We granted rehearing en bane in this case to consider the standards governing the admissibility of evidence proffered by a criminal defendant that another person or persons committed the crime alleged.1 The issue arises at the intersection of the defendant’s constitutional right to an opportunity to present a complete defense, Crane v. Kentucky, 476 U.S. 683, 690, 106 S.Ct. 2142, 2146, 90 L.Ed.2d 636 (1986), and the obligation of the trial court preliminarily to determine the relevance of proffered evidence and weigh its probative value against the potential it creates for undue prejudice. The issue arises specifically because of continued uncertainty, reflected in the trial judge’s ruling in this case, over the meaning and continued legitimacy of a principle enunciated by this court in Brown v. United States, 409 A.2d 1093 (D.C. 1979). See also Beale v. United States, 465 A.2d 796, 803 (D.C. 1983), cert. denied, 465 U.S. 1030, 104 S.Ct. 1293, 79 L.Ed.2d 694 (1984). Under the so-called Brown-Beale principle, “before evidence of the guilt of another can be deemed relevant and thereby admissible, the evidence must clearly link that other person to the commission of the crime.” Id. at 1097 (emphasis added).

This principle has long seemed in tension with the standard of relevance applied generally in this jurisdiction, which defines relevant evidence simply as “that which tends to make the existence or nonexistence of a fact more or less probable than would be the case without that evidence.” Punch v. United States, 377 A.2d 1353, 1358 (D.C. 1977), cert. denied, 435 U.S. 955, 98 S.Ct. 1586, 55 L.Ed.2d 806 (1978); see also Fed.R.Evid. 401. That the prior effort of a division to harmonize the “clear link” principle and the *3usual standard of relevance, see Johnson v. United States, 552 A.2d 513, 516 (D.C. 1989), has not settled the issue is apparent from this case and from the representation of the Public Defender Service, not seriously disputed by the government, that the former principle continues to “create[ ] confusion for trial judges.” Although we recognize the difficulty of laying down principles concrete enough to guide meaningfully decisions on the admissibility of nonprivileged, nonhear-say evidence, we undertake a modest effort at clarification here.

We conclude that the phrase “clearly linked” is unhelpful and should be discarded from our lexicon of terms governing the admissibility of third-party perpetrator evidence. We hold that there is only one standard of relevance, that set forth in Punch, supra, as explicated below. At the same time, as the division correctly recognized in this case, evidence of motivation of a third party to commit the crime charged risks distracting the jury from the issue of this defendant’s guilt or innocence, and in applying the relevance standard the judge may properly take account of that danger. We further hold, under the standards we announce and on the facts proffered here to support the claim of third-party perpetration, that the trial judge should not have excluded the evidence.

I. The Proffer

The evidence presented by the government, which resulted in appellant’s conviction for (inter alia) armed first degree murder, is summarized in the division opinion, supra note 1. While the proof of guilt was substantial, including the testimony of three eyewitnesses that appellant chased down and fatally shot Deborah Davis in Southeast Washington on the night of July 26, 1990, the evidence identified no motive for appellant to have committed the shooting.

The government and the defense each filed in limine motions (the government’s in opposition) concerning appellant’s desire to present evidence that a third person or persons had done the shooting without appellant’s complicity. The defense proffered (and the government did not dispute) that Deborah Davis, James Bias and Freddie Artis had committed an armed robbery together in the District of Columbia during the third week of June, 1990. All had been arrested and charged with the offense, but only Bias was detained. Upon her release from custody, Davis had begun cooperating with the government in its investigation of the crime. According to the defense proffer, on June 26, 1990, Davis was kidnapped from the District of Columbia and taken to Maryland by Artis, who was joined at some point in the abduction by Edward Huff. Artis told Davis that he was going to kill her because he feared she was “snitching.” He and Huff then stabbed her, shot her, and left her for dead in a wooded area in Prince George’s County.

Davis survived the murder attempt and identified Artis and Huff as her assailants. Artis, charged in both Maryland and the District, was arrested on June 28,1990 (Bias remained in jail), but Huff, though charged in Maryland, was still at liberty. Evidently cooperating with the government’s investigation of the abduction, Davis testified before the grand jury on July 26,1990. Later that day, according to the defense proffer, she called her mother and said, “They’re after me, They are going to get me. The word is out on the street.”2 That same night Davis was shot dead outside a housing project. According to the proffer, when the male assailant shot Davis he was heard to say, “You won’t tell this.”3 The defense contended that appellant was not connected with Artis, Huff or Bias, and in fact no evidence linking him to them was adduced at trial. The government, on the other hand, represented that a photograph of Huff had been placed in a photo array and shown to several witnesses, none of whom identified him as the shooter.

II. The Trial Court’s Ruling

The trial judge found the evidence that Huff had a motive to kill Davis “compelling,” *4but reasoned — in apparent reliance on the Brown-Beale principle — that the ultimate issue is “whether or not Mr. Huff or any other person is clearly connected to the shooting itself’ (emphasis added). Answering in the negative, the judge found significant the absence of any proffered evidence: (1) that anyone who had been able to view the shooter identified Huff as the shooter even though some had looked at a photo array that included his picture; (2) that Huff was anywhere near the scene of the murder that night or knew where Davis could be located at the time; and (3) that Huff knew that Davis had testified in the grand jury that day. The judge found the words assertedly spoken by the killer (“You won’t tell this”) “not so distinctive that they necessarily tie Mr. Huff to the shooting” (emphasis added). The judge reasoned, in sum:

The inherent ambiguity of this evidence supposedly linking Mr. Huff to the murder would not ... tend to create a reasonable doubt that the defendant who was apparently known by many of the eyewitnesses did not [sic ] commit the offense.

Furthermore, given the quality of the evidence proffered and even assuming its relevance, its probity ... in clearly linking Mr. Huff to the murder is so weak and the resulting potential of the evidence to confuse and mislead the jury so great that the Court concludes that on the state of the present record the defense should be precluded from presenting any evidence to show that Mr. Huff may have killed the decedent_ [Emphasis added].

III. Discussion

A.

The trial judge’s analysis reflects the lingering notion in our decisions that relevance means something different as regards evidence that a third party committed a crime than it does in other contexts. We now make clear that it does not. A division of this court endeavored to say so in Johnson v. United States, supra, by defining “[w]hat we mean by ‘clearly link,’ as used first by this court in Brown, supra, 409 A.2d at 1097.” Johnson, 552 A.2d at 516. That phrase, the division explained, means that there must be “proof of facts or circumstances which tend to indicate some reasonable possibility that a person other than the defendant committed the charged offense.” Id. The “focus” of the standard is not on the third party’s guilt or innocence, but on “the effect the evidence has upon the defendant’s culpability,” and in this regard it “need only tend to create a reasonable doubt that the defendant committed the offensé.” Id. at 517 (emphasis in original). By thus employing the verb “tend,” the division explicitly tied the relevance standard to the usual meaning of that concept. See Punch, 377 A.2d at 1358 (relevant evidence is “that which tends to make the existence or nonexistence of a fact more or less probable” (emphasis added)); Fed.R.Evid. 401 (relevant evidence is “evidence having any tendency to make the existence of any fact ... more probable or less probable than it would be without the evidence”). Correlatively, the division stated, “[t]here is no requirement that the proffered evidence must prove or even raise a strong probability that someone other than the defendant committed the offense.” Johnson, 552 A.2d at 517.

Sitting en banc, we now adopt the “reasonable possibility” formulation of Johnson and its conclusion that relevance here means what it generally does in the criminal context, requiring a “link, connection or nexus between the proffered evidence and the crime at issue.” Id. at 516. In the roughly reverse situation where the government seeks to introduce proof of other criminal acts by the defendant, the evidence must relate to an issue other than criminal propensity, Drew v. United States, 118 U.S.App. D.C. 11, 16, 331 F.2d 85, 90 (1964), but the test remains that of relevance.4 Similarly, where a party seeks to impeach a witness with a prior bad act not reduced to a conviction, the act must “bear[ ] directly upon the veracity of the witness in respect to the issue involved in the trial,” Kitchen v. United States, 95 U.S.App.D.C. 277, 279, 221 F.2d 832, 834 (1955), cert. denied, 357 U.S. 928, 78 *5S.Ct. 1378, 2 L.Ed.2d 1374 (1958), but the test is still relevance. No different rule than the Punch standard, explicated by Johnson’s reasonable possibility formulation, should govern the relevance decision in the case of evidence of third-party perpetration. That standard insures the exclusion of evidence that “is too remote in time and place, completely unrelated or irrelevant to the offense charged, or too speculative with respect to the third party’s guilt.” Johnson, 552 A.2d at 516. To the extent our decisions in Brown, Beale, and later cases, e.g., Watson v. United States, 612 A.2d 179, 182 (D.C. 1992), impose a more exacting standard of relevance, we disavow them.

A requirement that evidence “tend to indicate some reasonable possibility that a person other than the defendant committed the charged offense,” Johnson, 552 A.2d at 516 (emphasis added), sufficiently accommodates the concern voiced by the division majority here, that surmise as to third-party responsibility for a crime risks misleading the jury by distracting it from the issue of whether this defendant is guilty or not. First, the division was correct that a mere showing that another person possessed a motive to harm the victim as strong as the defendant’s, even stronger, usually will “not [be] sufficient to meet the foundation for admissibility ... which requires a nexus between the proffered evidence and the charged crime.” Winfield I, 652 A.2d at 608.5 Thus, we reaffirm the specific holding of Beale, supra, as summarized in Winfield I, “that a defendant’s proffer of evidence that other individuals had even stronger motives to murder the victim than the accused [is] insufficient, without more, to establish the [required] link to the offense charged_” Id. at 612 (emphasis added). Simple proof of motivation of others to commit the crime ordinarily does not create a “real possibility” that any of them was the perpetrator. It follows from this, as the division recognized, that the trial judge ordinarily may exclude evidence of third-party motivation unattended by proof that the party had the practical opportunity to commit the crime, including at least inferential knowledge of the victim’s whereabouts. See cases cited in footnote 5 of the division majority opinion, 652 A.2d at 613.

Second, as our decisions have repeatedly emphasized, the determination that evidence is relevant does not exhaust the trial judge’s responsibility in deciding whether to admit it. The judge must also balance the probative value of the evidence “against the risk of prejudicial impact.” Punch, 377 A.2d at 1358. See Johnson, 552 A.2d at 518 (“the probative value of the evidence must outweigh any tendency to create undue prejudice”). In the context of third-party perpetrator evidence, this means the trial judge will have discretion to exclude marginally relevant evidence creating the danger that proof of prior dealings or hostility between the victim and third persons will distract the jury from the issue in this case. That risk of confusion may be exacerbated if the government, in response to the defense proffer, asserts the need to present “rehabilitative” evidence disassociating the victim from other persons allegedly harboring a motive to harm her. The judge must retain full authority to prevent this sort of trial-within-a-trial. For that reason, too, even when admitting evidence of third-party perpetration, the judge preserves wide latitude to limit the number of witnesses and the extent of their testimony bearing upon the issue.

B.

Given the uncertainty in our decisions adverted to earlier, it is not easy to fault the trial judge’s exercise of discretion in this case. Applying the standards discussed above, however, we hold that the judge should have admitted the evidence in dispute. The peculiar aggregation of circumstances proffered convinces us that to bar evidence of the potential responsibility of Huff or a confederate (unrelated to Winfield) for the killing would undermine appellant’s right to present a defense.

*6Decisive for the trial judge was the absence of evidence placing Huff at or near the murder scene at the relevant time (Artis and Bias could not have been there since they were in jail), and evidence that Huff knew Davis’s whereabouts at the time or that she had testified in the grand jury that day. The judge thus concluded, and the division majority agreed, that the proffer boiled down to a showing that Huff had a motive to kill Davis — even a “compelling” one — and had attempted to do so in the recent past. This was too insubstantial a link to the murder, the judge and the division thought, to outweigh the potential confusion in setting before the jury Davis’s past and her violent relationship with parties not otherwise identified with the crime.

In our judgment, the combined force of the proffered circumstances requires admission of the evidence. In its brief to the en banc court, the Public Defender Service summarizes these circumstances as follows:

First, Bias and Artis, who had committed an armed robbery with Deborah Davis and had been arrested and charged with the crime, had the motive to silence Ms. Davis to prevent her from testifying for the prosecution against them. Second, Artis, in Huffs presence, threatened to kill Ms. Davis to prevent her from “snitching.” Third, Huff and his cohort had recently [a month before the murder] acted on the motive and threats by abducting Ms. Davis, stabbing and shooting her in an attempt to cause her death. Fourth, the timing of the murder — committed within hours of Ms. Davis’s appearance before a grand jury investigating her abduction— suggested that the reason for the murder was her cooperation with authorities. Fifth, Huff had the opportunity to commit the crime in that he was at liberty and the killing occurred in the public courtyard of a housing project where Ms. Davis was easily accessible to those who [might] want to harm her. Sixth, the method of the killing, by gunfire, was also one of the methods used in the attempt on Ms. Davis’s life by Artis and Huff. Finally, the very res gestae of the crime, the killer’s words as he accomplished the murder, directly demonstrated that this killing was about “snitching” — the motive held by Huff and his cohorts but not [proven to be held] by Mr. Winfield. [Citations and intervening argument omitted].

In the aggregate, these facts if proven would demonstrate more than “the coincidence of a motive, even a strong one, and prior assaults on the victim by a third party_” Winfield I, 652 A.2d at 613. The recent shooting stemmed from the same asserted motive of stifling Davis’s cooperation with the police — a motive intensified by the addition of kidnapping and attempted murder to the crimes for which Huff and Artis were being investigated. And while, as the trial court observed, expressions of an intent to kill for “snitching” are “really not so unusual” these days, the killer voiced that motivation on the very day Davis had testified against Huff and Artis in the grand jury. Lastly, having abducted and assaulted Davis before, Artis and Huff had demonstrated their ability to find her whether or not they actually knew her whereabouts. All told, as Judge Schwelb stated at the division level, exclusion of the evidence meant that the jury never learned “that the prosecutor’s case against Winfield rested on what an impartial trier of fact might reasonably view as a very unusual coincidence indeed.” Winfield I, 652 A.2d at 615 (Schwelb, J., dissenting).

The trial judge applied himself conscientiously to the admissibility issue. He received written in limine motions6 to admit and exclude the evidence, heard oral arguments and evaluated the proffered facts, and set forth his reasoning on the record. Even allowing for his reliance on a mistakenly heightened standard of relevance (Brown-Beale), his ruling was not capricious. Nevertheless, three considerations demonstrate why in our view the trial court must resolve close questions of admissibility in this setting in favor of inclusion, not exclusion. First, a substantial proffer that a third person com*7mitted the offense implicates the defendant’s constitutional right to “a meaningful opportunity to present a complete defense,” Crane v. Kentucky, 476 U.S. at 690, 106 S.Ct. at 2146. Second, evidence of the type appellant seeks to admit here — recent assaults against the victim stemming from identical motivation — has been considered highly probative by this court in the reverse setting of admission of “other crimes” evidence to support guilt. E.g., Hill v. United States, 600 A.2d 58, 61 (D.C. 1991); Green v. United States, 580 A.2d 1325, 1328 (D.C. 1990). Unduly restricting admission of third-party perpetrator evidence would raise concerns of unequal treatment. Finally, in a previous en banc decision, we cautioned against “excessive mistrust of juries” in the evaluation of relevancy. Allen v. United States, 603 A.2d 1219, 1224 (D.C. 1992) (en banc) (quoting Riordan v. Kempiners, 831 F.2d 690, 698 (7th Cir. 1987), cert. denied, 505 U.S. 1227, 112 S.Ct. 3050, 120 L.Ed.2d 916 (1992)). Beyond the point, made earlier, that the trial court retains broad discretion to prevent the cumulation of third-party perpetrator evidence, sifting the relevance of that evidence is largely about drawing commonsense inferences from uncomplicated facts, something we regularly entrust to juries. We should do so here as well, so long as the balancing of probative value versus prejudice does not clearly favor exclusion.

Reversed and remanded for a new trial.

separate statement:

As a member of the division in this case, I voted to affirm the conviction because the trial judge committed no error when he applied the Brown/Beale1 “clearly linked” test. The trial court was bound to apply that test, as was this court. M.A.P. v. Ryan, 285 A.2d 310, 312 (D.C. 1971) (prior decision of this court must be followed by a division of this court and may be overruled only by this court sitting en banc). Although I am still of the view that the trial judge correctly applied the test that then governed, I agree with the en banc court, as set forth in the opinion by Judge Farrell, that the “clearly linked” test is too restrictive and should be abandoned. Therefore, I endorse the new test adopted today. When that test is applied to the facts of this case, I am satisfied, for the reasons stated by the majority, that the judgment of conviction must be reversed. With that understanding, I join the opinion.

dissenting:

In my view, the majority has obscured, rather than clarified, the requirements for determining relevance for purposes of admitting third-party culpability evidence by injecting the “reasonable possibility” rule into the nexus requirement. The substitution of the “reasonable possibility ” concept for the “clearly linked ” formulation, the apparent source of past confusion in our caselaw, creates new ambiguity, in my opinion. Moreover, as applied by the en banc court in this case, this “reformulation” of the relevance test diminishes, if not eliminates, the nexus requirement in spite of the court’s expressed intention to the contrary. However, if the court intends, as it otherwise states, to retain the nexus requirement for admissibility and to preserve the authority of the trial court to “exclude evidence of third-party motivation unattended by proof that the party had the practical opportunity to commit the crime,”1 on the factual proffer actually before the trial court, as opposed to that supplied by impermissible inferences and speculation, it cannot be concluded that the trial court abused its discretion in excluding the evidence. For these reasons, the reasons stated in the division opinion of the majority in Winfield v. United States, 652 A.2d 608 (D.C. 1994), vacated, 661 A.2d 1094 (D.C. 1995) (Winfield I), and those which follow, I respectfully dissent from the opinion of the court.

I. The Proffer and the Trial Court’s Decision

In reviewing the trial court’s ruling, the majority relies upon a characterization of the *8proffer in the brief of amicus curiae, rather than the facts as presented in the trial court.2 One fact critical to the trial court’s decision, but omitted from the facts upon which the en banc majority relies, was that the witness who allegedly heard the killer state on the night of the murder, “You won’t tell this,” examined a photo array which included Huffs picture but did not identify Huff as the person who made the statement. This was an important factor which guided the trial court in the exercise of its discretion to exclude the evidence.3 In ruling on the motion in limine, the trial court, in its careful decision, made this point explicitly, stating:

A defense witness who viewed a photographic array which included the photograph of Mr. Huff did not identify him as the person who allegedly said, open quote, “You won’t tell this,” close quote.

Although appellant’s counsel contended that the photograph of Huff was not a recent one, she declined the court’s suggestion that she make arrangements to show the witness a more recent photograph. Later, in its ruling, the court again stated as a basis for its ruling that “a defense witness claimed that a person other than Mr. Huff said to [the victim], ... “You won’t tell this.’ ”

Other inadequacies in the proffer, including the ambiguity in the statement attributed to the shooter, also informed the trial court’s discretion. The shooter’s meaning in saying “[y]ou won’t tell this” is not clear. The shooter could have meant that the victim would not be able to tell on him as she had on others. In that sense, the statement did not reflect that the shooter’s motive for the murder was related to the motives of Artis, Bias and Huff. The trial court found the words not to be “so distinctive that they necessarily tie Mr. Huff to the shooting by relating back to the words spoken by Mr. Artis in Mr. Huff’s presence.” At the time of the prior shooting, these words were not used. Instead, Artis told Davis that she would be killed because she was the only one who could testify against him concerning an earlier robbery involving Artis and Bias in which Huff was not involved, according to the proffer. The trial court determined that the ambiguity of the proffered evidence including the shooter’s statement, was such that it would not “tend to create a reasonable doubt that [Winfield] who was apparently known by many of the eyewitnesses did not commit the murder.” See Johnson v. United States, 552 A.2d 513, 517 (D.C. 1989) (Third-party culpability evidence “need only tend to create a reasonable doubt that the defendant committed the offense.”) (emphasis in original). The trial court also determined that, in light of the ambiguity of the statement, the probative value of the evidence was outweighed by the tendency of the evidence to mislead the jury. See Brown v. United States, 409 A.2d 1093, 1097 (D.C. 1979).

However, the majority posits that Huff had the ability to locate the victim, Ms. Davis, on the night of the shooting, a fact which appellant neither proffered nor contended in the trial court. Moreover, the record is devoid of any evidence that would lead a reasonable factfinder to conclude that Huff knew how to locate Davis that night. Appellant did not claim that either Huff or the victim lived in the area where she was gunned down or that either ever had been there before. Appellant did not identify the area from which Davis was abducted earlier. In the motion in limine, the proffer states that it was appellant, not Huff, who lived approximately one-half block from the basketball court where Ms. Davis was killed.

In making its ruling, the trial court accurately took into account the absence of a proffer that Huff was near the scene of the crime or knew where Ms. Davis could be located that night. The trial court stated in this regard:

There is no evidence that Mr. Huff was anywhere near the scene of the murder of July 26th, 1990. There is no credible evidence that Mr. Huff knew that Ms. Davis had testified before the Superior Court grand jury on July 26th. There is no *9evidence that Mr. Huff knew where Ms. Davis could be located on the night of her death.

Further, as stated in the majority opinion in Winfield I:

[Although there was evidence that Huff had a motive to commit the crime and a proffer that Huff had committed other acts of violence against the victim on another occasion, no evidence was proffered tending to show that Huff was implicated in the commission of the offenses for which appellant was on trial. Appellant conceded in the motion that he was not connected to Artis, Huff, and Bias.... Not a single person who witnessed the crime, even those who viewed photo arrays with Huffs picture in it, identified him either as the perpetrator of the crime or as someone who was even in the area of the crime scene at any time that night.

(Emphasis added.) Id. 652 A.2d at 613. Therefore, consistent with our prior opinions, the trial court found lacking the requisite nexus between Huffs motive and prior crime and the crime charged.4

The majority supplies the critical fact missing from the proffer in order to find the requisite nexus to Davis’ murder. It does so by assuming that “having abducted and assaulted Davis before, Artis and Huff had demonstrated their ability to find her whether or not they actually knew her whereabouts.” Majority opinion at 4. The flaw in this assumption is that it has no factual basis in the record to support it. Indeed, it is inconsistent with the facts before the trial court when it made its ruling. The proffer, as pointed out in the majority opinion at 5, was that

Davis was kidnapped from the District of Columbia and taken to Maryland by Artis, who was joined at some point in the abduction by Edward Huff.5

(Emphasis added.) Thus, even at the time of the kidnapping, the proffer shows only Artis’ knowledge of Davis’ whereabouts one month earlier. Appellant’s counsel conceded that she was not claiming that Artis and Bias, other individuals allegedly involved in the prior crimes, were responsible for Davis’ murder because both were in jail at the time. Therefore, the parties and the trial court focused upon the possibility of Huffs culpability only. Although identifying Huff as the remaining likely suspect, appellant, unlike the en banc majority, did not proffer any facts tending to show that Huff knew how to find Davis.6

In finding that Huff also knew where to find the decedent the night of the murder, the en banc majority must first impute to Huff Artis’ knowledge, or speculate that Ar-tis told Huff where he had located Davis earlier. The majority then must assume either that the kidnapping and murder occurred in the same general area, or that armed with whatever knowledge Huff might have obtained from Artis, if any, Huff was able to trace the victim to the basketball court at the critical time when she met her death. The problem with these assumptions is that we have no information that Huff obtained any such information from Artis, or if he did, that he acted on it. The majority’s assumptions stretch well beyond the limits of inferences permissible by our jurisprudence, which requires that reasonable inferences be made from proof of other facts. See United Ins. Co. v. Nicholson, 119 A.2d 925, 927-28 (D.C. 1956). “Conclusions may be drawn by inference from established facts but not from other inferences, because conclusions so reached are merely conjectural.” Id.

*10The trial court relied upon legal principles extracted from Johnson, supra, 552 A.2d at 516-17, just as the en banc court does today. In that regard, the trial court adhered to the principles governing relevance for purposes of admissibility of third-party culpability evidence, including the rule “that relevance ... requires] a ‘link, connection or nexus between the proffered evidence and the crime at issue.’ ” See majority opinion at 5 (quoting Johnson, supra, 552 A.2d at 516); see also Winfield I, supra, 652 A.2d at 612. What accounts for the difference in the en banc court’s outcome can be attributed only to (1) the impermissible substitution of speculation for a proffer of facts, or (2) the “reasonable possibility” rule which the majority inserts into the nexus requirement. Having examined the former, I turn now to an examination of this new gloss which the majority places upon the nexus requirement.

II. The “Reasonable Possibility” Rule

The majority expresses its intention to adhere to central principles of our previously well-settled rules governing the admissibility of evidence purporting to show that another person committed the crimes charged. It preserves the evidentiary rule enunciated in our prior decisions, and applied by the panel majority in this case, that before evidence can be admitted that a person other than the accused committed the crime, there must be a “link, connection or nexus between the proffered evidence and the crime at issue.” See majority opinion at 4 (citing Johnson, supra, 552 A.2d at 516) (citations omitted); Winfield I, supra, 652 A.2d at 613. Likewise, the majority adheres to the rule that a third-party’s motive alone is not sufficient to establish the requisite link and that some other evidence tending to connect that person with the commission of the crime is required.7 Freeland v. United States, 631 A.2d 1186, 1189-90 (D.C. 1993); Shepard, supra note 3, 538 A.2d at 1118. These principles, which have guided our decisions in numerous cases over the years and which are consistent with the general rule elsewhere, should be preserved.8 However, in an effort to clarify the meaning of the nexus requirement and eliminate the confusion engendered by a “clearly linked” explanation found in our prior decisions, the court now reintexprets the “reasonable possibility” language from Johnson, supra. At the same time, the court disavows our prior decisions to the extent that they impose some “more exacting standard of relevance.”9 This reformulation of the relevance test attaches new meaning to Johnson’s “reasonable possibility” language, and it diminishes the well-recognized nexus requirement between the proffered evidence and the crime at issue as the foundation for admissibility of evidence that a person other than the accused committed the charged offense.10

In Johnson, this court used the “reasonable possibility” formulation in an effort to explain what we meant in earlier cases where we stated that the foundational requirement for the admission of third-party culpability evidence is that it be “clearly linked” to the commission of the charged offense. Johnson, supra, 552 A.2d at 516. Significantly, we retained in Johnson several other critical principles, including: (1) that the evidence must “establish[ ] the necessary link, connection or nexus between the proffered evidence and the crime”; (2) that the evidence “need only tend to create a reasonable doubt that the defendant committed the offense”; and (3) that our focus “is on the effect the evidence has upon the defendant’s culpability, and not the third-party’s culpability.” Id. at *11516-17 (emphasis in original). I do not understand the majority’s opinion to reject these tenets. However, the majority now holds that admissibility of third-party culpability evidence is determined by reference to the usual test of relevance articulated in Punch v. United States, 377 A.2d 1353, 1358 (D.C. 1977), cert. denied, 435 U.S. 955, 98 S.Ct. 1586, 55 L.Ed.2d 806 (1978), as “explicated by Johnson’s reasonable possibility formulation.”11 It is the latter part of the holding which creates the problem, in my view. First, it unnecessarily imposes an additional requirement to the relevance test, the meaning of which is not clear. Second, in spite of the en banc majority’s view that this formulation insures the exclusion of irrelevant and speculative evidence, this court has recognized previously the exact opposite view in Johnson, as reflected in the following passage from the opinion, which followed the court’s “reasonable possibility” explanation:

What we mean by “clearly link,” as used first by this court in Brown, supra, 409 A.2d at 1097, is proof of facts or circumstances which tend to indicate some reasonable possibility that a person other than the defendant committed the charged offense. This proof permits the admission of evidence which otherwise is generally excluded because it is too remote in time and place, completely unrelated or irrelevant to the offense charged, or too speculative with respect to the third party’s guilt.

Johnson, 552 A.2d at 516 (emphasis added).12 That acknowledged risk of admitting irrelevant evidence is heightened by today’s ruling.

Moreover, this approach is inconsistent with the majority’s efforts to harmonize the admissibility requirements for such evidence with the customary standard of relevance. See Punch, supra, 377 A.2d at 1358 (“[relevant evidence is that which tends to make the existence or non-existence of a fact more or less probable than would be the case without that evidence”) (citations omitted); accord, King v. United States, 618 A.2d 727, 728 (D.C. 1993). That standard requires generally that the evidence sought to be introduced have some connection with the defendant and the crime charged. Punch, 377 A.2d at 1358. Similarly, we have required consistently that, in order to be admissible, third-party culpability evidence must have “some link or nexus between [the] proffered events and the crimes involved in the case before the court.” Winfield I, supra, 652 A.2d at 613 (citing Johnson, supra, 552 A.2d at 516 and Beale, supra note 4, 465 A.2d at 803). Despite the “reasonable possibility” explication in Johnson, even in that case, this court did not relax the relevancy requirement which included a foundational proffer of “a set of facts or circumstances, which, in the aggregate, establishes the necessary link, connection or nexus between the proffered evidence and the crime at issue.” Johnson, 552 A.2d at 516.

Under the majority’s modified rule, it is apparently sufficient to proffer only the motive and prior crime of a third party against the victim,- provided the third party remains at large somewhere in the community, leaving any connection with the crime charged to mere conjecture. The majority’s approach casts an extremely wide net. Since being at large in the community can apply to virtually anyone, including others with a motive and a hostile history with the victim, the result of applying the new rule appears to be to eliminate the requirement for showing opportunity and presence, and therefore the nexus to the crime charged as a foundation for third-party culpability evidence. This opens the door for the introduction in a criminal trial of all manner of evidence about the criminal circles of which a victim might have been a part.

The difficulty with our prior decisions, as the majority points out, stems from the phrase “clearly linked,” found in Beale, supra note 4, 465 A.2d at 803 and Brown, supra, 409 A.2d at 1097, and I have no quarrel with *12the majority’s determination to discard it.13 However, there remains in our jurisprudence other principles, which the majority preserves, which are sufficient, in my view, to provide guidance for the trial court in meeting its obligation to assure that a defendant has a “ ‘meaningful opportunity to present a complete defense,’ ”14 and at the same time “accommodate other legitimate interests in the criminal trial process.” See Nelson v. United States, 649 A.2d 301, 304 (D.C. 1994) (quoting Chambers v. Mississippi, 410 U.S. 284, 294, 93 S.Ct. 1038, 1046, 35 L.Ed.2d 297 (1973)). The adoption of the “reasonable possibility” rule is not necessary to accomplish that purpose and provides no meaningful amplification. Moreover, the uncertainty left by today’s efforts to reformulate the evidentiary rule to include an undefined “reasonable possibility” standard, no less than the “clearly linked” lexicon which this court now rejects, portends continued difficulties for the trial court and counsel in attempting to apply the rule of relevance in this area, in my opinion. Stripped of the “explication] by Johnson’s reasonable possibility formulation,” as adopted by the majority, the remaining rules are workable and consistent with the general principles of relevance in this jurisdiction.

For the foregoing reasons and for the reasons stated in the majority opinion in Win-field I, I respectfully dissent from the opinion of the court in its formulation of the new rule and in its application to the facts of this ease.