concurring in the judgment:
I concur in the remand, and I agree with much of the discussion in Judge Ruiz’ opinion. I write separately, however, because my emphasis differs significantly from the court’s.
I.
As we reiterated in our opinion in Benn’s first appeal,
[a]n exercise of discretion must be founded upon correct legal standards. See, e.g., In re J.D.C., 594 A.2d 70, 75 (D.C.1991). “A [trial] court by definition abuses its discretion when it makes an error of law,” Koon v. United States, 518 U.S. 81, 100, 116 S.Ct. 2035, 135 L.Ed.2d 392 (1996).
Benn v. United States, 801 A.2d 132, 142 (D.C.2002) (Benn I) (quoting Teachey v. Carver, 736 A.2d 998, 1004 (D.C.1999)). In my opinion, the transcript of the hearing at which the trial judge decided not to admit Dr. Penrod’s testimony reveals that the judge departed from correct legal principles in three sepai'ate, though related, respects:
1. The judge seemed, at least initially, to treat the issue before him as being whether expert testimony is admissible in eyewitness identification cases generally. Indeed, the judge inquired as to the “score” as between judges who admit such testimony and those who exclude it. This approach effectively assumes that one rule fits all such cases, and that the answer is always yes or always no, regardless of the particular facts. This approach is contrary, inter alia, to Green v. United States, 718 A.2d 1042, 1051 (D.C.1998). Although the judge subsequently (and appropriately) disclaimed any absolute per se bar against receiving such expert testimony, he plainly disfavored such evidence, and he did not, at least explicitly, consider the specifics of the defense proffer in light of the record before him, as he was required to do by Green and by other authorities cited in the majority opinion.
*12852. The judge appeared to reject, as misinformed, the jurisprudence of appellate courts, including the Supreme Court, see, e.g., United States v. Wade, 388 U.S. 218, 228, 87 S.Ct. 1926, 18 L.Ed.2d 1149 (1967); and this court, see, e.g., Benn I, 801 A.2d at 145 n. 11, to the effect that, more than other evidence, “stranger-to-stranger identification is often untrustworthy and presents some danger of misiden-tification.” Indeed, the judge stated: “I think, you know, it just doesn’t.”
3. The judge declared that he was “not impressed” by our analysis in Benn I of the quality of the eyewitnesses identifications in this case. That analysis, however, was a part of our holding in Benn 1, 801 A.2d at 145. We concluded in Benn I, that the evidence for the prosecution was “fairly decent ... but by no means overwhelming,” and we ruled, substantially for that reason, that the trial court’s error in excluding Benn’s mother’s testimony (for a violation of the rule on witnesses) was prejudicial rather than harmless.
In my opinion, Benn has not shown that our precedents require reversal of his conviction or that, on the merits of the defense motion, the judge’s exclusion of Dr. Penrod’s testimony constituted an abuse of discretion. I am not satisfied, however, that the judge based his exercise of discretion on correct legal principles, and for that reason — and for that reason alone*1— I do not agree with the government’s view that we should affirm Benn’s conviction.
II.
This case is of ancient vintage. It arises from a crime which was committed in 1992. To some extent, almost seventeen years later, it remains shrouded in mystery. The record does not reveal, for example, how it came to be that Raymond Benn’s photograph was included in the photo spread that was shown to the five members of the Mahoney family several days after the kidnaping and murder of Charles (Sean) Williams. No connection between Benn and the Mahoney family, or between Benn and Williams, was established. Although the motive for the crime was apparently the alleged nonpayment of a debt, there was no evidence that Williams owed Benn any money. Thus, almost two decades after Williams’ death, we have yet to learn why Benn became a suspect.
Further, Benn secured a reversal in Benn I on the grounds that he was precluded from presenting an alibi defense, but he then asserted no such defense at the second trial. In other words, Benn initially claimed that he was denied a fair trial because his alibi was excluded, but when he had the opportunity to present his alibi to a jury, his attorneys declined to do so, preferring instead to attempt to attack the prosecution’s case with expert testimo*1286ny — a tactic which they could have used, but failed to employ, at Benn’s first trial. I suppose that this was a legitimate defense tactic, but I find the sequence less than edifying.
However, be that as it may, it is undisputed, that all five Mahoneys positively identified Benn at the first trial2 as one of the two men who abducted Williams from their home. This was not, by any stretch of imagination, a flimsy government case. If Benn is innocent, then all five Maho-neys, none of whom knew Benn or had any motive to lie, must have been mistaken in making their identifications. The five identifications tend to corroborate one another. Therefore, as we noted in Benn /, “the prosecution’s evidence seems, at first blush, to be quite formidable, for Benn was identified by five apparently disinterested witnesses.” 801 A.2d at 145.
But things are not always as simple as they seem. As we observed in Benn I, and as the court notes in the majority opinion,
closer scrutiny places the strength of the case in substantial doubt. All of the witnesses were strangers to Benn. When shown a photo spread which included Benn’s picture, four of the five witnesses said that the photograph “looks like” the tall man who accompanied the decedent. Common sense tells us that many people resemble one another, and in that sense, “looks like” is not really an identification at all.[3] The purported coincidence that three of the witnesses described themselves as 95% certain understandably troubled the judge and indicates, at least, that someone probably suggested something to somebody. Benn was never placed in a line-up, and although all five witnesses “positively” identified him in the courtroom, it is difficult to hypothesize a more suggestive setting for an identifying witness, when the individual whom the witness had selected from the photo array was seated at the defense table, and the witness could infer that the police obviously believed that the man whose photograph the witness had described as “looking like” the culprit was indeed guilty.[4]
Id. (footnote omitted). Because of our “substantial doubt” regarding the strength of the case, we concluded in Benn I that the erroneous exclusion of Benn’s mother’s testimony was not harmless and that it required reversal of Benn’s convictions and a new trial. Our assessment of the flaws in the government’s case at the first trial was a part of our holding and the law of the case, and it was binding on the trial court.
III.
At the beginning of the hearing regarding the admissibility of Dr. Penrod’s proposed testimony, the judge posed the following question to Benn’s attorney:
There are 59 judges on this court. What’s the score? Do you keep score who allows [expert testimony regarding eyewitness identification] and who *1287doesn’t. It is largely in favor of not allowing it, isn’t that true?
Counsel acknowledged that “it’s largely been in favor of not allowing it although not exclusively in favor of not allowing it.” The judge’s question was predicated on the assumption that expert testimony on this subject is either always admissible or always inadmissible, and that a majority of Superior Court judges who had confronted the question had ruled that it was inadmissible.
The prosecutor emphatically reinforced this assumption:
I don’t know the exact count, but my understanding of the 59 judges in this court is that only one has allowed it in. And that’s because the vast majority of the judges in this court believe that this is a not uniquely scientific area that’s appropriate for expert testimony.
Moreover, while acknowledging this court’s statement in Green that “there may be cases in which a jury would find such testimony helpful,” 718 A.2d at 1051, the prosecutor argued for a somewhat curious construction of the Green opinion:
It doesn’t say that there are cases where that would be helpful, but there may be at some point somewhere down the road a possibility.
This is not the case. It would be our position that in fact there's never a case where the type of expert testimony they are seeking to introduce is appropriate because it would usurp the very gut of what the jurors are here to do. The jurors are here to assess credibility.
(Emphasis added.)
In fact, and contrary to the prosecutor’s argument, we made it clear in Green 718 A.2d at 1051, that “Dyas5 and its progeny do not articulate a per se requirement that all expert testimony about the reliability of eyewitness identification must be exclucl-ed.” On the contrary, we explained in Green, 718 A.2d at 1051, that
the Dyas case and its progeny simply upheld discretionary calls by the trial court in the circumstances presented. Dyas does not exclude expert testimony about the reliability of eyewitness identification for all purposes and under all circumstances, even where a trial court, in its discretion, believes the jurors might find such testimony truly helpful. Conversely, a determination by the trial court excluding such testimony as not “beyond, the ken of the average layman” is a■ in,ling only that upon the particular proffer made and, in the concrete setting of that case, the possible assistance of the expert testimony to the jury is insufficient to outweigh the, potential for distracting the, junj or supplanting its customary role in evaluating credibility. Under Dyas, as under any case concerning the admissibility of expert testimony, we will review the trial court’s decision for abuse of discretion, whether the trial court admits or excludes the proffered testimony. See Oliver v. United States, 711 A.2d 70, 73 (D.C.1998) (per curiam) (“It is well established that a trial judge has broad discretion to admit or exclude expert testimony, and that a decision either way should be affirmed unless it is manifestly erroneous.”) (quoting Spencer v. United States, 688 A.2d 412, 417 (D.C.1997)); cf. General Elec. Co. v. Joiner, 522 U.S. 136, 118 S.Ct. 512, 139 L.Ed.2d 508 (1997) (“A court of appeals applying ‘abuse of discretion’ review to such rulings may not categorically distinguish between rulings allowing expert testimony and ruling which disallow it.”) In other words, Dyas and its progeny do not articulate a per se requirement that all expert testimony about the reliability *1288of eyewitness identification must be excluded.
(Emphasis added.) We further stated in a footnote:
In State v. Schutz, 579 N.W.2d 317 (Iowa 1998), the Iowa Supreme Court overruled its 1979 decision that established a per se rule excluding expert witness testimony on eyewitness identification. The court noted that it had found no appellate court other than Iowa that had such a per se rule. See id. at 320. It also noted that most of the scientific literature on the subject had been published subsequent to the 1979 decision. See id. at 319-20. The Schütz court cited a number of recent decisions that upheld the admission of such expert testimony or even held its exclusion to be an abuse of discretion. See id.
Id. at 1051 n. 9. Green thus stands for the proposition that, in exercising his or her discretion with respect to the admission of proffered expert testimony such as Dr. Penrod’s, the trial judge must consider the “particular proffer made” in the “concrete setting of [the particular] case,” in light of each of the three Dyas factors.6
In this case, the trial judge’s inquiry regarding the “score” among Superior Court judges suggests that he may initially have been leaning towards the inflexible view favored by the government, namely, that expert testimony is never admissible in an eyewitness identification case. Later in the hearing, however, the judge took a more nuanced position, and he sensibly declined to adopt a per se rule. The judge stated that in his experience “the worst kind of eyewitness cases are one-witness identification cases,” but that “when we have five eyewitnesses] ... who positively or almost positively identify somebody, you feel much more confident about the accuracy of the overall identification process than you do if it’s only one witness.” He added:
“I don’t think this is one of those cases that Green ... talks about where this kind of testimony would be helpful ... I don’t want to be construed as saying I would never allow it. But I don’t think this is the right case. I just don’t.”
Because the judge thus ultimately disclaimed adherence to any per se rule excluding expert testimony in all eyewitness identification cases, it cannot fairly be said that he failed to exercise discretion.
The judge did not, however, allude at all to the specifics of the defense proffer or to the proposed content of Dr. Penrod’s testimony. According to Benn’s counsel, Dr. Penrod was prepared to testify that, according to his empirical research, (1) there exists only a moderate correlation between a witness’ confidence or.certainty in an identification and its actual accuracy; (2) over time, witnesses may grow increasingly confident in the accuracy of their identification, on account of influences that have no bearing on reliability; (3) witnesses are highly susceptible to unconscious influences during identification procedures; (4) stress and the presence of a weapon reduce, rather than enhance, accurate recollection and perception; and (5) the accuracy of an identification diminishes when the *1289witness has only a short time to observe the assailant, especially when more than one person participated in the criminal activity. The gravamen of Dr. Penrod’s evidence would have been that many of the conclusions supported by the research are unknown to lay jurors, and that some of these conclusions are counter-intuitive. The judge did not analyze, or even mention, any of the specific subjects regarding which Dr. Penrod was prepared to testify, nor did he focus on how such specific testimony would bear on the three Dyas factors in the context of the identification testimony in this case.
In my view, it is fair to say the judge started with the conception that exclusion of expert testimony regarding the reliability of eyewitness identification of strangers was the norm, and that at least in most cases, such testimony would not be helpful. Further, although, as the hearing proceeded, the judge wisely disclaimed adherence to a per se rule of exclusion, and although he appropriately differentiated between a single positive identification and five positive identifications, he did not focus “upon the particular proffer made” in “the concrete setting of [this] case” to determine whether Dr. Penrod’s testimony would assist the jury rather than distract it. Green, 718 A.2d at 1051.
IV.
“The vagaries of eyewitness identification, and the potential for wrongful convictions or adjudications based upon such evidence, have long been recognized in the District of Columbia.” In re As.H., 851 A.2d at 459-60 (citing United States v. Telfaire, 152 U.S.App.D.C. 146, 149-51, 469 F.2d 552, 555-57 (1972) (per curiam)); Crawley v. United States, 320 A.2d 309, 311-12 (D.C.1974). In deciding Benn’s first appeal, we also took note of the dangers posed by eyewitness identification of strangers:
The identification of strangers is proverbially untrustworthy. The hazards of such testimony are established by a formidable number of instances in the records of English and American trials. Webster v. United States, 623 A.2d 1198, 1204 n. 15 (D.C.1993) (quoting Felix FRANKFURTER, THE CASE OF SACCO AND Vanzetti (1927)). This passage by Professor (later Justice) Frankfurter was also quoted in United States v. Wade, 388 U.S. 218, 228, 87 S.Ct. 1926, 18 L.Ed.2d 1149 (1967). See also Wehrle v. Brooks, 269 F.Supp. 785, 792 (W.D.N.C.1966), aff'd, 379 F.2d 288 (4th Cir.1967) (“[positive identification of a person not previously known to the witness is perhaps the most fearful testimony known to the law of evidence”); accord Webster, supra,, 623 A.2d at 1204 (quoting Wehrle)-, Jackson v. Fogg, 589 F.2d 108, 112 (2d Cir.1978) (“convictions based solely on testimony that identifies a defendant previously unknown to the witness are highly suspect”) (internal brackets omitted).
Benn I, 801 A.2d at 145 n. 11.
In light of “the possibility that a mistaken identification may send an innocent person to prison,” As.H., 851 A.2d at 460 n. 7, appellate judges have an obligation to draw upon their own experience and common sense to ensure that a verdict of “guilty” based on eyewitness identifications by strangers is in keeping with the facts. See Crawley, 320 A.2d at 313. The increasing judicial acceptance of expert testimony on this subject is grounded in the concern that an innocent defendant may be found guilty. In this case, however, the trial judge indicated that he found such concerns groundless. When Benn’s attorney cited this court’s statement in Benn I that the testimony of stranger eyewitnesses is “proverbially untrustworthy,” the judge, remarked:
*1290THE COURT: That seems one of the — to be one of these revealed truths in the law that somebody said about 100 years ago and it’s picked up in opinion after opinion after opinion. And we don’t know that that’s so. I don’t know that’s so. And I don’t know that Penrod knows that it’s so either.
Subsequently, the judge reiterated the same theme:
But I don’t — I’m not convinced in any way when an appellate court says stranger-to-stranger identification has some danger of misidentification. I think, you know, it just doesn’t — I don’t know why they are in any position to make that statement case after case, year after year, century after century, if you will.
The hazards of wrongful convictions resulting from mistaken identifications of strangers are not imaginary. As I have noted, Professor (later Justice) Felix Frankfurter wrote more than eighty years ago that these dangers had been established “by a formidable number of instances in the records of English and American trials.” FrankfurteR, The Case of Sacco and Vanzetti (1927). Justice Frankfurter’s thesis remained valid forty years later, when the Supreme Court quoted his work in Wade, 388 U.S. at 228, 87 S.Ct. 1926, and the number of miscarriages of justice resulting from faulty identifications had grown even more “formidable.” The problem remains with us today; indeed, the availability in recent years of DNA evidence has demonstrated the extent and pervasiveness of the problem:
Every major study of wrongful convictions in the last decade has concluded that eyewitness misidentification is the most common cause of wrongful convictions in America. Of the first 200 DNA-based exonerations, 79% of the cases involved an eyewitness misidentification.
Professor Cynthia E. Jones, The Right Remedy for the Wrongly Convicted: Judicial Sanctions for the Destruction of DNA Evidence, 77 Fordham L.Rev. 2893, 2929 (May 2009) (footnote omitted); see Jennifer Thompson-Cannino and Ronald Cotton, Picking Cotton: Our Memoir of Injustice and Redemption (St. Martin’s Press 2009).
The decisions of the Supreme Court and of this court recognizing the potential for misidentification when the accused is a stranger to the witness are grounded in reality. They provide the legal context in which judicial discretion must be exercised at the trial court level. Given the fact that at the time of the hearing, expert testimony regarding eyewitness identification had seldom, if ever, been admitted in the Superior Court, and considering that no decision excluding such testimony had ever been reversed by this court, the trial judge’s unwillingness to admit such evidence in this case is certainly understandable. Nevertheless, in rejecting the appellate precedents which defined the legal context under which the issue arose,7 the trial judge did not exercise his discretion in conformity with the applicable legal standards.
V.
Towards the conclusion of the hearing regarding Professor Penrod’s proposed *1291testimony, the trial judge, having previously indicated that he disagreed with the concerns of this court and of other appellate courts regarding the reliability (or lack thereof) of stranger eyewitness testimony generally, disclosed that he was likewise unpersuaded by this court’s evaluation of the strength of the specific eyewitness testimony at Benn’s first trial. The judge stated:
I’m not impressed by [the court’s] analysis from — as far as the quality of the eyewitness identifications in this case either. That does not impress me at all one way or the other.
In volunteering this comment, the judge was obviously referring to the passage in Benn I, quoted in Part II of this opinion, in which this court stated, that “closer scrutiny places the strength of the [prosecution’s] case in substantial doubt.” 801 A.2d at 145.
As I have previously noted, the court’s assessment of the government’s case was a part of its holding that the judge’s erroneous evidentiary ruling at Benn’s first trial was not harmless. Further, the decision whether to admit expert testimony on behalf of the defendant depends in substantial part on the strength of the government’s evidence. “The opinions that rule that the exclusion of the expert testimony was or may have been error are typically those where there was little or no evidence to corroborate the eyewitness identification.” Hager, 856 A.2d at 1149 (quoting Commonwealth v. Santoli, 424 Mass. 837, 680 N.E.2d 1116, 1119 (1997)).
At the time that the judge ruled on the admissibility of Dr. Penrod’s proposed testimony, the evidence at Benn’s first trial constituted the only material available to him regarding the strength of the government’s evidence. The Benn I trial was before a different judge, and the trial judge in the present case had no more information regarding the demeanor of the witnesses than this court had when it decided Benn I. If the judge believed, as he plainly did, that the identification testimony was more compelling than this court thought it was, then that belief necessarily encouraged him to exclude the expert evidence more readily than he would have done if he had believed the evidence to be weaker. It therefore follows that, having declared that he disagreed with this court’s evaluation of the strength of the identifications, the judge excluded the proposed expert testimony on the basis of a view of the relevant facts and law which was at variance with an important part of the reasoning of the appellate court in Benn I. In other words, on the narrow point at issue, the judge applied his own legal standard, rather than the standard articulated by this court.
As we noted in Benn I, 801 A.2d at 145, it is often difficult to make an accurate assessment of the strength of a case when the judge making the assessment has not seen or heard the witnesses. This difficulty confronts a successor trial judge just as it does an appellate court. To be sure, reasonable people can read the same transcript and come to different conclusions as to the persuasiveness of the government’s evidence. In this case, however, the appellate court had addressed the point in some detail and, for reasons set forth in some detail in its opinion, it had concluded that the eyewitness testimony was less than overwhelming. Indeed, this was an important reason for reversal in Benn I. Under these circumstances, I do not believe that the trial judge, who had already revealed that he did not share the concerns of appellate courts generally regarding the reliability of eyewitness identifications of strangers, was free, in exercising his discretion, simply to reject this court’s reasoning in Benn I.
*1292VI.
This is not an easy case. At the time the judge made his decision, a ruling admitting expert testimony such as that proffered here by the defense would have been a rarity, and certainly a departure from the norm in the Superior Court. Moreover, notwithstanding the problems with the particular eyewitness testimony enumerated in Benn I, the government presented five witnesses, all members of the same family, none of whom had any apparent reason to lie, and all of whom ultimately made positive identifications of Benn. The mutual corroboration provided by the five identifications did not necessarily make this an overwhelming prosecution case, but their collective weight also should not be minimized. Nevertheless, with the record as it stands, I cannot be confident that the judge exercised his discretion in conformity with correct legal principles. Accordingly, I join my colleagues in remanding the case.
. In their brief, counsel for Benn argued that the trial judge erred in relation to each of the three issues raised in this opinion, but they asked for reversal of Benn's conviction rather than for a remand. Counsel did not couch their argument explicitly in terms of the judge’s failure to apply correct legal principles in the exercise of his discretion. The government might plausibly argue that the grounds for reversal relied on in the opinion of the court and in this concurring opinion have not been properly preserved, and that the government has not had the opportunity to brief these precise issues.
In my judgment, however, the basic points discussed in this concurring opinion are before the court, and the court is effectively granting “lesser included” relief to Benn. Although reasonable people might differ on the point, I conclude that Benn has preserved the underlying claim of abuse of discretion, and that although he has not made the precise arguments on which the remand is based, that is not dispositive. Cf. Yee v. City of Escondido, 503 U.S. 519, 535-36, 112 S.Ct. 1522, 118 L.Ed.2d 153 (1992).
. And, again, at the second trial.
. If a photograph of the real perpetrator was not in the array, then the witnesses might well have identified the person in the photo spread who most resembled him.
. See also In re As.I-J., 851 A.2d 456, 462 (D.C.2004):
Where ... the police consider an individual to be a possible perpetrator and a witness makes an initially ambiguous identification, there may develop a process of mutual bolstering which converts initial tentativeness into ultimate certainty. In re Dwayne W., 109 Daily Wash. L. Rptr. 1901, 1906 (Super. Ct. D.C. 1981). ‘‘The victim relies on the expertise of the officer and the officer upon the victim's identification." Id. (quoting [Judge] Nathan Sobel, Eyewitness Identification, Legal and Practical Problems 12 (1972 & Supp. 1981)).
. Dyas v. United States, 376 A.2d 827 (D.C.1977).
. In Dyas, we identified criteria for the admissibility of expert testimony
(1) the subject matter "must be so distinctively related to some science, profession, business or occupation as to be beyond the ken of the average layman(2) "the witness must have sufficient skill, knowledge, or experience in that field or calling as to make it appear that his opinion or inference will probably aid the trier in his search for truth"', and (3) expert testimony is inadmissible if “the state of the pertinent art or scientific knowledge does not permit a reasonable opinion to be asserted even by an expert.”
376 A.2d at 832 (quoting McCormick on Evidence, § 13, at 29-31 (E. Cleary 2d ed. 1972)).
. It is true that the opinions which the trial judge questioned did not deal with the admissibility vel non of expert testimony. However, a major factor in determining whether such testimony should be admitted is the strength of the government’s case, for the less corroboration that there is for an identification by a stranger, the stronger the argument is for receiving expert evidence. See, e.g., Hager v. United States, 856 A.2d 1143, 1149 (D.C. 2004). The authorities which were the subject of the judge's skepticism were thus indisputably relevant to the issue, before us.