concurring in part and dissenting in part.
I fully join the majority’s treatment of Denise Shepard-Fraser’s challenges to the sufficiency of the evidence and the reasonableness of her sentence. I cannot join, however, the majority’s treatment of Jorge Correa-Osorio’s challenge to the in-court identification. In my view, that challenge has merit, and, accordingly, I would reverse the judgment of conviction on the ground that the District Court plainly erred in allowing the jury to weigh that evidence.
I.
Eyewitness testimony is undeniably powerful. That testimony is all the more powerful when the eyewitness identifies the defendant right in front of the jury. Ordinarily, we let juries weigh such testimony, just as they may weigh any other admissible evidence. But in certain circumstances, concerns about the reliability of an in-court identification — with all the persuasive force that comes from the witness identifying “that man” as the person who committed the crime — require more than faith in the jury’s capacity to evaluate what is reliable and what is not. See, e.g., Kampshoff v. Smith, 698 F.2d 581, 585 (2d Cir.1983) (“[Djoubts over the strength of the evidence of a defendant’s guilt may be resolved on the basis of the eyewitness’ seeming certainty when he points to the defendant and exclaims with conviction that veils all doubt, ‘[Tjhat’s the man!’ ” (second alteration in original) (quoting United States v. Wade, 388 U.S. 218, 235-36, 87 S.Ct. 1926, 18 L.Ed.2d 1149 (1967))).
One such circumstance arises when the government elicits the identification in court after having used suggestive out-of-*30court means to prompt the witness to make an earlier identification. See Simmons v. United States, 390 U.S. 377, 382, 88 S.Ct. 967, 19 L.Ed.2d 1247 (1968); see also Manson v. Brathwaite, 432 U.S. 98, 97 S.Ct. 2243, 53 L.Ed.2d 140 (1977). The classic out-of-court, government-designed, suggestive means are a stacked lineup, see Foster v. California, 394 U.S. 440, 442-43, 89 S.Ct. 1127, 22 L.Ed.2d 402 (1969), a highly suggestive photo array, see Simmons, 390 U.S. at 382, 88 S.Ct. 967, or, perhaps even worse, a show-up — in which the government brings the suspect before the witness in a one-to-one confrontation, see Stovall v. Denno, 388 U.S. 293, 295, 87 S.Ct. 1967, 18 L.Ed.2d 1199 (1967).
When the government uses such out-of-court prompts unnecessarily, due process bars the jury from weighing the in-court identification unless it survives review under the so-called Biggers factors. See Neil v. Biggers, 409 U.S. 188, 199-200, 93 S.Ct. 375, 34 L.Ed.2d 401 (1972); see also United States v. Maguire, 918 F.2d 254, 264-65 (1st Cir.1990) (“The Supreme Court has declared generally the same test for the admissibility of an in-court identification subsequent to a suggestive out-of-court identification as it has employed for admission of an allegedly suggestive pretrial out-of-court identification.”). Courts use those factors to decide whether the in-court identification arises from the witness’s pri- or encounters with the person identified rather than from the influence of the out-of-court suggestive prompt the government has used. See United States v. Castro-Caicedo, 775 F.3d 93, 97 (1st Cir.2014). And when those factors indicate a substantial risk that the suggestive prompt did corrupt the in-court identification — say because the witness encountered the person identified years before and never since, and then only in conditions not likely to make the memory stick with any accuracy — then the jury may not consider the in-court identification. Biggers, 409 U.S. at 198, 93 S.Ct. 375.
Correa does not argue to us that an out-of-court suggestive prompt preceded this in-court identification. He instead argues that what happened in the courtroom was alone so suggestive as to necessitate review under Biggers. And I conclude that, on these facts, Correa is right.
In reaching this conclusion, I recognize courtroom identifications are a traditional feature of criminal trials. But tradition should not distract from what to me seems obvious and what I do not understand the majority to deny. A prosecutor who orchestrates an in-court identification does at least risk exposing the jury to a very misleading form of unusually powerful and prejudicial testimony. For that reason, judges must be on the lookout for the case in which that risk is realized — even if it is realized through means other than the government’s prior use of egregious out-of-court suggestive prompts.
Here, we confront government-selected, in-court means for prompting that, though subtle, were plenty suggestive-and unnecessarily so. So much so, in my view, that even on plain error review, an inspection of the identification’s reliability under Biggers is required. See United States v. De León-Quiñones, 588 F.3d 748, 753 (1st Cir.2009) (“To establish plain error, a defendant ‘must show an error that was plain, (i.e., obvious and clear under current law), prejudicial (i.e., affected the outcome of the district court proceedings), and seriously impaired the fairness, integrity, or public reputation of the judicial proceedings.’ ” (quoting United States v. Griffin, 524 F.3d 71, 76 (1st Cir.2008))). And, after undertaking that inspection, I further believe that due process demands a greater degree of assurance that Correa was “that man” than this identification can possibly *31supply. For that reason, I cannot agree that this in-court identification — on which the government’s case almost entirely rests — may be the cause of Correa’s long-term loss of liberty.26
II.
To begin, I must first explain why Perry v. New Hampshire, — U.S. —, 132 S.Ct. 716, 181 L.Ed.2d 694 (2012), does not shield from Biggers review any in-court identification that is untainted by a prior suggestive out-of-court prompt — the seemingly categorical position the Eleventh Circuit takes. See United States v. Whatley, 719 F.3d 1206, 1216 (11th Cir.2013). But the explanation is not hard to give. Simply put, Perry did not involve an in-court identification at all. Perry thus cannot set the standard for how we should treat one.
Perry concerned only whether a prior out-of-court identification should have been subjected to review under the Biggers factors. And Perry concluded that due process did not require such review in that ease because the government did not orchestrate the out-of-court identification. See Perry, 132 S.Ct. at 725-27. Rather, as Perry explained, the witness had made a spontaneous out-of-court identification of the suspect while the suspect stood next to a police officer in a parking lot. Id. at 721-22. The Court noted that defense counsel — in briefing and in argument — explicitly conceded that the government “did not arrange the suggestive circumstances surrounding [the] identification,” id. at 725, and that defense counsel did “not allege any manipulation or intentional orchestration by the police,” id. (citing Tr. of Oral Arg. 5). There was, as the New Hampshire Supreme Court found, “a complete absence of improper state action.” New Hampshire v. Perry, No. 2009-0590, 2010 WL 9105720, at *1 (N.H. Nov. 18, 2010) (quoting New Hampshire v. Addison, 160 N.H. 792, 8 A.3d 118, 125 (2010)). Thus, notwithstanding that the circumstances of the identification may have been suggestive, the Court concluded that the deterrence rationale that underlies the whole Biggers line had no application. Perry, 132 S.Ct. at 726.
In my view, therefore, Perry is no per se bar to finding plain error here. Perry is instead best read to affirm what the Court had said before about when the Biggers test must be applied. Due process requires the Biggers review for reliability “when law enforcement officers use an identification procedure that is both suggestive and unnecessary.” Id. at 724. For it is only when the government is responsible for the suggestiveness that due process requires an inquiry into the reliability of the identification.
III.
With Perry out of the way, the issue reduces to the following. When, if ever, should an in-court identification be subject to Biggers review by virtue of the suggestive attributes of what happened in the courtroom itself? Substantial precedent shows that a government-orchestrated, in-court identification may, in some circumstances, be so suggestive as to trigger Biggers review, even absent a prior, out-of-court suggestive prompt. See, e.g., United States v. Greene, 704 F.3d 298, 307-08 (4th *32Cir.2013); United States v. Rogers, 126 F.3d 655, 657-58 (5th Cir.1997) (“[I]t is obviously suggestive to ask a witness to identify a perpetrator in the courtroom when it is clear who is the defendant.”); United States v. Hill, 967 F.2d 226, 232 (6th Cir.1992); United States v. Rundell, 858 F.2d 425, 426 (8th Cir.1988); see also United States v. Beeler, 62 F.Supp.2d 136, 140-45 (D.Me.1999) (suppressing, in advance of trial, in-court identification because it would be impermissibly suggestive and unreliable). And the government does not argue otherwise.
The majority contends, however, that such review is required only when the in-court identification involves special features of concern. The majority then notes that, here, the prosecutor’s in-court questions or comments did not expressly draw the witness’s attention to the defendant or directly suggest the hoped-for result. See Greene, 704 F.3d at 307-08. Nor was the defendant in this case of a different race or gender from all other persons in the courtroom or at counsel table. See United States v. Archibald, 734 F.2d 938, 941 (2d Cir.1984). But, as I will explain, the government still orchestrated this identification to occur in circumstances that clearly were unnecessarily suggestive — at least when the circumstances are considered in their totality.
As to orchestration, the challenged in-court identification did not just happen. Instead, the prosecutor brought it about through a cooperating witness, José Vega-Torres, who had been prepped and promised (according to his own testimony) that his cooperation would protect his wife from prosecution. And Vega was asked to provide testimony that was — the government concedes — the whole of the case against Correa.
As to suggestiveness, the prosecutor asked Vega to identify Correa only on Vega’s second day of testifying, after the lunch break. Thus, the prosecutor asked Vega to make the identification only after Vega had spent a day and a half on the stand, with Correa — the only male defendant- — seated before him at counsel table. And during Vega’s day-plus time on the stand, defense counsel rose to object at numerous points. By the time of the identification, then, the object of the prosecution would have been obvious.
The majority points out that the government did not put Correa at counsel table. He took that seat on his own, as. it was his right to do. But the government still chose to seek the identification from Vega fully aware that the defendant was so positioned — and thus fully aware that Vega would be asked to identify Correa at a moment when he was at that table and after Vega had observed him there at substantial length. Indeed, it is fair to say the longest look that Vega — by his own account — ever had of Correa was during the time the government had asked Vega to appear on the stand prior to asking him to make the identification. Cf. United States v. Montgomery, 150 F.3d 983, 992 (9th Cir.1998) (“permitting [the witness] to view [the defendant] in the courtroom the day before the witness was scheduled to testify” found to be a “suggestive procedure[ ]”).
In these circumstances, I do not believe Correa is overstating things in contending that the government’s presentation of the identification was “the functional equivalent of the one person show-up, a classically suggestive method of identification in which an eyewitness is confronted with only one option to cho[o]se from.” See Greene, 704 F.3d at 307-08 (4th Cir.2013); Hill, 967 F.2d at 232; Beeler, 62 F.Supp.2d at 144-45; Commonwealth v. Crayton, 470 Mass. 228, 21 N.E.3d 157, 166 (2014) (“Where, as here, a prosecutor asks a wit*33ness at trial whether he or she can identify the perpetrator of the crime in the court room, and the defendant is sitting at counsel’s table, the in-court identification is comparable in its suggestiveness to a showup identification.”). And, for that reason, I believe it clear that the identification’s reliability must be tested under the Biggers factors.
After all, the government cannot show that these suggestive means were somehow “necessary.” See Perry, 132 S.Ct. at 724 (finding “[cjrucial” to the Court’s allowance of a suggestive procedure in Stovall that it was a “necessity”); Stovall, 388 U.S. at 302, 87 S.Ct. 1967 (finding a suggestive encounter necessary because “[n]o one knew how long [the hospitalized witness] might live” and thus “the police followed the only feasible procedure”); United States ex rel. Kirby v. Sturges, 510 F.2d 397, 403-04 (7th Cir.1975) (Stevens, J.) (concluding, despite there being “no evidence of bad faith or excessive zeal to obtain a conviction,” that a showup was “unnecessarily suggestive” because it was “not justified by any exigent circumstances, or even by any minimal showing of inconvenience”). True, as a general matter, the prosecution is entitled to present its case in the order it thinks best. But that standard feature of the way that we organize criminal trials does not mean it was necessary for the government to elicit the identification of Correa in the-highly suggestive way that it chose.
The government could have relied on an identification by Vega that he made out of court through a non-suggestive means. In fact, the government introduced photo arrays showing that Vega had made out-of-court identifications of other alleged conspirators. Yet the government chose not to pursue that same approach at trial in Correa’s case. And, even if the government wished to proceed with an in-court identification, the government acknowledged at oral argument that there were mitigating measures that could have been taken in court but were not. See Beeler, 62 F.Supp.2d at 144-45 (noting “that a number of courts have reasoned that the preferred remedy for a suggestive in-court identification is not, necessarily, the suppression of the identification but an in-court lineup or some other protective measure to ensure the fairness of the identification and cross-examination of the eyewitness”); see also Kirby, 510 F.2d at 405-06 (describing “unanimity among scholars” that “evidence of, or derived from, a show-up identification should be inadmissible unless the prosecutor can justify his failure to use a more reliable identification procedure”).
I recognize that, in one respect, the situation is different when, unlike here, an in-court identification is made only after a prior, suggestive out-of-court identification. In that type of case, the jurors do not have the opportunity to witness the complained of suggestive circumstance themselves. Here, by contrast, the jurors did, as they obviously watched what happened in court. But, that does not show that an in-court identification of this sort must be left to the jury to weigh. For while in theory the jurors were well-positioned to evaluate the suggestiveness of what they saw, in fact the jurors were exposed to a seemingly certain identification made in a very suggestive setting. In other words, the fact that the jury witnessed this particular identification does not solve the problem. It potentially is the problem. See Kampshoff, 698 F.2d at 585 (“There can be no reasonable doubt that inaccurate eyewitness testimony may be one of the most prejudicial features of a criminal trial. Juries, naturally desirous to punish a vicious crime, may well be unschooled in the effects that the subtle compound of suggestion, anxiety, and for*34getfulness in the face of the need to recall often has on witnesses.” (footnote omitted)).
In light of the “obvious[ ] suggestivefness]” of the circumstances in which the government chose to elicit the identification, I conclude — as the Fifth Circuit did more than a decade ago in considering an unobjected-to, in-court identification — that even on plain error review, the Biggers test applies. See Rogers, 126 F.3d at 658. In that case, the Fifth Circuit confronted circumstances not unlike those at issue here and undertook such reliability review because the witness had been asked to make the identification “when it [was] clear who [was] the defendant.” Id. Rogers explained that the witness initially provided testimony without making any identification but was then called back to the stand the next day to identify the defendant, the only black man at counsel table, as the culprit.27 Id. at 657-58 & n. 1. And while the Fifth Circuit noted that the concern about suggestiveness was “heightened” in that case because the witness was of a different race from the defendant, id. at 658, here that same concern is heightened by the duration of the witness’s preidentification testimony and the gender of the defendant relative to the other defendant at counsel table.
In both cases, therefore, the key fact is the ultimate one that the Biggers line suggests to me should matter: each time, it was obvious who the defendant was when the prosecution asked the witness to make the in-court identification. So although I grant that there is no case finding plain error on facts exactly like these, it seems to me that the relevant inquiry into suggestiveness that Biggers requires yields an answer no less obvious in Correa’s case than the Fifth Circuit found in Rogers. And given that “the ‘plainness’ of [an] error can depend on well-settled legal principles as much as well-settled legal precedents,” United States v. Brown, 352 F.3d 654, 664 (2d Cir.2003) (emphasis omitted), I do not believe counsel’s failure to object at trial should shield this in-court identification from the scrutiny that Biggers plainly requires in comparably suggestive circumstances.
IV.
In consequence of the unnecessarily suggestive means that the government used, I believe we must apply the Biggers factors to test whether Vega’s identification was corrupted by the suggestiveness of the setting. See 409 U.S. at 199-200, 93 S.Ct. 375.28 And it is plain to me that, after doing so, we should have no confidence that Vega’s identification rested on his rec*35ollection of Correa’s appearance from prior encounters — to the extent they qualified as “encounters” — rather than from the suggestive circumstances in which the prosecutor asked Vega to make his choice.
Vega testified to viewing Correa, known to him in the conspiracy as “El Don,” on only three occasions, each of which occurred long ago. The only physical interaction was brief, at most lasting a matter of minutes. And the encounters between the two men could be described, at best, as indirect: putting luggage in a car, a sighting through a car window, and finally, Vega’s glimpse of “El Don” handling baggage on the tarmac while Vega waited in the airport from what must have been a substantial distance. The witness never purported to have had a personal relationship with the defendant. Nor did Vega offer any previous description that matched Correa’s.
Furthermore, Vega claimed to have interacted with Correa only in 2006, while Vega’s in-coxxrt identification of Correa occurred five years later, in 2011. Biggers suggested, however, that a far shorter, seven-month gap between sighting and identification counted against the identification’s reliability. Id. at 201. And while we have sometimes permitted lengthy time gaps, we have not been so forgiving when the encounters were as fleeting and indirect as these. See, e.g., United States v. Flores-Rivera, 56 F.3d 319, 331 (1st Cir.1995) (excusing seven-year time gap when “other reliability criteria were sufficiently persuasive”).
Vega did not appear to hesitate in identifying Correa from the stand. But certainty on the part of the witness does not reveal much. See United States v. Jones, 689 F.3d 12, 18 (1st Cir.2012) (“[L]ack of confidence is certainly a reliable warning sign, while the presence of confidence is probably closer to a neutral factor.”). Such apparent certainty may result from the suggestive circumstances, and certainty after suggestive prompting cannot show reliability. See, e.g., Raheem v. Kelly, 257 F.3d 122, 139 (2d Cir.2001) (finding it “difficult to view [a witness’s] certainty as an indicator of reliability independent of the suggestive lineup”).
Not surprisingly, therefore, the government does not argue that the Biggers factors support the identification’s reliability. This identification, which should have triggered the Biggers test, clearly cannot survive it.
V.
That leaves the issue of prejudice. See United States v. Delgado-Marrero, 744 F.3d 167, 184 (1st Cir.2014) (describing the “infrequent case[] in which reversal is warranted” under the plain error standard of review); see also Rogers, 126 F.3d at 658-60 (finding error in the admission of an obviously suggestive and unreliable in-court identification, . but holding error harmless in light of overwhehning evidence of guilt). But here, the prejudice is clear.
The government conceded, that Vega’s testimony provided the only direct evidence that linked Correa to this conspiracy. And a review of the record also reveals many — in some cases, inexplicable— weaknesses in-the government’s case. These weaknesses include: the inability of the testifying investigative agent to explain how and why Correa was identified as a member of the conspiracy; the evidénce suggesting Correa worked for a different baggage handling company than any of the other coconspirators tasked with the same alleged role in the conspiracy; the testimony suggesting Vega made out-of-court identifications of many of the co-conspirators without apparently making an out-of-court identification of Coxrea; another tes*36tifying co-conspirator’s non-recognition of Correa; and the failure of the government to call any co-conspirators — despite having entered into plea agreements with them— who had dealt personally with Correa.
There is some corroborating evidence of Correa’s involvement in the conspiracy: a resume listing his email address as including the words “El Don” and his wife’s ownership of the same car Vega testified that “El Don” had at Vega’s one interaction with him. But that evidence itself traces back only to Vega’s testimony. Thus, the other evidence against Correa gives no assurance that, even absent the identification by Vega, the jury would have convicted. See, e.g., United States v. Casas, 356 F.3d 104, 123-24 (1st Cir.2004) (vacating conviction of defendant Cunningham when — apart from improper testimony — only evidence of involvement with conspiracy was his identification by three, co-conspirators with whom he had few contacts).
This case, then, is like those that have given courts the. most pause about the due process implications of admitting in-court identifications. The problematic identification here was essential to the government’s case. See Raheem, 257 F.3d at 142 (“The identification testimony of [the witnesses] clearly bore on an essential issue, the identity of the shooter. And that testimony was crucial to the prosecution’s case, for the State presented no evidence other than the testimony of [the witnesses] to tie [the defendant] to the events.”); Kampshoff, 698 F.2d at 588 (“[A]ll told, absent the identification testimony, the evidence against Kampshoff is simply not overwhelming.”); cf. United States v. Williams, 436 F.2d 1166, 1168 (9th Cir.1970) (“[W]here the question of guilt or innocence hangs entirely on the reliability and accuracy of an in-court identification, the identification procedure should be as lacking in inherent suggestiveness as possible.”).
Still, to succeed on plain error review, the defendant must make one more showing. He must demonstrate that the error “seriously impaired the fairness, integrity, or public reputation of the judicial proceedings.” De León-Quiñones, 588 F.3d at 753 (quoting Griffin, 524 F.3d at 76). But that additional requirement poses no obstacle for Correa. When the government introduces an unnecessarily suggestive in-court identification on which nearly the whole of the prosecution depends, it seems to me clear that this last part of the plain error test is met.
VI.
I recognize the concern about opening the door to other successful attacks on in-court identifications. But not every in-court identification is similarly staged, with the cooperating witness prepped to testify, brought to the stand, and then prompted to give an identification of the one male defendant at counsel table and then only after having provided a day and a half of testimony. Not every in-court identification rests on minimal contacts that occurred long ago and thus plainly lacks sufficient indicia of reliability to survive the Biggers test. Not every prosecution rests so heavily on an in-court identification this shaky. And fewer still are the prosecutions that are so dependent on in-court identifications that share these concerning attributes. I thus do not think it too much to require the government to make the case against Correa on evidence less likely to mislead the jury in this needless way. And, in my view, the Due Process Clause requires that same conclusion.
I respectfully dissent.
. Correa received a prison sentence of eleven years. That term, I might add, is greater than the sentences received by all other members of the conspiracy, save for the leader, who received a sentence of 135 months — only three months longer than Correa’s. Correa and his co-defendant in this trial, Denise Shepard-Fraser, were the only alleged conspirators who went to trial. All others listed in the indictment, with the exception of one individual against whom charges were dismissed, entered into plea agreements.
. The prosecutor identified "something odd” in the witness’s demeanor during her initial testimony and afterwards had the FBI case agent approach her. Rogers, 126 F.3d at 657. The witness told the agent during that conversation that she recognized the defendant as the culprit and the government then asked— and was granted permission — to recall her to the stand. Id. And while the defendant’s counsel objected to the witness being recalled, counsel did not object to her identification after she took the stand. Id. at 657-58.
. The factors to consider in evaluating whether the identification is reliable despite the unnecessary suggestiveness of the identification procedure 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.” Biggers, 409 U.S. at 199-200, 93 S.Ct. 375. Of course, as the majority notes, courts need not apply these factors to test the reliability of an identification not orchestrated by the government to occur in an unnecessarily suggestive way. See Maj. Op. 23 n. 17.