concurring in the judgment:
I find this case considerably more difficult than my colleagues do, for the argument by the defense for the proposition that the excluded expert testimony was relevant to Smiths’ guilt or innocence or to reasonable doubt strikes me as counter-intuitive, and the notion that its exclusion constituted prejudicial error is, at least for me, somewhat problematic. Nevertheless, because the prosecution was permitted to introduce expert testimony on an issue which, though remote from actual guilt or innocence, was treated by the government as being of great importance; because the defendant was precluded from attempting to refute this testimony by calling his own expert; because, as we have recognized, expert testimony may sometimes “be given undue deference by jurors,” Smith v. United States, 389 A.2d 1356, 1359 (D.C. 1978), cert. denied, 439 U.S. 1048, 99 S.Ct. 726, 58 L.Ed.2d 707 (1978); and because this danger is especially acute when one side calls an expert witness and the other cannot and does not, I am constrained to agree that Smith’s convictions must be reversed.
The principal theme of George Wynn’s proposed expert testimony for the defense was that the “powder” method utilized by the police to develop any potential fingerprint evidence was inferior to the “Super Glue” method contemplated by the MPD’s Special Order and favored by Wynn. Thus, according to the defense, the police investigation was defective, and the police may have failed to uncover relevant evidence. Therefore, defense counsel argue, Wynn’s evidence would have generated or supported a reasonable doubt of Smith’s guilt.
Wynn could not have testified, however, that if the Super Glue method had been used, this would have made it more likely that Smith would be exonerated, for we do not know, and nor could he, if any print on the pistol or the bicycle (whether or not such a print could have been retrieved and identified by better police work) was that of Smith or of someone else. The trial judge was of the opinion that under these circumstances, “what could have been done, what might have been determined ...” lacked probative value, because there was “no evidence to suggest that whether [the police] used powder or used Super Glue ..., the result would have been different-” The judge added that “actually, I think it’s going to confuse the jury.”
“It is quintessential^ the function of the trial judge to determine whether expert testimony is likely to be helpful to the jury.” Benn v. United States, 978 A.2d 1257, 1279 (D.C.2009) (citations omitted). “It is well established that a trial judge has broad discretion to admit or exclude expert testimony, and that a decision ei*1200ther way should be affirmed unless it is manifestly erroneous.” Oliver v. United States, 711 A.2d 70, 73 (D.C.1998) (per curiam); Spencer v. United States, 688 A.2d 412, 417 (D.C.1997). This deferential standard of review is important, and I have more difficulty than my colleagues apparently do with the notion that the judge’s ruling can fairly be characterized as “manifestly erroneous.” A reasonable person could, I think, agree with the judge that Wynn’s proposed testimony would make it neither more likely nor less likely that Smith had possessed the pistol. For this reason, in my view, the relevance of the proposed evidence, which depends entirely on inferences which can allegedly be drawn from perceived shortcomings in the police investigation, is not at all obvious, especially when there is no evidence that these shortcomings, if that is what they were, are likely to have made any difference to the outcome of the trial.
Moreover, even assuming, arguendo, that the proposed testimony did have some attenuated bearing on the question of Smith’s guilt or innocence, “[t]he trial judge must balance the probative value of the evidence against the risk of prejudicial impact, including the risk of jury confusion from a trial within a trial, and may exclude marginally relevant evidence if it will distract the jury from the issue in this case.” Hager v. United States, 791 A.2d 911, 914 (D.C.2002) (emphasis added; citation and internal quotation marks omitted). The trial court may therefore take reasonable steps to avoid the injection of “collateral and confusing questions into the proceedings.” Bennett v. United States, 876 A.2d 623, 635 (D.C.2005), cert. denied, 546 U.S. 1123, 126 S.Ct. 1134, 163 L.Ed.2d 914 (2006) (citation and quotation marks omitted). In this case, if Wynn had been permitted to testify, then the prosecution would doubtless have attempted to rebut his testimony, with the potential consequence that the jury’s attention might well have been distracted from the question of Smith’s innocence or guilt, which Wynn’s testimony could not and would not have addressed, to the merits of the competing methods of fingerprint retrieval and identification.
Both before the trial court and on appeal, Smith’s attorneys have attempted to make the methodology used by the police the focus of the case. Revealingly, defense counsel argued to the trial judge that “the issue of the case is the investigative work by the police.” This was, no doubt, an ingenious tactic, but we should not lose sight of the fact that it was Smith who was on trial, and not the investigative techniques of the police. It surely should not be in every case, to quote Justice (then Judge) Cardozo, that “[t]he criminal is to go free because the constable has blundered.” People v. Defore, 242 N.Y. 13, 150 N.E. 585, 587 (1926).
Even if the trial judge’s exercise of his discretion was unreasonable, it is not as clear to me as it apparently is to the majority that Smith suffered any appreciable prejudice. To be sure, Wynn’s testimony would have reinforced the defense contention, with which an impartial juror could reasonably agree, that the police investigation was not as thorough as it should have been, but much of this was established on cross-examination and by the MPD’s Special Order. The record provides no reason to believe that the use of the Super Glue method, or a more thorough investigation generally, would have retrieved a hypothetical real culprit’s fingerprint (rather than Smith’s) and would thus have made acquittal more likely, or that the jurors would have determined that any additional evidence that the Super Glue method might have produced was likely to be exculpatory.
*1201The failure to investigate adequately may in appropriate cases give rise to an inference favorable to the defense, for putting on a weak case, when a strong one could have been presented, is assuredly something that the trier of fact is not required to ignore. See, e.g., Greer v. United States, 697 A.2d 1207, 1212 (D.C.1997). Suppose, for example, that five people had witnessed a shooting. Four of them were members of the clergy and of unimpeachable character. The fifth was a near-sighted drug addict with a long criminal record, which included convictions for perjury. Suppose further that the prosecution did not interview any of the righteous and upstanding citizens, but relied solely on the addict’s identification of the defendant. In such a case, the jury could surely infer, from the severely flawed investigation, that reliance on evidence gathered by the police during that investigation was unwarranted, and that there was a reasonable doubt of the defendant’s guilt. But use of the powder method rather than Super Glue, if blunder at all, is not a comparably colossal kind of blunder, and there is no basis upon which one can reasonably infer in this case that the kind of investigation favored by Wynn would have made it more likely that any additional evidence would have pointed to Smith’s innocence.
Smith describes the case for the prosecution as “thin” and “weak,” and the majority implicitly agrees. In my view, it is not quite as simple as that. To begin with, it is “never easy” to assess the closeness of a case from our “sheltered appellate perch.” Ford v. United States, 616 A.2d 1245, 1250 (D.C.1992) (citations omitted). “Since appellate courts cannot view the demeanor of the witnesses, it is difficult to determine the strength of the case, because the result turns in large part on ... credibility.” Brooks v. United States, 599 A.2d 1094, 1102 n. 17 (D.C.1991) (cita tion and internal quotation marks omitted). The legal sufficiency of the case for the prosecution has not been challenged, and the government presented substantial evidence of Smith’s guilt. When the officers arrived at the scene, Smith was the only one of the young men who matched the broadcast description of the suspect. Although there were some inconsistencies in the accounts provided by the officers, there was testimony that Smith straddled or sat on and held the bicycle in question. After the police arrived, Smith dropped or left the bicycle and walked away rapidly, ignoring several police orders to stop. One officer thought that Smith was “about to run,” and another described him as “running away.” The pouch on the bicycle from which Smith separated himself with such apparent dispatch contained a loaded pistol. If the jurors believed the above-described police testimony — and it is their assessment of credibility (including reconciliation of contradictions and inconsistencies) that matters, not ours— then there was ample evidence of Smith’s having carried the pistol on his bicycle and of his consciousness of guilt. Although the prosecution was not able to present eye-witness testimony to the effect that the weapon was seen in Smith’s possession, circumstantial evidence may often be more persuasive than direct evidence. See, e.g., Janifer v. Jandebeur, 551 A.2d 1351, 1352 (D.C.1989).
Nevertheless, the exclusion of Wynn’s proposed evidence, after expert testimony on the same subject had been presented and emphasized by the prosecution, created what I must agree was an unfair imbalance. Although Wynn’s testimony would not have shown that any fingerprint evidence that could have been retrieved and analyzed by use of the Super Glue method would have been favorable to Smith, the government focused heavily during its *1202presentation of evidence, and again at closing argument, on the difficulties allegedly encountered in lifting and identifying fingerprints. Because, according to the prosecution and its witnesses, such evidence could be recovered from a handgun only in rare cases, the failure of the government to produce any fingerprint evidence here did not, so the government insisted, appreciably weaken the case for the prosecution.
Had Wynn been permitted to testify, he would have provided expert testimony directly contradicting a contention to which the prosecution attached great importance. Because, as the government points out in its brief, “there is an authoritative quality that surrounds expert opinion,” Smith, 389 A.2d at 1359, and because expert testimony may “be given undue deference by jurors and ... could thereby usurp the truth seeking function of the jury,” id., it is difficult to have confidence in the verdict when the prosecution has been permitted to adduce expert evidence regarding a point that it regards as extremely important while the defense has been precluded from responding in kind. Accordingly, although my analysis differs somewhat from that of my colleagues, I agree that Smith’s convictions must be reversed and that a new trial must be ordered.