In the
United States Court of Appeals For the Seventh Circuit ____________________ No. 20-1432 JUAN REYES, Petitioner-Appellant, v.
MINDI NURSE, Respondent-Appellee. ____________________
Appeal from the United States District Court for the Central District of Illinois. No. 16-cv-2346 — Colin S. Bruce, Judge. ____________________
ARGUED OCTOBER 25, 2021 — DECIDED JUNE 29, 2022 ____________________
Before EASTERBROOK, ROVNER, and WOOD, Circuit Judges. WOOD, Circuit Judge. In 2004, six men decided to rob a ma- rijuana dealer, William Thomas, at gunpoint in his home. Two of the robbers shot Thomas, who died. One of the two shoot- ers also shot Timothy Landon, Thomas’s business partner and guest, but Landon survived. In 2007, an Illinois jury convicted Juan Reyes of Thomas’s murder, Landon’s attempted murder, and home invasion. On the murder and attempted murder counts, the state’s evidence against Reyes included Landon’s 2 No. 20-1432
identification of Reyes as the shooter after viewing a photo array. But that identification was far from ironclad. It took the police five attempts to extract it from Landon, and on several occasions, he seemed to confuse Reyes with another man who was not a suspect in the robbery. Reyes moved, unsuccess- fully, to suppress the identification. After Reyes exhausted state-court review of his convic- tion, he moved for federal collateral relief pursuant to 28 U.S.C. § 2254. 1 As he had done in state court, he argued that the identification procedure was impermissibly suggestive and that Landon’s identification was too unreliable to pass constitutional muster. The district court denied his petition, and Reyes appealed. We agree with Reyes that the identifica- tion procedure the state employed was unnecessarily sugges- tive; the state court also expressed concern on this point. But in the end that court found that these flaws did not taint the conviction. Moreover, error alone is not enough to entitle Reyes to relief. A section 2254 petitioner must also show prej- udice. Reyes cannot, because the jury that convicted him heard significant evidence of his guilt beyond the identifica- tion and, critically, had the opportunity to evaluate most of the evidence bearing on the reliability of the identification. We affirm the district court’s judgment.
1 Reyes is incarcerated at Pontiac Correction Center, and so the proper
respondent to his section 2254 application is the current warden of that facility. See Bridges v. Chambers, 425 F.3d 1048, 1049 (7th Cir. 2005). We have substituted Mindi Nurse, presently the Acting Warden of Pontiac, for Leonta Jackson, Pontiac’s warden at the time the appeal was briefed and argued. See Fed. R. Civ. P. 25(d). No. 20-1432 3
I We draw the details of the events leading to Reyes’s con- viction from the trial transcripts and the Illinois Fourth Ap- pellate District’s order affirming the conviction. See People v. Reyes, No. 4-07-0412 (Ill. App. Ct. Oct 7, 2008). A We begin with the robbery. In January 2004, Troy Hutchins, a marijuana dealer, learned that his supplier, Wil- liam Thomas, had a large supply of drugs and at least $40,000 in cash on hand at his home in Danville, Illinois. Hutchins and Thomas had quarreled after Hutchins gave Thomas counter- feit money for drugs. On January 28, Hutchins told another Danville drug dealer, Kenneth Wright, about the cash and drugs in Thomas’s home. At least four other men were visit- ing Wright that day: Reyes, Alex Garcia, Joseph Hernandez, and Andre Smith. At some point, someone—possibly Hutchins—proposed robbing Thomas. All six men agreed to participate and split the proceeds. Thomas’s friend Timothy Landon—like Hutchins, a mari- juana dealer who relied on Thomas for his supply—was vis- iting Thomas’s home that same afternoon. Thomas’s two daughters, Emily (nine years old at the time) and Alyssa (four years old at the time), were also in the house. 2 Thomas and Landon spent the afternoon smoking marijuana, drinking beer, and playing video games. Around dinner time, they
2 Thomas was not Emily or Alyssa’s biological father and, so far as we
can tell from the record, never formally adopted them. Thomas nonethe- less referred to the children as his daughters and they referred to him as their father. We do the same. 4 No. 20-1432
went out with the girls to get food, but came home a short time later. Around 9 p.m., the robbers arrived at Thomas’s house. They had driven over in two vehicles. Reyes, Garcia, Hernan- dez, and Smith came in a maroon van, while Wright and Hutchins took a blue (or possibly white) Cadillac. The plan was for the four men in the van to carry out the robbery, be- cause Hutchins and Wright knew Thomas and believed that he would recognize them. Events from this point on are disputed. We first recount the agreed details, and then we revisit the remainder as we describe the testimony at trial. After parking the van near Thomas’s home, two or three of the four men approached the enclosed porch. At least one of them then entered the house through the front door. Inside, Landon was seated on the couch, while Thomas was at a table near the television rolling a joint. Emily was also in the living room, doing homework; Alyssa was taking a bath. The first robber through the door stood inside for a few seconds. Whether, and how soon after, a second robber entered the home is disputed. Putting that point to one side, however, it is clear that the first robber soon pulled out a gun. According to one account, he immediately shot Landon once in the stomach; according to another, either he or a second gunman shot Landon in the stomach a bit later. Thomas jumped over the table and began to struggle with the gunman, while Emily fled down the hallway to find her sister. Landon also fled but glanced back at one point and saw Thomas still struggling with the shooter. As Landon was leav- ing through the back door, he heard the front door open again, and then heard two more gunshots. No. 20-1432 5
During the struggle, Thomas was hit seven times by bul- lets from two different guns. He wound up outside, where he collapsed in the driveway before dying from his wounds. Landon escaped to a neighbor’s porch. He survived but had to undergo surgery and spent over two weeks in the hospital recovering. Both of Thomas’s daughters also survived. The robbers fled after the shooting without taking anything. B Police undertook a protracted investigation. On the night of the shooting, they collected physical evidence from the scene, including blood, bullets, spent casings, and clothing. They also took a statement from Emily. She described only one intruder, whom she characterized as “large,” “fat,” possi- bly Black, and wearing a hat or mask. They were not able to interview Landon right away—he had been sent straight into surgery—but police took a statement from him a few days later. In that statement, he described two intruders, one of whom he characterized as Black with a light complexion and the other as large and probably also Black. Landon was likely quite impaired by medication and the after-effects of surgery when he gave that statement; he later testified that he did not recall speaking to the police at all. Investigators also inter- viewed several neighbors and Thomas’s girlfriend (Emily and Alyssa’s biological mother). About a week after the shooting, Keith Garrett, one of the investigating officers, paid a follow-up visit to Landon in the hospital. Landon was able to self-administer morphine at the time, but the record does not establish how muddled or clear- headed he was. Garrett showed Landon four 3-by-2 photo ar- rays, one of which was built around Reyes. Landon did not identify Reyes or anyone else as the man who had shot him. 6 No. 20-1432
Around ten days later, shortly after Landon was discharged but while he was still medicated, Garrett visited him and showed him the same arrays. Again, Landon did not identify a shooter. Another ten days later, on February 23, Garrett showed Landon two new 3-by-2 arrays. One of the two included a more recent photo of Reyes. Landon again failed to identify anyone in the arrays as the shooter. He did, however, com- ment on a photo of a man named Peaslee (who was not a sus- pect) that he wasn’t “the guy” but was “close.” Peaslee’s photo was in the same array as Reyes’s photo, directly below it. A week after that, Garrett showed Landon five new arrays, none of which included Reyes. For the fourth time, Landon made no identification. Nearly five months later, and six months after the robbery and shooting, Garrett made a fifth attempt to procure an iden- tification from Landon. This time, he showed Landon a single 3-by-2 array that included the older photo of Reyes, as well as photos of Hernandez and Garcia. Peaslee’s photo was not in- cluded. It seems Reyes’s inclusion in the array was a mis- take—Garrett had meant to build it around Hernandez. But for the first time, Landon identified Reyes as the shooter. That same day, Landon showed Emily—the only other surviving eyewitness to the events inside the house—an array including the same picture of Reyes, but she did not make an identifica- tion. In August, the state presented its evidence to a grand jury. It also secured arrest warrants for Reyes, Smith, and the other conspirators. Smith was arrested soon after, and in January 2006 he pleaded guilty to one count of murder, for which he received a 20-year sentence. Reyes was not located until No. 20-1432 7
February 2006, when he was arrested in Detroit. He was charged with Thomas’s murder, Landon’s attempted murder, home invasion, and several other offenses related to the rob- bery. C Before trial, Reyes moved to suppress Landon’s identifica- tion. Landon testified at the suppression hearing. His cross- examination by defense counsel resulted in several confusing, still-disputed exchanges about Reyes’s and Peaslee’s photos. Counsel asked Landon several times whether Landon recog- nized Reyes in his photo. Sometimes Landon seemed to re- spond that the man in the photo was “not the individual [he] picked out”; sometimes he said that the picture “look[ed] like” the guy; sometimes he responded “yeah” to counsel’s question whether he said the photo “looks like him but it’s not him” (referring to the Peaslee picture). Eventually he identi- fied the “gentleman in the lower left-hand-corner” (i.e. Reyes) as the shooter. It is fair to say that the entire exchange was a muddle. Reyes’s counsel argued that Landon’s identification had been a product of suggestive procedures and was unreliable, and he urged that the hearing colloquy illustrated the point. But the court disagreed and denied the motion to suppress. A new judge took over the case prior to trial and reaffirmed the first judge’s decision to deny suppression. D The case went to trial in January 2007. The prosecution called 27 witnesses; Reyes called just one. We summarize only the testimony relevant to the issues before us. 8 No. 20-1432
Testifying for the prosecution, Emily—12 years old by the time of the trial—recounted that just before the shooting she had been in the living room doing homework. A man came in, whom she described as masked and wearing “all black.” Her father initially mistook the man for a friend of his named Glenn, but then realized he was an intruder. They then began fighting, and Emily ran into her bedroom. (Notably, Emily did not testify that the intruder shot Landon in the first few seconds after he entered the home.) She heard a single gun- shot while hiding under her bed. After she heard the shot, she remembered that Alyssa was in the bath and went to find her. While Emily was getting Alyssa, she heard a few more shots. She then hid her younger sister in a bedroom and went out into the hall. There, she encountered Landon running toward the back door. Landon told her to stay put so that she would not get hurt. But Emily instead went toward the living room, found it empty, and then continued out onto the porch, where she saw her dad “rolling around in the snow.” On cross-ex- amination, Emily added details to her description, confirming that the man she saw enter the living room was big—six feet tall—and heavy-set in addition to being masked. She also con- firmed that she had described him as Black to police when she was interviewed on the night of the robbery. Nathan Howie, the Danville police officer who had inter- viewed Emily, largely confirmed her account of the interview during his testimony. He recounted that she had described the intruder as “5’ 10” to 6 foot,” “fat build,” and Black or “mixed” with a light complexion. But his recollection, after reviewing his contemporaneous report, was that she de- scribed the intruder as wearing “a black stocking cap,” not a mask. No. 20-1432 9
Landon was the next significant witness. Describing the robbery and shooting, he testified that a single man walked in and “shut[] the door nonchalant, like he was supposed to be there.” He then stood near the door for 15 to 20 seconds before pulling out a gun and shooting Landon in the stomach. Lan- don pointed to Reyes when asked whether he saw the shooter in the courtroom. Landon’s direct-examination testimony about events after the shooting matched up with Emily’s. On cross, Landon conceded that the shooter had been wearing a hoodie with the hood up, obscuring part of his fore- head. He claimed not to remember describing a second in- truder or calling the shooter a “light-skinned” Black man, as- serting that he had always identified the shooter as a “very light-skinned Mexican or Puerto [R]ican.” Regarding the ar- ray identification, defense counsel cross-examined Landon at length about his repeated exposure to photos of Reyes and his focus on Peaslee at the suppression hearing. During redirect, Landon once again mixed up the two men. Landon kept in- sisting that he was relying on facial hair in his identifications, and he indicated that he had singled out Peaslee from the ar- ray. Citing those exchanges and their inconsistency with Lan- don’s courtroom gesture, the defense once again moved to suppress the identification, but the motion was denied. Next came Officer Garrett. On direct, he explained array procedures to the jury. He also explained that he had included Reyes in the first sets of arrays he showed to Landon in the hospital because Reyes’s name “had come up already” at that stage in the investigation. Like Landon, Garrett pointed out Reyes as the man Landon had identified using an array. On cross, Garrett conceded that Landon had changed his 10 No. 20-1432
description of the shooter over time, from a “light-skinned” Black male to “mixed race or Hispanic.” Hernandez testified that he was at first meant to serve only as the getaway driver, while Reyes and Smith were to carry the guns. Garcia was meant to be the “tape man,” mean- ing he was in charge of binding anybody found inside with duct tape while Reyes and Smith held them at gunpoint. But Garcia got cold feet at the last second and wanted to stay in the van, and so Hernandez agreed to play the tape-man role instead. He approached the house with Reyes and Smith. He then waited on the sidewalk and watched them go onto the porch, but he testified that he could not see whether or when they each entered the house. He fled down a nearby street when he heard gunshots, as shooting had not been part of the plan—Hutchins had described Thomas and Landon as “some sweet-ass white boys, and they wasn’t gonna do nothing.” On cross, Hernandez admitted that he was testifying with the benefit of a very favorable plea deal. In exchange for his testimony and plea of guilty to conspiracy to commit armed robbery, the state agreed to drop charges for “home invasion, aggravated battery with a firearm, first degree murder, and attempted first degree murder.” It also abandoned an unre- lated felony theft charge against Hernandez. Thanks to the deal, Hernandez was sentenced to only six years in prison and was scheduled for early release less than a year after the trial. Hernandez also admitted that Garrett had made clear to him that such deals would be on offer only to “people that were outside the house.” Hernandez conceded that he knew that Garcia and Wright already had given statements when he gave his, but that Reyes was then still at large and was the subject of an active manhunt. Finally, he confirmed that he No. 20-1432 11
had the same lawyer as Garcia and had consulted with that lawyer after Garcia gave a statement but before giving his own. Garcia testified that when the robbers arrived at the house, he suggested abandoning the plan. The others rejected that idea. As Garcia watched, Reyes grabbed a pistol from the van floor. Then Reyes, Hernandez, and Smith approached Thomas’s home while Garcia hung back. A few minutes later, Smith came running back to the van, shouting that “Reyes shot the guy.” Reyes followed shortly thereafter, and then the three drove off. Back at Wright’s house, Reyes was concerned that he had been shot, because he was covered in blood. Reyes also said, according to Garcia, that he had to shoot “the guy”—that is, Thomas—because Thomas had attacked him. Garcia recalled that Smith shot Thomas, and “they shot an- other guy running out the house.” On cross, Garcia admitted to testifying with the benefit of a plea deal similar to Hernan- dez’s. Indeed, Garcia’s deal was arguably better, as he already had been paroled by the time of the trial. And, like Hernan- dez, he conceded that he knew Reyes was still at large when he implicated Reyes as one of the shooters. But he denied re- calling Garrett’s telling him during one interview that “I think my main thing right now is looking at Juan and Dre” (mean- ing Reyes and Smith). Hutchins also testified, but because he had stayed in the Mercedes, he was not able to give a first-hand account of who played what role in the robbery. Several neighbors gave testi- mony that corroborated the broad contours of Hernandez’s and Garcia’s testimony. One reported seeing the maroon van parked near Thomas’s house. Another witnessed a man who appeared to her to be Hispanic fleeing in the direction in 12 No. 20-1432
which Hernandez claimed to have fled. And a neighbor with a clear view of Thomas’s porch saw several gunshot flashes. Smith did not testify. The defense called a single witness, Officer Garrett. His brief testimony as a defense witness concerned only a few stray details, such as the colors of the two vehicles the robbers had used. The jury convicted Reyes of first-degree murder, at- tempted first-degree murder, home invasion, and various ag- gravating factors. The judge sentenced Reyes to a life sentence for the murder of Thomas, a second life sentence to run con- secutively for the home invasion, and 30 years, also to run consecutively, for the attempted murder of Landon. E Reyes appealed, challenging the admission of Landon’s identification. The Illinois Appellate Court affirmed. People v. Reyes, No. 4-07-0412 (Ill. App. Ct. Oct 7, 2008). Although it held that the fourth array had been suggestive, it concluded that the suggestiveness was not “‘so impermissibly suggestive as to give rise to a very substantial likelihood of irreparable misidentification.’” Id. at 31 (quoting Simmons v. United States, 390 U.S. 377, 384 (1968) (emphasis added by the Illinois Ap- pellate Court). It explained that suggestiveness normally is a jury question, that the standard of review was deferential, and concluded that the record did not “absolutely compel” rever- sal. Id. Reyes unsuccessfully pursued both further direct review and state-court collateral review. He then filed an application for relief under 28 U.S.C. § 2254 in federal court. The district court denied his petition and declined to issue a certificate of No. 20-1432 13
appealability. Reyes then asked this court to grant the certifi- cate. We did so, and the present appeal followed. II We review de novo the district court’s denial of a section 2254 petition. Pruitt v. Neal, 788 F.3d 248, 264 (7th Cir. 2015). But because Reyes was convicted by a state court, “[o]ur re- view is governed (and greatly limited) by the Anti-terrorism and Effective Death Penalty Act of 1996,” or AEDPA. Hicks v. Hepp, 871 F.3d 513, 524 (7th Cir. 2017). Under AEDPA, Reyes is entitled to relief only if his conviction “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1). Reyes contends that the methods Officer Garrett em- ployed to procure Landon’s identification were unnecessarily suggestive and resulted in an unreliable identification, that AEDPA deference does not save the state court’s contrary holding, and that admission of the identification at trial was so prejudicial as to justify a retrial. We consider his arguments in that order. A In Stovall v. Denno, 388 U.S. 293, 302 (1967), the Supreme Court recognized that the introduction of identification evi- dence resulting from police conduct that is “unnecessarily suggestive and conducive to irreparable mistaken identifica- tion” violates due process. This principle applies to photo ar- rays as well as line-ups, show-ups, and other similar identifi- cation procedures. Simmons, 390 U.S. at 383. When a criminal defendant moves to suppress an identification based on a photo array on reliability grounds, the analysis proceeds in 14 No. 20-1432
two steps. At the first step, we ask whether the disputed array and accompanying procedures were “unnecessarily sugges- tive,” or words to that effect. Manson v. Brathwaite, 432 U.S. 98, 104 (1977). (Some cases ask whether the procedures were “un- duly” or “impermissibly,” rather than “unnecessarily,” sug- gestive. See, e.g., United States v. Gregory-Bay, 332 F.3d 1036, 1045 (7th Cir. 2003) (“unduly”); United States v. Williams, 522 F.3d 809, 810 (7th Cir. 2008) (“impermissibly”). But whatever term one chooses, the point is, as one justice of the Supreme Court has put it, that all three terms “reinforce our focus not on the act of suggestion, but on whether the suggestiveness rises to such a level that it undermines reliability.” Perry v. New Hampshire, 565 U.S. 228, 254 n.3 (2012) (Sotomayor, J., dis- senting).) If the array and procedures were unnecessarily sugges- tive, we proceed to the second step, asking whether the un- necessary suggestiveness rendered the identification so unre- liable that its admission violated the defendant’s right to due process. Brathwaite, 432 U.S. at 114. At both steps, the analysis turns on “the totality of surrounding circumstances.” Sim- mons, 390 U.S. at 383. We often have informed our consideration of identifica- tion evidence with insights from social-science literature. See United States v. Acox, 595 F.3d 729, 730 (7th Cir. 2010); see also United States v. Hall, 165 F.3d 1095, 1118–20 (7th Cir. 1999) (Easterbrook, J., concurring). As we have explained, “[i]t often takes evidence from psychology and statistics to decide whether a photo spread or lineup is ‘unduly suggestive’ and, if so, whether the suggestiveness is ‘irreparable.’” Acox, 595 F.3d at 730. “Lawyers’ assertions that the effects of a photo spread are ‘clear’ or ‘obvious’ are no substitute for evidence.” No. 20-1432 15
Id. Nonetheless, this remains an area devoid of clear bench- marks. We look first at the procedures Garrett used to procure an identification from Landon. The parties devote considerable time to a debate over just how much suggestiveness qualifies as “unnecessary.” There is no precise formula we can use, but we are willing to say for present purposes that the procedures the state used here were sufficiently suggestive that they would flunk any test. Garrett showed Landon pictures of Reyes on at least four separate occasions. On the first and second viewings, Reyes’s picture was one of twenty-four options; on the third viewing, his picture was one of twelve; and the fourth time, Reyes’s picture was one of only six. Worse, after Landon suggested that Peaslee greatly resembled the shooter, Garrett removed Peaslee’s photograph from future arrays, thereby eliminating the best available confounder from consideration. In short, Garrett’s chosen procedures transparently homed in on Reyes, whose picture kept showing up even as the number of available alternatives decreased. This progression of arrays would have suggested to any attentive witness that the police suspected Reyes. And it might have caused even an inattentive witness to make a mis- taken identification because of the well-documented psycho- logical process of “unconscious transference”: recognizing Reyes and attributing the familiarity to having seen him on the day of the shooting, when in truth any recognition was a product of prior exposure to his photograph. See Young v. Conway, 698 F.3d 69, 81–82 (2d Cir. 2012) (discussing uncon- scious transference); Kenneth A. Deffenbacher, Brian H. Born- stein, & Steven D. Penrod, Mugshot Exposure Effects: 16 No. 20-1432
Retroactive Interference, Mugshot Commitment, Source Confusion, and Unconscious Transference, 30 Law and Hum. Behav. 287, 299–306 (2006) (same). Indeed, the Supreme Court has acknowledged the risk of unconscious transference since the 1960s, albeit without employing the vocabulary of social sci- ence. See Simmons, 390 U.S. at 383–84 (warning that an eye- witness may be “apt to retain in his memory the image of the photograph rather than of the person actually seen”). Under either party’s proposed standard for suggestiveness, the pro- cedures Garrett used to procure Landon’s identification were unnecessarily suggestive. That takes us to reliability. Here we ask whether the indi- cators of Landon’s “ability to make an accurate identification” were “outweighed by the corrupting effect of law enforce- ment suggestion.” Perry, 565 U.S. at 239 (cleaned up). If so, Landon’s identification was unconstitutionally unreliable and should have been suppressed. Id. We evaluate reliability, like suggestiveness, by the totality of the circumstances. Simmons, 390 U.S. at 383. The Supreme Court has identified five specific circumstances that usually will be probative: the opportunity of the witness to view the crimi- nal at the time of the crime, the witness’ degree of attention, the accuracy of the witness’ prior description of the criminal, the level of certainty demonstrated by the witness at the confronta- tion, and the length of time between the crime and the confrontation. Neil v. Biggers, 409 U.S. 188, 199–200 (1972).
If it were up to us, we would have no trouble saying that each of these considerations supports Reyes’s claim that No. 20-1432 17
Landon’s identification was unreliable. We explain why be- fore turning to the impact of AEDPA on the case. The facts of Brathwaite—a case in which the Supreme Court found no sub- stantial likelihood of irreparable misidentification, 432 U.S. at 116—provide a useful counterpoint. The witness’s opportunity to view the suspect. Landon had very little time to view the shooter’s face—just “15, 20, 25 sec- onds.” This was far less than the officer in Brathwaite. See id. at 114 (“two to three minutes”). And the record reflects an un- resolved dispute about how clear Landon’s view was—he tes- tified himself that the shooter was wearing a hoodie that “[m]aybe covered his forehead.” The witness’s degree of attention. Landon was not, like the officer in Brathwaite, a trained observer who knew at the time that he had good reason to fix the shooter’s face in his memory. See id. at 115. And Landon had been consuming al- cohol and marijuana for several hours when he saw the shooter, and so may have been impaired. The accuracy of the description. Landon’s initial description of the shooter was limited to a comment about skin color that might have fit at least two other robbers (Hernandez and Gar- cia). His characterization of the shooter’s race also changed over time. Contrast the description in Brathwaite, which was both consistent and thorough, covering race, height, build, hair color, hairstyle, clothing, and cheekbone structure. See id. An eyewitness’s description need not “satisf[y] Proust,” Big- gers, 409 U.S. at 200, but it should at least provide the basics. Landon’s did not. The level of certainty. The state points out that Landon iden- tified Reyes immediately and with confidence in July. But we 18 No. 20-1432
cannot take that point in isolation. He did so only after he failed to identify Reyes on at least three prior occasions. We are willing to disregard Landon’s inability to identify Reyes on February 3, when he may have been on a heavy dose of morphine. And his failure to identify Reyes on February 23 may have stemmed from the fact that Garrett used a different photograph, and so Reyes’s appearance was slightly differ- ent. Even so, Landon’s certainty in July is impossible to square with his uncertainty on February 13. The same photograph was used on both dates, and we find it implausible that the later identification is more probative of the shooter’s identity than the earlier failure to identify. To be sure, the state sug- gests one reason to discount the February 13 failure: Landon was still medicated at the time. But all we know is that he was “taking pain pills,” not that he was substantially impaired. Nor did the Illinois courts make a finding of fact to that effect. Time between the crime and the identification. The state sug- gests that the time between the February 13 viewing and the July identification mitigated any suggestiveness from Lan- don’s first exposure to the photo. Some authority supports that view. See United States v. Harris, 281 F.3d 667, 671 (7th Cir. 2002). But the passage of time cuts two ways: it may mitigate suggestiveness, as the memory of the prior exposure to the suspect’s photo fades; but it undermines reliability, as memory of the incident also fades. Notably, more time passed between the shooting and the July identification than between February 13 and that identification. Two other circumstances beyond those identified in Big- gers also favor Reyes. First, Landon leaned heavily on his memory of the suspect’s facial hair to identify Reyes, a dubi- ous heuristic given the ease with which facial hair can be No. 20-1432 19
changed. That problem is magnified by Landon’s inability to identify Reyes in the more recent (February 13) photograph, the one depicting Reyes (and his facial hair) at a moment much closer in time to the shooting than the older photo- graph. And second, in his colloquies with defense counsel about the Peaslee photograph, Landon comes across as con- fused and manipulable (if not incoherent). Landon’s initialing of the Peaslee photograph at the suppression hearing was not a clear repudiation of his prior identification of Reyes. But nei- ther was it suggestive of a highly reliable, confident eyewit- ness. In short, if we were reviewing this on direct appeal, we would conclude that the circumstances of Landon’s identifi- cation of Reyes show that it was an unreliable product of un- necessary police suggestion and that its admission was error. B But we are not on direct appeal. Instead, this case comes to us on appeal from the district court’s denial of a section 2254 application, and so we must apply the filter of AEDPA deference. We may not grant Reyes relief just because the Illi- nois Appellate Court’s affirmance of the identification’s ad- mission strikes us as wrong; we may do so only if it “was con- trary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.” 28 U.S.C. § 2254(d)(1). A state court’s determination is “contrary to” clearly established law if it ap- plies a rule that contradicts that law. Cullen v. Pinholster, 563 U.S. 170, 182 (2011). If the state court identifies the proper law, then the question becomes whether its application of that rule is unreasonable. Id. An “unreasonable application” occurs only when “‘there was no reasonable basis’” for the state 20 No. 20-1432
court’s decision. Id. at 188 (quoting Harrington v. Richter, 562 U.S. 86, 98 (2011)). In conducting either inquiry, we bear in mind that “[o]ur own decisions, as well as those of other cir- cuits or state courts, are informative only insofar as they may shed light on our understanding of the authoritative Supreme Court precedents.” Lewis v. Zatecky, 993 F.3d 994, 1000 (7th Cir. 2021). Reyes argues that the Illinois Appellate Court’s opinion was “contrary to” law because that opinion, as he reads it, found suggestiveness but conducted “no reliability analysis at all.” But he is not pointing to the state court’s failure to con- sult the correct Supreme Court precedents. He is arguing in- stead that the state court failed to apply all of the elements of that governing law—both the suggestiveness branch and the reliability branch. We thus conclude that, if Reyes is to avoid AEDPA deference, he must show that the state court unrea- sonably applied the applicable law. In this connection, the problem for Reyes is that the Su- preme Court has warned us, time and again, to interpret state- court opinions in the section 2254 context with a large meas- ure of charity. Indeed, even a summary denial that cites no rule of law at all may merit deference in certain circumstances. See Pinholster, 563 U.S. at 187–88. Although the Illinois Appel- late Court never used the word “reliability” in any of its forms, it did conclude in the end that there was no “‘very sub- stantial likelihood of irreparable misidentification.’” People v. Reyes, No. 4-07-0412, slip order at 31 (Ill. App. Ct. Oct 7, 2008) (emphasis deleted). The internal quotation comes directly from Simmons, see 390 U.S. at 384, and articulates that case’s standard for reliability, not suggestiveness. On the generous reading required by AEDPA, it is possible to understand the No. 20-1432 21
state court as holding that the identification was not so unre- liable that it led to “irreparable misidentification.” It is difficult, as the Supreme Court has recognized, for a petitioner to demonstrate an unreasonable application of the relevant law. Richter, 562 U.S. at 101. Reyes recognizes this, and for good reason: the reliability inquiry is fact-bound. Once a court finds unnecessary suggestiveness in identifica- tion procedures, it encounters a line-drawing problem. Often, suggestiveness can be mitigated by the simple expedient of ventilation before the jury. The reliability inquiry aims to fer- ret out the particularly egregious cases in which that solution will not suffice. But reasonable jurists will differ about where, precisely, to draw the line between flawed identifications that can go to the jury and those that cannot. It follows that AEDPA deference will often be appropriate in close cases so long as the jury heard and considered the evidence tending to impeach reliability. This is such a case. Reyes’s defense at trial hinged in large part on his attempts to impeach Landon’s identification. And the Illinois Appellate Court rested its affirmance in significant part on that fact, declaring that, “[n]ormally, suggestiveness is a matter to be argued to the jury.” Although we may not have come to the same conclusion as the state court, we can- not say that it was so far out of bounds as to justify granting relief in the face of AEDPA. Indeed, as we explain below, even without the extra thumb on the scale furnished by AEDPA, the impeachment of Landon was so central to the defense that 22 No. 20-1432
we could also affirm on the alternate ground that Reyes can- not show prejudice from the flawed identification. C A party seeking relief under section 2254 must demon- strate not only error that eludes AEDPA deference but also “actual prejudice”—that the state court’s error had a “sub- stantial and injurious effect or influence in determining the jury’s verdict.” Czech v. Melvin, 904 F.3d 570, 577 (7th Cir. 2018) (quoting Brecht v. Abrahamson, 507 U.S. 619, 623 (1993)). Under Brecht, “actual” means something like real, discernible, or outcome-determinative—i.e., “more than a ‘reasonable possibility’ that the error was harmful.” Davis v. Ayala, 576 U.S. 257, 268 (2015) (quoting Brecht, 507 U.S. at 637). We must therefore consider whether, in light of the error, we are left with “grave doubt” about whether it had a substantial and injurious effect on the verdict. Richardson v. Griffin, 866 F.3d 836, 845 (7th Cir. 2017) (quoting Ayala, 576 U.S. at 268). Reyes’s argument stumbles at this juncture. For at least two reasons, the state court’s erroneous admission of Lan- don’s identification did not cause Reyes actual prejudice. First, even without Landon’s identification, the jury heard persuasive evidence establishing that two shooters entered the house and that Reyes was one of them. Second, the defects in Landon’s identification were presented to the jury in detail. We begin with the other evidence of Reyes’s guilt. Any at- tentive juror would have been certain of at least two things by the close of trial. First, Andre Smith was a shooter. In their interviews, both Emily and Landon described one of the in- truders as a larger Black man matching Smith’s appearance. Every witness who testified about which of the robbers No. 20-1432 23
entered Thomas’s home agreed that Smith had done so. And none of Reyes’s evidence suggested otherwise. Second, our hypothetical juror would have been confident that there had been a second shooter. Multiple witnesses put two armed rob- bers in the house. Thomas was hit by bullets fired from two different guns. And none of Reyes’s evidence suggested only one shooter. In light of those facts, two questions would have re- mained: Who was the second shooter? And which of the two shooters—Smith or the second man—shot Landon? As to the first question, all the evidence pointed to Reyes. Garcia and Hernandez both testified that Reyes had entered the house alongside Smith. Both were able to describe the rob- bery’s planning and its botched execution at a high level of particularity. Their accounts were internally consistent and consistent with one another. And both accounts were corrob- orated by the testimony of disinterested witnesses who con- firmed hard-to-falsify details—for example, the timing and direction of Hernandez’s flight from the scene just after gun- fire broke out. A jury hearing such evidence and knowing that two shooters entered the home could readily have concluded that Reyes was one of them. Reyes counters that Garcia and Hernandez had both the motive and the opportunity to fabricate their testimony. The motive was lenience: Garrett made clear to both that a deal was available if they were willing to testify that Reyes entered the house. (And both were staring at potential murder charges and so had every incentive to cooperate.) The oppor- tunity was also easy to see: Garcia and Hernandez shared a lawyer who could have ferried information back and forth around the time of their conversations with the investigating 24 No. 20-1432
officers. To be sure, these points tended to impeach both wit- nesses to some degree. But some uncertainty is not “grave doubt.” And, crucially, the jury itself knew all the reasons for doubting Garcia and Hernandez, yet it overrode them when it reached its verdict. It no doubt did so after evaluating the demeanor and credibility of the two witnesses, “a role re- served to the jury,” not a post-conviction court. United States v. Henderson, 736 F.3d 1128, 1131 (7th Cir. 2013). We conclude that even without Landon’s unreliable identification, the jury reasonably could have decided that Reyes was the second shooter. That conclusion disposes of any claim of prejudice Reyes has relating to two of the three charges: Thomas’s murder and the home invasion. Thomas was hit by bullets from two guns, and nobody ever has suggested that both guns were fired by a single shooter. If the jury was entitled to find that Reyes was the other robber in the house with Smith—as it was—it fol- lows that it was also entitled to find that he fired some of the shots that killed Thomas. And that finding would suffice to convict Reyes of both murder and home invasion—charges for which Reyes received consecutive life sentences. That leaves Reyes’s third conviction, for the attempted murder of Landon. That charge presents a closer question than the other two. Unlike Thomas, Landon was struck by bullets from just one gun. The only surviving witnesses to that shooting (excluding the possible shooters) were Landon and perhaps Emily. Emily’s testimony tended to undermine Lan- don’s; she described the first intruder in a manner consistent with Smith’s appearance and did not recall any shooting early in the encounter or testify to having seen a second intruder. Landon, in contrast, claimed to have been shot by the first No. 20-1432 25
intruder just a few seconds after the intruder entered Thomas’s home. He identified that shooter as Reyes, not Smith. Given these facts, the jury could have concluded that Reyes, rather than Smith, shot Landon only if it credited Lan- don’s unreliable identification. Nothing else gave the jury any grounds to distinguish between the two men. Still, we conclude that Reyes was not prejudiced on this count by the erroneous admission of Landon’s identification. Once again, the most important point is that the impeachment evidence—in this context, that impeaching Landon’s identifi- cation—was fully available to the jury. As we have said, it was one of the central themes of the defense in both its cross-ex- aminations and its closing argument. That airing considerably mitigates the risk of prejudice. Moreover, Reyes himself has never argued that we should separate his attempted-murder conviction from his other con- victions. His claim throughout has been that Landon’s identi- fication was crucial evidence putting him in the house, thereby implicating him in all three crimes. He has thus forfeited any argument that relies on the differences in evidence support- ing the different convictions. Nor can we overlook that forfei- ture, because prejudice vis-à-vis the attempted murder would not necessarily follow even if we concluded that Smith, rather than Reyes, was the one who shot Landon. In that circum- stance, Reyes might still have been liable for his accomplice’s attempted murder; both were in the house shooting people at the same time. Whether he was would depend on Illinois state law and the jury instructions. But neither party has briefed those issues because Reyes has not argued, either in the state courts or here, that he might be entitled to relief from the at- tempt conviction alone. 26 No. 20-1432
To sum up: The jury had adequate evidence to convict Reyes of the murder and home-invasion charges without put- ting any stock in Landon’s identification. Although the at- tempted-murder charge might have presented a closer ques- tion on direct review, that is not the posture of this case. The jury there, too, had a rational path leading to its decision. Reyes has not shown prejudice. III We thus conclude that the state courts did not unreasona- bly apply established law to this case, and in the alternative, with or without AEDPA deference, Reyes cannot show preju- dice from the admission of Landon’s identification. The judg- ment of the district court is AFFIRMED.