dissenting.
Because I cannot agree with the majority’s narrow view of the record in this case, I respectfully dissent. The majority requires that Garvey articulate his objections with near-surgical precision, yet casts aside the inconvenient fact that the trial court prevented Garvey’s counsel from explaining the full scope of his suppression motion after the close of evidence — arguably the most important time. Even so, Garvey still succeeded in raising the issue of suggestive civilian conduct through the testimony elicited during the suppression hearing. Indeed, Garvey’s counsel so readily raised the issue that the trial court’s findings of fact and conclusions of law specifically address it in language that the majority glosses over. Moreover, the trial court rendered a ruling on the issue of suggestive civilian conduct by erroneously dismissing it as immaterial to the constitutional suppression analysis. Ironically, and unfortunately, the majority endeavors to transform the trial court’s dismissal of the question into Garvey’s failure to raise it.
Principally as a result of my view of the facts, I also conclude that the majority’s application of N.Y. C.P.L. § 470.05(2) is exorbitant and thus inadequate to bar our review of Garvey’s constitutional claim. Finally, perceiving no obstacle to our review of the merits, I conclude that the *721state appellate court unreasonably applied clearly established Supreme Court law. I therefore would vacate the judgment of the District Court and remand for a determination, in the first instance, of whether the trial court’s error had a “substantial and injurious effect or influence in determining the jury’s verdict.” Wray v. Johnson, 202 F.3d 515, 525 (2d Cir.2000) (internal quotation marks omitted).
I. There Is No Independent and Adequate State Procedural Bar
For simplicity’s sake, I shall assume that the majority’s strict view of § 470.05(2) is correct. See supra, at 714-717. I note that reasonable minds can differ on this point. See N.Y. C.P.L. § 470.05(2) (“[A] party who without success has either expressly or impliedly sought or requested a particular ruling or instruction, is deemed to have thereby protested the court’s ultimate disposition of the matter ... sufficiently to raise a question of law with respect to such disposition or failure regardless of whether any actual protest thereto was registered. ” (emphases added)); Cotto v. Herbert, 331 F.3d 217, 247 (2d Cir.2003). Nevertheless, this issue is not the crux of my disagreement. Taking for granted the majority’s exacting view of § 470.05(2)—ie., that in seeking suppression, Garvey had to challenge suggestive civilian conduct as opposed to suggestive conduct generally— Garvey satisfied that standard for two reasons: because he raised the issue of suggestive civilian conduct at the suppression hearing, and because the trial court rendered a ruling on the issue, albeit an erroneous and dismissive ruling.
A. Garvey Raised the Issue of Suggestive Civilian Conduct
I wish to make clear something that the majority has downplayed: the trial court abruptly cut short Garvey’s counsel’s explanation of all the grounds of the suppression motion. After the close of testimony at the suppression hearing, Garvey’s counsel began to explain her motion, as follows:
Counsel: Your Honor, I would move to suppress the identification of Mr. Garvey at the time of his arrest. Based on the fact that it was, that the information that Officer Davis had at the time was unreliable, Mr. Garvey should have been taken to the precinct and put in a lineup and afforded the opportunity of having, you know, suggestive—
The Court: Rest on the record.
Counsel: Yes. We rest on the record.
The trial court’s directive to “rest on the record” most certainly hampered counsel’s ability to fully articulate the grounds of the suppression motion once all the evidence was in. This handicap alone raises serious questions as to the adequacy of the purported procedural bar that the majority embraces. To me, it seems unwise to require Garvey to lodge pinpoint objections when the trial court directs his counsel to quit explaining his objections and rest on the record, especially given that the independent and adequate state bar doctrine “is prudential rather than jurisdictional.” Cotto, 331 F.3d at 238.
Even setting aside this pragmatic concern, prior to the trial court’s instruction to rest on the record, Garvey’s counsel’s questions and Officer Davis’s responses more than adequately raised the issue of suggestive civilian conduct. Specifically, Garvey’s counsel questioned Davis as follows:
Counsel: Now, you said that you arrived at the scene and you saw a crowd of people; is that correct?
Davis: Yes, ma’am.
*722Counsel: And where was Mr. Garvey in relation to that crowd?
Davis: He was — he was surrounded by the crowd.
Counsel: Okay. And was anybody [ie., any civilian] holding him at the time?
Davis: Not that I recall, no.
Shortly after that exchange, Garvey’s counsel again queried “the location where [Davis] came upon Mr. Garvey surrounded by the neighbors?”
Garvey’s counsel further probed the civilian-created, and suggestive, conditions under which Davis first encountered Garvey, as follows:
Counsel: Now, when you came upon this crowd surrounding Mr. Garvey, did you notice any merchandise ... [i]n the area?
Davis: There was a machine which I later learned to be a video compressor.
Counsel: Where exactly was this located?
Davis: It was on the ground in the same area where they were all standing.
Garvey’s counsel then questioned the manner in which McKenzie identified Garvey under the conditions described above. In particular, Garvey’s counsel elicited testimony that after Garvey had been held by Gaines and surrounded by a crowd of neighbors with the purloined compressor at his feet, one member of that crowd, McKenzie, approached Davis and identified Garvey as the man who “had burglarized her home earlier that day.” Once McKenzie made that identification, Davis arrested Garvey.
It was that identification, made under the circumstances described above, that Garvey challenged in his suppression motion as “unconstitutionally suggestive.” The fact that Garvey’s counsel also asserted in the motion papers that Garvey was handcuffed at the time of McKenzie’s identification — which was not true, as the testimony showed — does not mean that Garvey challenged only police conduct. Nowhere does the motion contain any limitation as to police conduct only; rather, the motion papers challenge the identification on the broad “ground that the ‘showup’ was unconstitutionally suggestive,” among others. Read fairly and in context, the motion challenged and the testimony specifically addressed all the circumstances surrounding the identification, including those related to civilian and police conduct.
Given the broad wording of the motion papers and the scope of the testimony, it is no wonder, then, that the trial court’s factual findings bear directly, and in some instances only, on the issue of suggestive civilian conduct. In particular, the trial court found that when the arresting officer arrived, he
observed] a small group of people, in the midst of which was the defendant. The officer did not observe the defendant being held by anyone. The officer was already in the driveway, and Mr. Theodore Ga[i]nes, approached him and told the officer that defendant had been trying to move property from his backyard; that he grabbed him and was holding him until the police came.
Indeed, the trial court concluded that the crowd surrounding Garvey was aggressive enough that Garvey “was moved [by the officer] to a doorway some distance away from the crowd for his own protection,” and that shortly thereafter, “Violet McKenzie approached the officer and said that she had observed the defendant remove property from her home earlier and that also the property that was in Ga[i]nes’ backyard was property that belonged to her.”
Similarly, the trial court’s conclusions of law bear directly on the issue of suggestive *723civilian conduct, specifically the conclusion that,
No suggestive acts occurred by the police department. The holding of the defendant initially was by a private citizen and when the officer was investigating it, another private citizen, identifying herself, approached him and said that she was a witness to [the] complaint of a burglary occurring shortly before in her premises.
I cannot help but wonder why, if Garvey had failed to raise the issue of suggestive civilian conduct, the trial court rendered factual findings and legal conclusions on that very subject.
The record passages cited above establish two points. First, Garvey raised the issue of suggestive civilian conduct because his counsel elicited specific testimony as to (1) whether, prior to McKenzie’s identification, her neighbors surrounded Garvey in the driveway, (2) whether any of the neighbors, such as Gaines, were holding Garvey, (3) whether, at the time the crowd surrounded Garvey, McKenzie’s stolen property was lying at Garvey’s feet, and (4) whether McKenzie came forth from the crowd to identify Garvey under the circumstances already described. Simply put, the motion sought to suppress the identification on the ground of suggestiveness, and the testimony, not to mention the trial court’s findings of fact and legal conclusions, specifically addressed suggestive civilian conduct. Accordingly, I cannot understand how the majority concludes that Garvey failed to raise the issue in a manner sufficient “to provide the trial court with a fair opportunity to consider” it. Supra, at 715.
B. The Trial Court Issued a Ruling on the Issue of Civilian Suggestiveness
Further, the trial court actually ruled on whether suggestive civilian conduct required suppression of McKenzie’s identification testimony. I do not pretend that the trial court’s ruling is perfectly neat; brief oral bench rulings, such as the one issued here, rarely are. Nevertheless, the trial court said enough to show that it considered the issue of suggestive civilian conduct and found it irrelevant. Specifically, the trial court’s rulings that “[n]o suggestive acts occurred by the police department” and that “[t]he holding of the defendant initially was by a private citizen,” establish that the trial court (1) was quite aware of the issue of suggestive conduct by private citizens, such as Gaines’s holding of Garvey, but (2) believed that suggestive civilian conduct — as opposed to police conduct — raised no constitutional concerns. That is, counsel’s questions brought the issue of civilian conduct to the trial court’s attention, but the trial court considered civilian conduct to be legally immaterial, and therefore dismissed it in passing. Hence its conclusion, styled as one of law, that the “[t]he holding of the defendant initially was by a private citizen,” and thus not a proper ground for suppression. Also telling is what the trial court did not say: after noting that some suggestive civilian conduct had occurred-— again, Gaines holding Garvey — the trial court did not rule, or even imply, that Garvey had failed to challenge such conduct. Instead, the trial court erroneously treated the question, properly raised for decision, as if it were irrelevant to the constitutional suppression analysis.
Lest one think that I am adopting a strained reading of the trial court’s decision, I note that it is unsurprising that the trial court took this view; the state of the law in our circuit at the time of the trial court’s ruling left open the possibility that suggestive civilian conduct was legally immaterial. It was not until one month after *724the trial court’s ruling that we expressly-held that suggestive civilian conduct, just as much as police conduct, may raise constitutional concerns. See Dunnigan v. Keane, 137 F.3d 117, 128 (2d Cir.1998) (“The linchpin of admissibility, therefore, is not whether the identification testimony was procured by law enforcement officers, as contrasted with civilians, but whether the identification is reliable.”).1 What is surprising, however, is the majority’s effort to morph the trial court’s erroneous ruling on an issue into Garvey’s failure to raise it.
In sum, as I see the facts of this case, Garvey’s counsel raised the issue of suggestive civilian conduct, the trial court considered it in its factual and legal findings, and the trial court actually, although improperly, ruled on it by dismissing the question as legally irrelevant. Each of those acts satisfies § 470.05(2)’s preservation requirement, even as the majority strictly interprets it. Indeed, both the Magistrate Judge and the District Court Judge who reviewed Garvey’s petition similarly concluded.
C. Section 470.05(2), as Applied by the Majority, Is An Inadequate Bar to Review
In addition to my factual disagreement with the majority, its application of § 470.05(2) is exorbitant and thus inadequate to bar our review of Garvey’s constitutional claim. As the majority recognizes, “there are ‘exceptional cases in which exorbitant application of a generally sound rule renders the state ground inadequate to stop consideration of a federal question.’ ” Cotto, 331 F.3d at 240 (quoting Lee v. Kemna, 534 U.S. 362, 376, 122 S.Ct. 877, 151 L.Ed.2d 820 (2002)). Three factors guide the inquiry into whether a case fits “within that limited category.” Id. (internal quotation marks omitted).
First, “whether the alleged procedural violation was actually relied on in the trial court, and whether perfect compliance with the state rule would have changed the trial court’s decision.” Id. Second, “whether state caselaw indicated] that compliance with the rule was demanded in the specific circumstances presented.” Id. Third, “whether petitioner had ‘substantially complied’ with the rule given ‘the realities of trial,’ and, therefore, whether demanding perfect compliance with the rule would serve a legitimate government interest.” Id. (quoting Lee, 534 U.S. at 381-85, 122 S.Ct. 877).
My view of the facts drives my analysis of these three factors. In respect of the first factor, the trial court did not rely, and could not have relied, on Garvey’s purported default because there was no default. Further, Garvey’s compliance with § 470.05(2) had no effect on the trial court’s decision because Garvey raised the issue of suggestive civilian conduct and the trial court erroneously brushed it aside. This factor, therefore, favors Garvey.
As to the second factor, I have accepted, for the sake of argument, the majority’s strict interpretation of § 470.05(2), and found it satisfied here, which, in my view, quite decisively favors Garvey. Finally, *725regarding the third factor, Garvey substantially complied with the rule given the reality of the suppression hearing. His counsel elicited testimony regarding the civilian-created circumstances surrounding McKenzie’s identification and, before counsel could fully explain the bases of the motion, the trial court instructed her to rest on the record. In light of the trial court’s instruction to stop speaking, Garvey’s counsel substantially complied with the state’s preservation requirement by raising the issue through the prior testimony. The third factor thus favors Garvey as well. Accordingly, the majority’s application of § 470.05(2) to the facts of this case is exorbitant and consequently inadequate to bar our review of Garvey’s constitutional claim.
For all those reasons, I conclude that there is no independent and adequate state bar to our review of Garvey’s constitutional claim.
II. The Appellate Division Unreasonably Applied Clearly Established Supreme Court Law
Having concluded that there exists no bar to our review of the merits of Garvey’s petition, I also conclude that the judgment of the Appellate Division, First Department constitutes an unreasonable application of clearly established Supreme Court law, as explained below.
A. Standards Applicable to Garvey’s Habeas Petition
In order to prevail on a habeas petition, Garvey must show that the state court’s “adjudication of the claim ... resulted in a decision that ... 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). Clearly established Supreme Court law consists of the Supreme Court’s holdings, as opposed to dicta. See Kennaugh v. Miller, 289 F.3d 36, 42 (2d Cir.), cert. denied, 537 U.S. 909, 123 S.Ct. 251, 154 L.Ed.2d 187 (2002).
An “unreasonable application” occurs when a “state court identifies the correct governing legal principle ... but unreasonably applies that principle to the facts of the prisoner’s case.” Cotto, 331 F.3d at 247 (internal quotation marks omitted). The application of federal law must be “objectively unreasonable,” which requires the petitioner to demonstrate “some increment of incorrectness beyond error.” Id. at 248 (internal quotation marks omitted). Nevertheless, “[t]his increment need not be great; otherwise, habeas relief would be limited to state court decisions so far off the mark as to suggest judicial incompetence.” Jenkins v. Artuz, 294 F.3d 284, 292 (2d Cir.2002) (internal quotation marks omitted).
Although there exists no test to determine whether a state court ruling is objectively unreasonable, three factors guide the inquiry: (1) whether there exists a “lack of any precedent supporting [the state court’s] result in the Supreme Court or any federal court of appeals”; (2) whether the state court has given “specific reasons” for its conclusion; and (3) whether the state court’s conclusion is consistent with the “purpose behind” the relevant rule. Cotto, 331 F.3d at 251-52.
B. Substantive Standards Governing the Admission of Identification Testimony Following a Suggestive Confrontation
The Due Process Clause forbids the admission of identification testimony where there exists a “very substantial likelihood of irreparable misidentification.” Neil v. Biggers, 409 U.S. 188, 198, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972) (internal quotation marks omitted). Once a suggestive con*726frontation occurs between the identifying witness and the suspect, “the central question” is “whether, under the totality of the circumstances, the identification was reliable even though the confrontation procedure was suggestive.” Id. at 199, 93 S.Ct. 375 (internal quotation marks omitted). In Biggers, the Supreme Court set forth “the factors to be considered” in determining whether an identification is sufficiently reliable notwithstanding some degree of suggestiveness. Those factors “include the opportunity of the witness to view the criminal 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 confrontation, and the length of time between the crime and the confrontation.” Id. at 199-200, 93 S.Ct. 375. In Manson v. Brathwaite, 432 U.S. 98, 114, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977), the Court added another step to the analysis: “Against these [Biggers ] factors is to be weighed the corrupting effect of the suggestive identification itself.” See also id. at 116, 97 S.Ct. 2243.
C. The Decisions of the Appellate Division and the District Court
On the merits, the Appellate Division concluded, without elaboration, that “the identification was sufficiently reliable under all the circumstances.” After unsuccessfully seeking leave to appeal in the Court of Appeals, Garvey filed this habeas petition. He principally urges that pursuant to the Biggers analysis, McKenzie’s near-total failure to initially describe the burglar to the police required the District Court to severely — -if not entirely — discount other factors that might weigh in favor of admissibility, such as McKenzie’s opportunity to view the burglar and the certainty of her later identification.
The District Court agreed that “[clearly, the identification of Mr. Garvey was suggestive.” The Court also “agree[d] that McKenzie’s inability to describe the burglars weighs against finding her later identification sufficiently reliable.” Nevertheless, the Court rejected Garvey’s claim that McKenzie’s initial failure to describe the burglar undercuts other Biggers factors, mainly because “[i]f the Court were to weigh the Biggers factors as Garvey urges, the ability of a witness to describe a suspect to the police prior to the identification would be dispositive” by itself, whereas the Biggers analysis is based upon “the totality of the circumstances.”
Accordingly, the District Court analyzed each Biggers factor independently and concluded that the identification was sufficiently reliable. In particular, the District Court relied upon the strength of three Biggers factors: McKenzie’s opportunity to observe Garvey, her attentiveness during the crime, and the certainty of her later identification. The District Court reasoned that,
(a) McKenzie had a clear opportunity to observe Garvey at the time of the crime; (b) she was not a casual or inattentive observer, as she descended the stairs for the specific purpose of investigating the loud noise she had heard; (c) after turning on the downstairs lights, McKenzie had a clear, unobstructed view of [Garvey] as he was trying to escape through McKenzie’s front door; (d) McKenzie also observed [his] face again, a few moments later, from a bedroom window, as he looked back repeatedly towards McKenzie’s house while fleeing; (e) McKenzie expressed certainty that Garvey was the burglar she saw flee[]ing her home; (f) it was just a few hours after the burglary that McKenzie told [the police] that Garvey was one of the burglars and was wearing the same *727clothes that one of the burglars had worn.
Despite its conclusion, the District Court granted a certificate of appealability on the ground that “the resolution of the constitutional issue underlying Garvey’s petition can be debated.”
D. The District Court Misapplied the Biggers Factors
As explained below, the District Court committed two legal errors in applying Biggers and Brathwaite to Garvey’s claim. In turn, those errors caused the District Court to ratify a conclusion that amounts to an unreasonable application of established Supreme Court precedent.
First, the District Court erred by considering the Biggers factors in isolation. As we have pointed out before, one Big-gers factor affects another — especially in cases where the victim initially fails to describe the perpetrator — and this interplay is not to be ignored. For example, in Raheem v. Kelly, 257 F.3d 122, 139-40 (2d Cir.2001), cert. denied, 534 U.S. 1118, 122 S.Ct. 930, 151 L.Ed.2d 892 (2002), we reasoned that a witness’s inability to describe the perpetrator shortly after the crime devalued the certainty of his later identification. There, a man in a black leather coat entered a bar in which two witness, Shiloh and Cooke, were drinking. Id. at 125. Shortly after the man shot and killed the owner of the bar, Shiloh and Cooke attempted to describe him to police, but could give only vague descriptions. Id. at 125-26, 139. They later picked Raheem out of a suggestive lineup in which he was the only person wearing a black leather coat, and the state court admitted their identification testimony. Id. at 126-27, 130. After his conviction, Raheem brought a habeas petition challenging the admission of that testimony, and the District Court dismissed the petition. Id. at 125.
In reversing, we reasoned that, “To the extent that either Cooke or Shiloh exhibited certainty [in their later identification of petitioner], we find it difficult to view that certainty as an indicator of reliability independent of the suggestive lineup, given their lack of recollection as to any physical features of the shooter’s face (except, as to Cooke, its roundness). Whatever their certainty, it was engendered by the suggestive element itself, the black leather coat.” Id. at 139.
Likewise, in Dickerson v. Fogg, 692 F.2d 238, 245-47 (2d Cir.1982), we reasoned that a witness’s initial failure to describe the perpetrator to police undercut the Biggers factors that might otherwise weigh in favor of reliability. There, a man named Colon was carjacked and robbed at gunpoint. Id. at 240-41. As the perpetrators ordered him out of the car, he briefly glanced the face of a man sitting in the rear passenger seat. Id. at 241. After the crime, Colon, much like McKenzie, could describe his assailants to the police only as “four black males.” Id. at 242.
Colon later saw Dickerson at the arraignment of another man who had been arrested while driving Colon’s stolen car. At that arraignment, Colon identified Dickerson as the man in the rear passenger seat of his car at the time of the carjacking. Id. at 241. Colon’s identification testimony was admitted at Dickerson’s trial. Id. at 239-40. After Dickerson was convicted, he petitioned for a writ of habe-as corpus. Id.
In affirming the District Court’s grant of the petition, we noted that Colon’s initial, vague description devalued other relevant factors, such as the degree of Colon’s attentiveness during the crime: “Lastly, any possibility that Colon’s training as a security supervisor contributed to his attentiveness is belied by his inability to *728describe the back seat passenger to the police with any degree of specificity.” Id. at 245. As in Raheem, we also concluded that the poor initial description undercut the certainty of the later identification: “In sum, not only does the technically accurate but unduly vague description of the back seat passenger furnished to the police before the confrontation diminish the reliability of [Colon’s] later identification, but, considering that Colon did not describe [Dickerson] in greater detail until after he viewed [Dickerson] in the courtroom, his second specific description was unmistakably the product of the [suggestive] viewing, and questionably reliable as well.” Id. at 246.
In this case, McKenzie almost completely failed to describe the burglar to the police when they first arrived at her house shortly after the crime occurred. According to the complaint report, McKenzie described the perpetrator only as a black male. When the responding officer asked her to describe the burglar’s age, height, weight, hair color, eye color, whether he had facial hair, and what clothes he was wearing, McKenzie could provide no information whatsoever, except that his clothes were “dark.” Her recollection of his appearance was so lacking that she would not even view photographs of suspects in an attempt to identify him. Pursuant to Ra-heem and Dickerson, McKenzie’s inability to describe the perpetrator shortly after the crime diminishes the extent to which one could conclude that she had ample opportunity to view the perpetrators or was attentive during the event. The District Court erred by failing to account for the manner in which McKenzie’s initial descriptive failure affects those other Big-gers factors.
Second, the District Court erred by failing to weigh the corrupting effect of the suggestive confrontation against the Big-gers factors, especially the factor addressing the certainty with which McKenzie later identified Garvey. Both Supreme Court and Second Circuit precedent expressly require this additional analytical step, yet the District Court simply did not undertake it. See Brathwaite, 432 U.S. at 114, 97 S.Ct. 2243 (“Against these [Biggers ] factors is to be weighed the corrupting effect of the suggestive identification itself.”); see also id. at 116, 97 S.Ct. 2243; Solomon v. Smith, 645 F.2d 1179, 1185 (2d Cir.1981) (“[A]s Brathwaite makes clear, the constitutional assessment of reliability requires a balancing of the factors outlined in Biggers, against the degree of suggestiveness in the impermissible procedures.”) (internal citation omitted).
Here, an extraordinary “degree of suggestiveness,” see Solomon, 645 F.2d at 1185, permeated McKenzie’s identification; one might go so far as to call these circumstances accusatory rather than merely suggestive. When McKenzie’s sister came to McKenzie’s door a few hours after the burglary, she told McKenzie, “they caught a guy.” McKenzie and her sister then went to Gaines’s house. When they arrived, Gaines was “holding” Garvey. Further, an agitated crowd had surrounded Garvey. Gaines testified that he apprehended Garvey after Garvey entered Gaines’s backyard in an attempt to remove property that Garvey had placed there “last night,” specifically, the video compressor. At the moment McKenzie first encountered Garvey, her stolen property quite literally lie at his feet. By failing expressly to weigh the degree of suggestiveness against the other Biggers factors — especially in light of these rather extreme circumstances — the District Court erred.
Correcting for the two errors identified above, it is apparent that there is no independent basis of reliability for McKenzie’s *729identification. As I have said, her inability to give but the most rudimentary and generic description of the perpetrator, even shortly after the crime, virtually precludes one from concluding that she had a reliable opportunity to view him or was attentive to any reliable degree. Likewise, the certainty of her later identification owes to the remarkably suggestive circumstances under which she encountered Garvey. Accordingly, both the Appellate Division and the District Court erroneously concluded that McKenzie’s identification was reasonably reliable.
I pause briefly to note that, as this analysis demonstrates, the District Court was mistaken in believing that if it allowed McKenzie’s initial descriptive failure to affect the other Biggers factors, then “the ability of a witness to describe a suspect to the police prior to the identification would be dispositive” by itself. Instead, it is the combination of three relatively rare facts that is dispositive here: (1) McKenzie’s almost total failure to describe the perpetrator to the police (2) despite the fact that she was interviewed shortly after the burglary, plus (3) the severely suggestive circumstances that prevailed during her later identification. If any of those facts were tempered, the outcome might be different.
As things stand, however, the next question is whether the Appellate Division’s judgment is more than merely erroneous, such that it constitutes an unreasonable application of Supreme Court law within the meaning of § 2254. Pursuant to the three factors we set forth in Cotto, I believe that it is. The first factor — whether there exists a “lack of any precedent supporting [the state court’s] result in the Supreme Court or any federal court of appeals” — is somewhat difficult to evaluate because cases such as this one are very fact-specific. Cotto, 331 F.3d at 251. Nevertheless, as explained above, our precedent strongly indicates that when a witness fails to describe a perpetrator shortly after the crime, and then later identifies someone under highly suggestive circumstances, the identification testimony generally should be suppressed. Raheem, 257 F.3d 122; Dickerson, 692 F.2d 238. This factor therefore somewhat favors a finding of objective unreasonableness.
The second factor — whether the state court has given “specific reasons” for its conclusion — strongly favors a finding of objective unreasonableness, as neither the state trial court nor the Appellate Division gave a single reason why McKenzie’s identification was reliable. Id. The third factor asks whether the state court’s conclusion is consistent with the purpose behind the rule. The Supreme Court has explained that the purpose of the Biggers analysis is to avoid the “primary evil” of “a very substantial likelihood of irreparable mis-identification.” Biggers, 409 U.S. at 199, 93 S.Ct. 375 (internal quotation marks omitted). The state court’s conclusion in this case — that “the identification was sufficiently reliable under all the circumstances” — is inconsistent with the goal of avoiding misidentification. Allowing testimony about identifications, such as McKenzie’s, that occur after an almost total failure of recollection and arise from highly suggestive circumstances, only increases the chances that a misidentification will occur. This factor also weighs in favor of a finding of objective unreasonableness. On balance, therefore, I conclude that the Appellate Division’s judgment is an unreasonable application of clearly established Supreme Court law.
For the above reasons, I would vacate the judgment of the District Court. I would remand for consideration of the only remaining question in this case, and one that the District Court did not reach given its conclusion that no constitutional error *730occurred: whether the error I have identified had a “substantial and injurious effect or influence in determining the jury’s verdict.” Raheem, 257 F.3d at 142 (internal quotation marks omitted); see also Wray v. Johnson, 202 F.3d 515, 525 (2d Cir.2000).
. The fact that we did not expressly clarify this point until shortly after the trail court's decision does not impact the merits of Garvey's petition. Were we to reach the merits, we would look to the law in effect "at the time of the Appellate Division’s judgment,” not that in effect at the time of the trial court's decision. Harris v. Kuhlmann, 346 F.3d 330, 345 (2d Cir.2003). By the time the Appellate Division rejected Garvey's appeal, we had decided Dunnigan, in which we interpreted Supreme Court precedent to require suppression of identification testimony when the identification initially was made under unduly suggestive, civilian-created circumstances. See Dunnigan, 137 F.3d at 128-30.