Gomez and Garcia testified honestly, their testimony does not prove Macklin's identity as the offender beyond a reasonable doubt.

¶ 52 The majority is concerned that my approach would operate to "second-guess" the trial court's findings of credibility and guilt, in essence disregarding the standard of review. Supra ¶ 18. Not so. Our disagreement can be reduced to a difference in approach. The majority looks to the lineup and trial identifications and assures itself that any infirmities in those identifications have been ameliorated. I look to the fleeting nature of the offense coupled with the inherent distraction caused by being shot and witnessing a relative being shot and find no assurances about the reliability of the identifications. I remain faithful, as I must, to the standard of review and reach a different conclusion-that Macklin's conviction, based exclusively on problematic eyewitness testimony, is unreasonable.

¶ 53 Gomez and Garcia had a few seconds to view the men who robbed them. In that instant, one of the men pulled a gun and shot Garcia. At the end of it all, the only description they could give police was of three black men wearing black hoodies and baseball caps. Given this cursory and generic description, the fleeting nature of the offense, and the inherent distraction caused by the firing of a weapon, one might ask: How then can we rely solely on Gomez's and Garcia's assurance that they were certain, by the time of trial, of Macklin as the shooter? I am unwilling to place substantial reliance on inherently malleable testimony. See People v. Fountain , 2016 IL App (1st) 131474 , ¶ 159, 407 Ill.Dec. 185 , 62 N.E.3d 1107 (Hyman, J., dissenting).

¶ 54 As troubling is Garcia's treatment during the lineup procedure. Garcia spoke "Spanish only" and was unsure whether any Spanish-speaking officers were present at the lineup. He figured out the purpose of the lineup based on piecing together the few words he knew and the "signs"

*1261 *243 that suggested he should identify the shooter. Gomez, also present for the lineup procedure, explained the English-language lineup advisory form to Garcia in Spanish and the purpose of the lineup, which he understood as intending Garcia to see if he recognized the person who committed the crime. The officer responsible for conducting the lineup insisted a Spanish-speaking officer was present but could not recall who. That officer also did not give a Spanish-language advisory form to either Gomez or Garcia, even though one was available.

¶ 55 In light of the questionable conditions surrounding Gomez's and Garcia's initial observation of the offenders, coupled with the problems attendant to the lineup procedures, I cannot agree with the majority's conclusion that the identifications possessed sufficient reliability to amount to proof beyond a reasonable doubt.

¶ 56 Whenever I am called on to review the issue of reasonable doubt, I take heed of the cautionary words of U.S. Supreme Court Justice John Marshall Harlan, II: "I view the requirement of proof beyond a reasonable doubt in a criminal case as bottomed on a fundamental value determination of our society that it is far worse to convict an innocent man than to let a guilty man go free." In re Winship , 397 U.S. 358 , 372, 90 S.Ct. 1068 , 25 L.Ed.2d 368 (1970) (Harlan, J., concurring). This case presents just the sort of situation that Harlan's admonition evokes.

¶ 57 The Biggers Factors

¶ 58 Macklin has not challenged the continued vitality of the Biggers test (see Biggers , 409 U.S. at 199-200 , 93 S.Ct. 375 ; Slim , 127 Ill. 2d at 307-08 , 130 Ill.Dec. 250 , 537 N.E.2d 317 ), but I remain skeptical of many aspects of that analysis. People v. Starks , 2014 IL App (1st) 121169 , ¶¶ 85-93, 382 Ill.Dec. 588 , 13 N.E.3d 1 (Hyman, P.J., specially concurring, joined by Pucinski, J.); see also Fountain , 2016 IL App (1st) 131474 , ¶¶ 151-60, 407 Ill.Dec. 185 , 62 N.E.3d 1107 . The majority suggests that I have gone out of my way to issue "wholesale indictments of the reliability of eyewitness identifications." Supra ¶ 20. The majority overstates my position to criticize it. True, I share the same concerns about some of the Biggers factors that have been recognized by courts in other states ( e.g. , State v. Henderson , 208 N.J. 208 , 27 A.3d 872 , 904-05 (2011) ; State v. Lawson , 352 Or. 724 , 291 P.3d 673 , 695 (2012) ). Nonetheless, Macklin has not questioned Biggers , and the majority frames its analysis by Biggers . So, in addition to explaining my concerns about the continuing utility of some Biggers factors, I also have analyzed them as they currently exist and explained why each exposes a disturbing lack of reliability under the facts before us.

¶ 59 Again, Biggers serves as a litmus test for affirmance, not reversal. Christian W. , 2017 IL App (1st) 162897 , ¶¶ 82-85, 416 Ill.Dec. 794 , 86 N.E.3d 1079 . In other words, to affirm in an identity case, we must be satisfied that the Biggers factors have been met, but we can reverse for "other reasons independent of the five-factor test, such as inconsistencies and discrepancies in the overall testimony" of the eyewitnesses. Id. ¶¶ 84-85.

¶ 60 Opportunity to Observe

¶ 61 Illinois courts repeatedly describe a victim's opportunity to observe an offender as "the most important" of the Biggers factors. E.g. , People v. Wehrwein , 190 Ill. App. 3d 35 , 39, 137 Ill.Dec. 244 , 545 N.E.2d 1005 (1989) ; People v. Moore , 115 Ill. App. 3d 266 , 270, 71 Ill.Dec. 167 , 450 N.E.2d 855 (1983). Reliability or unreliability hinges initially on witness's proximity to the perpetrator and the length and conditions for sound observation. See People v. Tomei , 2013 IL App (1st) 112632 , ¶ 40, 369 Ill.Dec. 209 , 986 N.E.2d 158 .

*1262 *244 ¶ 62 Garcia and Gomez disagree on whether the offense lasted three seconds (Garcia) or three to five minutes (Gomez). The majority assumes the encounter lasted only seconds. In so doing, the majority tacitly calls into question the reliability of Garcia's and Gomez's accounts. This court has previously remanded for the trial court to consider the presentation of expert identification testimony where, under stressful conditions, a witness had no more than a few seconds to observe an offender. Starks , 2014 IL App (1st) 121169 , ¶ 72, 382 Ill.Dec. 588 , 13 N.E.3d 1 (opinion of Mason, J.). Confronted with a similar scenario, the majority concludes that Garcia and Gomez had a sufficient opportunity to observe their attackers. The majority also tries to convey the length of several seconds by asking the reader to count out the seconds ("one, one thousand, two, one thousand, three, one thousand" ( supra ¶ 30) ) as if the dramatic surrounding circumstances did not exist. It is not only that Gomez and Garcia had a fleeting view of the offenders. Their attention was pulled in several directions-they watched the car the offenders got out of drive away, there were three attackers, one brandished a gun, and Garcia was shot. We are not confronted with a scenario where Gomez and Garcia had "three one thousands" to stare at the shooter uninterrupted and distraction and stress free.

¶ 63 Given the brevity of the encounter and that Macklin's conviction hinges entirely on the identification, this factor exposes a lack of reliability. But see People v. Herrett , 137 Ill. 2d 195 , 204-06, 148 Ill.Dec. 695 , 561 N.E.2d 1 (1990) (affirming where witness only had "several seconds" to observe attacker but other circumstantial evidence, including proceeds of robbery, confirmed defendant's identity). Gomez's and Garcia's opportunities to observe were insufficient to support Macklin's identification as the shooter.

¶ 64 Degree of Attention

¶ 65 Not only did Gomez and Garcia have a fleeting view of their attackers, one of the attackers brandished and fired a gun. We do not consider whether they focused only on the gun; we consider whether the overall stress of the crime "contribute[d] to the unreliability of [their] testimony." In re J.J. , 2016 IL App (1st) 160379 , ¶ 30, 407 Ill.Dec. 151 , 62 N.E.3d 1073 (citing People v. Lerma , 2016 IL 118496 , ¶ 26, 400 Ill.Dec. 20 , 47 N.E.3d 985 ). The presence of a gun tends to "focus [a victim] on weapons rather than the offender's face." Id. (citing People v. Allen , 376 Ill. App. 3d 511 , 525, 314 Ill.Dec. 934 , 875 N.E.2d 1221 (2007) ). The phenomenon of "weapon focus" has been amply documented. E.g. , Henderson , 27 A.3d at 904-05 . When a weapon is absent, identification becomes significantly more reliable. Id. at 905.

¶ 66 In J.J. , the victim was still able to be "detailed and descriptive," and video of the incident corroborated many of the ancillary details she remembered, minimizing the effect of the gun's presence. 2016 IL App (1st) 160379 , ¶ 30, 407 Ill.Dec. 151 , 62 N.E.3d 1073 . But, Garcia's and Gomez's fleeting visual views yielded not a single pertinent facial or physical or distinct indicator of identification other than race, clothing, and a differential in height of one of the offenders. See Henderson , 27 A.3d at 905 (brief interaction exacerbates weapon focus). Additionally, we are not confronted with a situation involving the mere presence of a gun; one of the offenders fired the gun and injured Garcia. This becomes particularly relevant because, after being shot, Garcia fell to the ground facedown, abbreviating his ability to observe the offenders.

¶ 67 In an encounter lasting just seconds, during which witnesses are fired at *1263 *245 by a gun-wielding assailant, the likelihood of a reliable identification drastically diminishes. The vagueness of Gomez's and Garcia's initial descriptions of the offenders, at the most crucial point of observation for purposes of identification, confirms their inability to focus on details and undermines the evidentiary value of all their later identifications of Macklin. See Nat'l Research Council, Identifying the Culprit: Assessing Eyewitness Identification 55 (2014), https://nap.edu/read/18891 [https://perma.cc/DA9N-HZA3] ("weapon focus is a real-world case in point for eyewitness identification, in which attention is compellingly drawn to emotionally laden stimuli, such as a gun or a knife, at the expense of acquiring greater visual information about the face of the perpetrator").

¶ 68 Prior Description

¶ 69 A witness's prior description of an offender, even if brief, must be specific. See J.J. , 2016 IL App (1st) 160379 , ¶ 32, 407 Ill.Dec. 151 , 62 N.E.3d 1073 . Here, that description offers nothing more than the presence of three black men wearing black hoodies and baseball caps, one taller than the other two. This is an exceedingly nonspecific description that would not even support reasonable suspicion to arrest Macklin. See Florida v. J.L. , 529 U.S. 266 , 268, 272, 120 S.Ct. 1375 , 146 L.Ed.2d 254 (2000). Descriptions at this level of generality weigh against a finding of reliability. People v. Simmons , 2016 IL App (1st) 131300 , ¶ 95, 408 Ill.Dec. 568 , 66 N.E.3d 360 . The description easily applies to tens of thousands of black men in Chicago on any given day.

¶ 70 The majority takes issue with the sufficiency of the record regarding the initial description of the offenders and yet, at least two times, accepts that Garcia and Gomez first described their attackers only as three men in black hoodies and baseball caps. In the statement of facts ( supra ¶ 4), the majority's description of the offense includes: "All three men wore dark sweatshirts and baseball caps with the hoods pulled up over the caps." In its description of the suppression hearing ( supra ¶ 10), the majority acknowledges that the parties agreed " the victims had described the men as wearing black hoodies and baseball caps." (Emphasis added.) The majority faults Macklin for failing to include the police report in the record and criticizes me of indulging a presumption in Macklin's favor based on evidence that is not before us. Supra ¶ 25. Inexplicitly, the majority discounts the affirmative representations made by an assistant state's attorney that the only description given to police was of three men in black hoodies and baseball caps. That representation most certainly is in the record, and the majority offers no reason that we cannot rely on it. I would find the record supports concluding that Gomez and Garcia gave only a generic description to the police.

¶ 71 In a puzzling detour, the majority drops a footnote and takes "judicial notice" of a decision in a different case to explain that Macklin was not arrested because he matched the description given to the police by Gomez and Garcia. Supra ¶ 25 n.1 (citing People v. Macklin , 2016 IL App (1st) 140697-U , ¶ 27). Never mind that the circumstances of Macklin's arrest are nowhere in the record before us. That aside, the majority's reference to another case involving Macklin is at best irrelevant and at worst a sidelong attempt to portray Macklin as an unsympathetic criminal.

¶ 72 The majority's citation to Macklin's other case is irrelevant primarily because the issue before us is not whether the officers had received a sufficient identification to locate and arrest him or whether they arrested him based on the descriptions given. At this juncture, we are only concerned with Gomez's and Garcia's descriptions *1264 *246 to the extent they give us a basis to evaluate their ability to observe and compare their initial descriptions with their later ones. See J.J. , 2016 IL App (1st) 160379 , ¶ 32, 407 Ill.Dec. 151 , 62 N.E.3d 1073 (noting, prior description increases reliability where it is specific). I simply cannot see the point in citing Macklin's other case when the reason for his arrest has no bearing on the reliability of Gomez's and Garcia's identifications.

¶ 73 Even under the majority's view of the record this factor does not favor the State. See People v. Green , 2017 IL App (1st) 152513 , ¶ 111, 421 Ill.Dec. 474 , 100 N.E.3d 491 (third factor does not apply at all when no prior description given). In other words, accepting the majority's assertion that there is nothing in the record providing Gomez's and Garcia's prior descriptions (there is), this factor would drop out of the analysis entirely.

¶ 74 I also am troubled because, contrary to the majority, I find well supported Macklin's argument that Gomez "embellished" his description after later observations of him. Even if unintentional, Gomez augmented his description at trial from a man in a black hoodie and baseball cap to include the offender's eyes, nose, mouth, and small beard and moustache. We should be particularly suspicious of descriptions that improve with time, as memory rarely works that way. See Henderson , 27 A.3d at 907 . The majority attempts to draw a distinction because Gomez's additional descriptors of Macklin arose at the time of the lineup as opposed to trial. Supra ¶ 27. The majority points out that it is quite easy for defendants to change their appearance at trial in an attempt to frustrate an in-court identification. That makes no difference here, where nothing in the record suggests Macklin's appearance had changed. Gomez's description spectacularly advanced in detail and became sharper and more vivid at trial; whether the specifics came from observing Macklin during the lineup or in the courtroom, Gomez's description somehow blossomed when compared to what it was at the time of the incident.

¶ 75 In all, the generic description combined with its inexplicable and expansive improvements weighs heavily in finding Gomez's and Garcia's identifications as unreliable.

¶ 76 Level of Certainty

¶ 77 The reliability of a witness's certainty about his or her identification has been roundly criticized in this court and elsewhere. E.g. , Starks , 2014 IL App (1st) 121169 , ¶ 72, 382 Ill.Dec. 588 , 13 N.E.3d 1 (describing "the weak correlation between a witness's confidence in his or her identification and its accuracy"); id. ¶¶ 85-89 (Hyman, P.J., specially concurring, joined by Pucinski, J.) (collecting cases). As I have observed, the most confident witnesses are still wrong 20% to 30% of the time. Fountain , 2016 IL App (1st) 131474 , ¶ 159, 407 Ill.Dec. 185 , 62 N.E.3d 1107 . Witness confidence is influenced both by overestimation of the amount of time that a witness observed an offender and by inadvertent feedback from officers during identification procedures. Id. Many jurisdictions that have reconsidered this aspect of their Biggers factors equivalents have drastically altered the way they do so. E.g. , Henderson , 27 A.3d at 920-22 ; Lawson , 291 P.3d at 695 .

¶ 78 The majority seeks to diminish the persuasive value of cases like Henderson and Lawson because those courts analyzed suggestive identification and lineup procedures as opposed to the sufficiency of the evidence. Supra ¶¶ 20-21. The problem for the majority is that so did Biggers , the case from which we derive the test. See Biggers , 409 U.S. at 190 , 93 S.Ct. 375 (noting, certiorari was granted to determine "whether the identification procedure *1265 *247 violated due process"). Without question, we apply Biggers to sufficiency claims in Illinois ( Slim , 127 Ill. 2d at 307-08 , 130 Ill.Dec. 250 , 537 N.E.2d 317 ), and there is no principled reason that a critique of Biggers in one context would not equally apply to all others.

¶ 79 Any distinction between this case and cases like Lawson and Henderson , to which the majority refers, does not affect my argument. As the court in Lawson explained, the scientific literature divides the factors used to determine identification reliability into two groups: system variables and estimator variables. 291 P.3d at 700 . The first group, system variables, "refer[s] to the circumstances of the identification procedure itself that generally are within the control of those administering the procedure." Id. They include factors like blind administration of the procedure, preidentification instructions, lineup construction, simultaneous versus sequential lineups, showups, multiple viewings, suggestive questioning, cowitness contamination, and suggestive feedback. Id. at 705-11. It is self-explanatory that these factors would only apply to a claim of a suggestive lineup procedure, and it is telling that Biggers itself does not employ them.

¶ 80 But, the second set of factors-the estimator variables-are concerned with "characteristics of the witness, the perpetrator, and the environmental conditions of the event that cannot be manipulated or adjusted by state actors ." (Emphasis added.) Id. at 700. They include stress, witness attention, duration of exposure, environmental viewing conditions, witness characteristics and condition, description, perpetrator characteristics, speed of the identification, level of certainty, and memory decay. Id. at 700-05. As the court in Lawson took care to clarify, these factors do not depend on any claim that law enforcement used suggestive identification procedures because these factors are immune from manipulation. Notably, the five factors used in the Biggers decision come from this subset of factors described in Lawson . Again, I do not see how a critique of this set of factors would not be applicable to all contexts in which a defendant challenges his or her identification because these factors apply regardless of the behavior of law enforcement personnel.

¶ 81 For that same reason, the majority's attempt to distance itself from comments made in Starks is unpersuasive. The majority correctly notes that Starks was decided in a different context than a reasonable doubt challenge. But, the conclusion in Starks -that it was an abuse of discretion to disallow expert testimony on eyewitnesses-was supported by the court's conclusion that jurors may not understand that a witness's certainty cannot be trusted. 2014 IL App (1st) 121169 , ¶ 72, 382 Ill.Dec. 588 , 13 N.E.3d 1 . The underlying premise is the same in either context. The entire point of presenting expert testimony on the science related to eyewitness fallibility is to aid the jury in their consideration of a defendant's guilt beyond a reasonable doubt; if jurors are entitled to information about the scientific criticism of the Biggers factors, we can rely on the same criticism in our consideration of a reasonable doubt challenge. Indeed, the case relied on by the majority in Starks for the proposition, State v. Guilbert , 306 Conn. 218 , 49 A.3d 705 , 721 (2012), itself cites Henderson, 27 A.3d 872 , as the leading case in this area. Starks , 2014 IL App (1st) 121169 , ¶ 72, 382 Ill.Dec. 588 , 13 N.E.3d 1 .

¶ 82 Accepting the factor of witness certainty as it is, however, still provides no assurance of reliability. Contrary to the majority's assertion that I have simply dismissed Gomez's certainty "out of hand" ( supra ¶ 31), I find it alarming and illuminating *248 *1266 that Gomez said he was 100% sure of his lineup identification and at the same time also vastly overestimated the amount of time he had to observe the offenders. The majority picks and chooses what it accepts.

¶ 83 Besides, Garcia was 70% sure of his identification immediately after the lineup and then it suddenly jumped to 100% sure later. The circumstances that surround Garcia's increased assuredness greatly heighten doubt and confusion. Problematic conduct during lineups should give anyone pause before accepting an identification as reliable. See Fountain , 2016 IL App (1st) 131474 ¶¶ 162-65, 407 Ill.Dec. 185 , 62 N.E.3d 1107 (Hyman, J. dissenting); see also People v. Delamota , 18 N.Y.3d 107 , 936 N.Y.S.2d 614 , 960 N.E.2d 383 , 391 (2011) (expressing skepticism about lineup procedures because, among other concerns, family member with prior exposure to perpetrator had to translate for witness).

¶ 84 Garcia spoke "Spanish only." Detective Wallace, who spoke "very limited" Spanish, knew that both Gomez and Garcia spoke little English. There was a Spanish-speaking officer available to conduct the lineup, but Wallace could not recall his or her name. Garcia could not remember if there was a Spanish-speaking officer and Gomez testified that there was none. Instead, Gomez, who could read English sufficiently, had to translate the lineup advisory form into Spanish and read it to Garcia. On cross-examination, Garcia could not remember when or if he signed the form. Spanish-language lineup forms were available, but Wallace did not use them. Equally distressing, Gomez had to explain the purpose of the lineup procedure in Spanish so that Garcia could understand that he was there to identify the person who committed the crime. With this understanding of the lineup procedure, Garcia's shifting level of certainty casts the reliability of his identification into greater-not less-doubt.

¶ 85 As a final point, the majority finds it "a remarkable coincidence that [Gomez and Garcia] both separately picked Macklin out of lineups in which he sat in different positions." Supra ¶ 29. Remarkable? The phenomenon of exonerations based on faulty identifications, even where multiple witnesses identified the same person or the evidence was otherwise thought to be overwhelming, has been well documented. E.g. , District Attorney's Office for the Third Judicial District v. Osborne , 557 U.S. 52 , 98-99, 129 S.Ct. 2308 , 174 L.Ed.2d 38 (2009) (Stevens, J., dissenting, joined by Ginsburg and Breyer, JJ.) ("DNA evidence has led to an extraordinary series of exonerations, not only in cases where the trial evidence was weak, but also in cases where the convicted parties confessed their guilt and where the trial evidence against them appeared overwhelming."); Kansas v. Marsh , 548 U.S. 163 , 208-10, 126 S.Ct. 2516 , 165 L.Ed.2d 429 (2006) (Souter, J., dissenting, joined by Stevens, Ginsburg, and Breyer, JJ.) (citing examples and noting, "[m]ost of these wrongful convictions and sentences resulted from eyewitness misidentification, false confessions, and (most frequently), perjury"). Even the briefest perusal of cases at the Innocence Project reveals that what the majority describes as "remarkable" is eminently possible. E.g. , Stephan Cowans , The Innocence Project, https://www.innocenceproject.org/cases/stephan-cowans/ (last visited Feb. 19, 2019) [https://perma.cc/5VDD-TWPU]; Cody Davis , The Innocence Project, https://www.innocenceproject.org/cases/cody-davis/ (last visited Feb. 19, 2019) [https://perma.cc/R79D-DZQM]. What is truly remarkable is reliance on highly suspicious witness certainty coupled with the victims' patently insufficient opportunity to view the offenders.

*1267 *249 ¶ 86 Time Between Offense and Identification

¶ 87 The only factor that arguably weighs in favor of reliability is the time between the offense and the lineup identifications. Gomez and Garcia viewed lineups 10 days after the offense, a time frame on par with others endorsed by Illinois courts. See Simmons , 2016 IL App (1st) 131300 , ¶ 97, 408 Ill.Dec. 568 , 66 N.E.3d 360 (approving time gap of up to two weeks) (citing People v. Williams , 221 Ill. App. 3d 1061 , 1068, 164 Ill.Dec. 411 , 582 N.E.2d 1355 (1991) (approving time gap of 10 days to two weeks) ).

¶ 88 Despite Illinois cases suggesting otherwise, the reality is that an interval of 10 days before a lineup can alter and impair a person's memory. See Lawson , 291 P.3d at 705 ("An aspect of memory decay that is less well known, however, is that decay rates are exponential rather than linear, with the greatest proportion of memory loss occurring shortly after an initial observation, then leveling off over time."). It is not likely that Gomez's and Garcia's original description of three men in black hoodies and baseball caps would dramatically improve with the passage of time, whether a few days or 10 days. That the identification did dramatically intensify indicates an extreme risk of misidentification. See id. Accordingly, with mere seconds to view the offenders under stressful conditions, we have firm reason to doubt the retained accuracy of a 10-day-old identification. See Nat'l Research Council, supra ¶ 67, at 99 (duration of exposure and degree of attention may enhance effect of interval between observation and identification).

¶ 89 Conclusion

¶ 90 In the eyes of the majority, "the true measure of a court's fidelity to the rule of law is its acknowledgment of the difficult decision the trial judge was called upon to make, but recognition of the duty the law imposes to afford that decision deference." Supra ¶ 18. In my eyes, the true measure of our fidelity to the rule of law is a court's willingness to consistently apply the most fundamental constitutional precept in a criminal case: a conviction is not to be had except on proof beyond a reasonable doubt. We must never erode that most crucial of procedural protections in service to a deferential standard of review.

¶ 91 We need a continued conversation, from the bench and elsewhere, to explore the utility of the Biggers factors as our understanding about eyewitness identifications continues to evolve. But, I must emphasize again that my disagreement with the majority is primarily connected to the facts here, not academic disagreements about Biggers . A reviewing court may use common sense and engage in a searching evaluation of the record. Cunningham , 212 Ill. 2d at 280 , 288 Ill.Dec. 616 , 818 N.E.2d 304 ("The reviewing court must carefully examine the record evidence while bearing in mind that it was the fact finder that saw and heard the witnesses."). The two witnesses viewed their attackers for mere seconds: traumatic and tense seconds. One of the offenders pulled a gun and shot one of the victims. Immediately after the offense, the only description the two victims could offer was of three men in black hoodies and baseball caps. Ten days later, relying on fleeting glimpses of the offenders, the victims identified Macklin in a lineup muddled by cross language confusion.

¶ 92 The court must consider all of the Biggers factors together along with the surrounding circumstances. Simmons , 2016 IL App (1st) 131300 , ¶ 89, 408 Ill.Dec. 568 , 66 N.E.3d 360 . Not one Biggers factor weighs in favor of finding Gomez's and Garcia's identifications reliable.

*1268 *250 ¶ 93 I conclude that a reasonable fact finder could not find Macklin guilty based on the identification before us and would reverse.