(dissenting). The court today holds that a “victim’s prior violent conduct may be probative in determining whether the victim was the first aggressor where a claim of self-defense has been asserted and the identity of the first aggressor is in dispute.”1 Ante at 650. In particular, the court creates a new rule of evidence in which prior acts of violence “reasonably alleged to have [been] initiated” by a victim, and unknown to a defendant, may be admitted to establish a victim’s tendency toward violence. Id. Because I believe that past acts of violence by a victim, unknown to a defendant, do little to help a jury resolve the issue whether a defendant was the first aggressor and have no place in our consideration of a defendant’s guilt or innocence, I respectfully dissent.
It has long been the rule in the Commonwealth that “evidence of a person’s character is not admissible to prove that [that person] acted in conformity with that character on a particular occasion.” P.J. Liacos, M.S. Brodin, & M. Avery, Massachusetts Evidence § 4.4.1, at 130 (7th ed. 1999) (Liacos). See Commonwealth v. Baker, 440 Mass. 519, 529-530 (2003). Consistent with this guiding principle, we allow juries to consider evidence of a victim’s prior acts of violence or reputation for violence, provided that the defendant knew about such acts or reputation at the time of the altercation. See Commonwealth v. Fontes, 396 Mass. 733, 733-734 (1986); Commonwealth v. Edmonds, 365 *668Mass. 496, 502 (1974). This “known” evidence bears directly on the controversy being tried — it is nonpropensity evidence that speaks to a defendant’s reasonable belief at the time of the alleged crime. Today’s decision represents a fundamental break from this limitation on character evidence by allowing a defendant to present evidence that has no direct connection to the case at hand and is relevant to nothing but a victim’s general tendency toward violence.2 While the court largely avoids the use of the term “character evidence,” the evidence to be admitted after today is precisely that: evidence of a person’s prior acts used to show his or her propensity to commit such acts. Cf. Commonwealth v. Helfant, 398 Mass. 214, 224 (1986) (character evidence is “evidence that a defendant previously has misbehaved, indictably or not, [introduced] for the purposes of showing his bad character or propensity to commit the crime charged”).
As an initial matter, I am troubled by the premise underlying the court’s decision: that a victim’s violent character (as evidenced through prior violent acts) is sufficiently probative of his or her action on a particular occasion to justify its admission. See ante at 658-660. Although authorities have acknowledged that character evidence may be probative of a person’s behavior, see, e.g., Liacos, supra at § 4.4.1, at 131; 1A J. Wigmore, Evidence, § 57, at 1180-1181, 1211, 1366-1367 (Tillers rev. ed. 1983), I believe we should be hesitant to admit such evidence in the manner proposed by the court today. That “[t]he relevance threshold for the admission of evidence is low,” see ante at 657 n.11, quoting Commonwealth v. Arroyo, 442 Mass. 135, 144 (2004), does not satisfy our doctrine regarding the admissibility of character evidence. Our traditional rule prohibiting the admission of prior act evidence to prove a person’s character, while primarily motivated by our concern regarding prejudice, has *669also been rooted in our belief that prior acts may not be adequately probative of a person’s action on any particular occasion. See Liacos, supra at § 4.4.1, at 131, 1A J. Wigmore, supra at § 54.1, at 1150; Commonwealth v. Stone, 321 Mass. 471, 473 (1947). And the widely acknowledged concern over the risk of prejudice attendant to the admission of character evidence necessarily arises from, and is connected to, its questionable value. How many fights must a person initiate before being considered a “violent” person? One? Ten? Does the person who tends to pick fights as a young adult necessarily grow to be a fifty year old initial aggressor?
A guiding principle of our legal system is that we try cases and controversies, not people or their characters. Because we understand that good people sometimes do bad things, that bad people do not always do bad things, and that circumstances greatly influence behavior, it follows that, with certain well-defined exceptions, a person’s prior actions, no matter how vile, should not be considered in judging their action at any other particular time. Commonwealth v. Stone, supra (“It does not follow that, because the defendant committed a similar offence on another occasion, he committed the crime for which he is being tried. . . . [I]t is not fair that a defendant in the course of a trial should be called upon to defend himself against accusations not set forth in the indictment”). See Maillet v. ATF-Davidson Co., 407 Mass. 185, 188 (1990) (evidence that plaintiff had once consumed “a beer” while at work irrelevant to prove plaintiff consumed beer on day of accident). See also Henderson v. State, 234 Ga. 827, 829 (1975) (“A single act may have been exceptional, unusual, and not characteristic and thus a specific act does not necessarily establish one’s general character . . .”). Precisely because a person’s prior actions can be misleading and inaccurate predictors of future behavior, we have long held that it is unfair to judge a person based on past transgressions. See 1A J. Wigmore, supra at § 55.1, at 1160 (“it cannot be argued, ‘Because A did an act X last year, therefore he probably did the act X as now charged.’ Human action being infinitely varied, there is no adequate probative connection between the two. A may do the act once and may never do it again: and not only may he not do it again, but it is in no *670degree probable that he will do it again. The conceivable contingencies that may intervene are too numerous”). For these reasons, I have considerable misgivings about adopting the view that character evidence is adequately relevant where such a view seems inconsistent with our long-held judicial philosophy against condemning people on the basis of past behavior. I believe we should adhere faithfully to this basic philosophy and remain committed to considering the crime charged and the facts of the case at hand, not the character or prior acts of the victim or the defendant.
Putting aside my doubts about the relevancy of the prior act evidence, I reject the idea that, in weighing the dubious probative value of such evidence against the risk of prejudice, we ought to adopt such evidence against one party and not the other. Although some scholars take the position that character evidence about victims, as opposed to defendants, poses few risks, see 1A J. Wigmore, supra at § 63.1, at 1382, I believe our well-settled rejection of character evidence should be applied with equal force to both defendants and victims. Prior acts of violence, possibly occurring years earlier and involving different parties, are no more probative or less damaging because they are committed by a victim rather than a defendant. If we accept, as we have historically, that a defendant’s prior behavior is insufficiently predictive of her actions years later to outweigh the potential for prejudice, it follows that a victim’s prior acts are also inappropriate for consideration. Conversely, if character evidence is, as the court proclaims today, highly relevant as to victims, it must be similarly so for defendants.
The court limits this newly admissible type of character evidence to prior violent acts initiated by a victim. See ante at 650. This does little to resolve the underlying deficiency of this type of character evidence. I do not believe that a single past act in which a victim was the aggressor is probative on the first aggressor issue. If the act must have been initiated by the victim to be relevant, why must it not involve similar circumstances to be meaningful? Why not the use of the same weapon, for example? Even accepting, arguendo, the court’s premise that prior acts of violence are sufficiently relevant to outweigh any risk of prejudice, a victim would have had to engage in a pat*671tern or series of unprovoked attacks for the evidence to be adequately probative. I also disagree with the premise that prior bad act evidence in this context is somehow appropriate because it is being used to prove the “identity” of the first aggressor. See id. Evidence that is probative of identity is generally that evidence that distinguishes a person’s behavior from others because of its uniqueness or distinctiveness. Commonwealth v. Baker, 440 Mass. 519, 530-531 (2003), quoting Commonwealth v. Jackson, 417 Mass. 830, 836 (1994). In contrast, the evidence to be admitted today need not indicate any modus operandi or reveal any signature style of behavior by the victim. It is not suggestive of identity, but rather allegedly indicates that one of the two parties — the victim — is the violent “type.”
Even if we assume that such prior acts are sufficiently probative of future behavior, today’s rule would be fair only if victims were equally able to explore defendants’ violent histories. Instead of creating an even-handed approach, the court today constructs a one-sided rule that is prejudicial to victims. In cases involving two parties with violent pasts (not uncommon in murders and assaults), defendants may now introduce evidence of the victims’ violent histories, while the Commonwealth will remain powerless to introduce similar evidence concerning defendants. See Commonwealth v. Helfant, 398 Mass. 214, 224 (1986). The court asserts that today’s ruling will provide juries with “as complete a picture of the (often fatal) altercation as possible.” Ante at 658-659. But by constructing a lopsided rule that permits consideration of only one side of the story, the decision does little to paint a “complete” picture for the jury and much to promote a biased view of the parties. And what of a victim’s propensity toward peacefulness? Will prosecutors be able to rebut damaging character evidence with more favorable character evidence?
Hoping to neutralize this imbalance, the court notes (one would presume favorably) that Fed. R. Evid. 404(a)(1) now allows the Commonwealth to respond to victim character evidence with evidence of a defendant’s violent character. See ante at 666 n.19. Admittedly, the Federal approach, if adopted by the Commonwealth, would resolve the imbalance cited above. *672However, the court today does not adopt this more equitable Federal approach. Even if the court or the Legislature were eventually to adopt the Federal approach, this “solution” to the unfairness problem would be more troublesome than the court’s decision today. The Federal approach, while equitable, would greatly diminish our traditional evidentiary protections that prevent defendants from being “reconvicted” of their prior acts. See, e.g., Commonwealth v. Baker, supra; Commonwealth v. Brusgulis, 406 Mass. 501, 505 (1990); Commonwealth v. Triplett, 398 Mass. 561, 562-564 (1986); Commonwealth v. Stone, supra. I believe today’s decision, by creating a gross imbalance and necessitating some sort of “fix,” sets us on precisely this dangerous course toward the erosion of long-held evidentiary safeguards for defendants. Given this court’s valid concern for defendants’ rights, see ante at 662, I would reject the court’s new rule in favor of our well-balanced traditional approach of excluding most character evidence concerning both victims and defendants.
Today’s decision will have other undesirable consequences. First, the admission of character evidence against victims will unduly prejudice juries against victims with violent pasts. See Liacos, supra at § 4.4.1, at 131 (rule against use of character evidence for propensity purposes premised on “high risk that such evidence will have a prejudicial impact on the jury and will result in a decision motivated by something other than the particular facts of the incident before the court”). Our new rule unreasonably invites the fact finder to evaluate the relative worth of a deceased victim without logical basis.3 As the court acknowledges, “[t]he deep tendency of human nature to punish not because [the defendant] is guilty this time but because he is a bad man and may as well be condemned now that he is caught is a tendency that cannot fail to operate with any jury, in or out of court.” 1A J. Wigmore, supra at § 57, at 1185. While such prejudice raises due process concerns when applied to a defendant, we should be no more willing to allow its applica*673tion to a victim. If we find it unacceptable to imprison people for their prior bad acts, how is it any more acceptable to punish people for their prior bad acts by sanctioning their deaths?
Second, today’s ruling will result in jury distraction and confusion, contribute to judicial delay, and increase litigation costs for the Commonwealth and the defense. Fact finders considering this newly permitted character evidence will be asked to wade through multiple incidents of violence (perhaps having to weigh conflicting evidence concerning several such incidents) before turning to the basis of the prosecution. See Commonwealth v. Fontes, 396 Mass. 733, 736 (1986) (“Negative information about the victim may divert the jury from focusing on their basic task by causing them ... to consider the victim’s character and worth . . .”); Williams v. Lord, 996 F.2d 1481, 1483-1484 (2d Cir. 1993), cert. denied, 510 U.S. 1120 (1994), quoting People v. Miller, 39 N.Y.2d 543, 551 (1976) (New York’s rule against allowing evidence of victim’s prior violent actions serves legitimate interest of “preventing ‘undue diversion to collateral matters’ ” by jury); People v. Miller, supra at 552 (“progress of a criminal trial may become stalled by evidentiary conflicts over matters of tangential relevance, thereby impeding the expeditious administration of justice”).4 It is sometimes difficult for juries to develop a complete and accurate picture of the circumstances before them. It will no doubt be even more difficult for juries to assess comprehensively the circumstances attending myriad prior incidents of violence involving victims, about which they will have far less information.
The court acknowledges some of these concerns in passing but concludes they do not “require an unbending rule excluding all of the victim’s specific acts of violence when relevant to the *674identity of the first aggressor.” Ante at 662. The court’s contention that we currently employ an “unbending rule” is erroneous. Prior to today’s ruling, defendants could introduce a range of evidence of a victim’s past conduct in support of a self-defense claim. See Commonwealth v. Rodriquez, 418 Mass. 1, 5-7 (1994) (evidence of specific recent violent acts by victim directed at defendant); Commonwealth v. Fontes, supra at 735-737 (victim’s violent acts directed at third parties but known to defendant); Commonwealth v. Edmonds, 365 Mass. 496, 502 (1974) (victim’s reputation for violence known to defendant). Defendants have even been permitted, in limited circumstances, to introduce some evidence of victims’ prior acts even though they were unknown to these defendants. See Commonwealth v. Rubin, 318 Mass. 587, 589 (1945) (evidence of threats against defendant, “even though the threat was unknown to the defendant”). Moreover, it is generally recognized that prior acts, regardless whether defendant knew them, may be introduced not as evidence of bad character, but to show common scheme, pattern of operation, absence of accident or mistake, identity, intent, or motive. See Commonwealth v. Triplett, supra at 562-563. I believe these existing evidentiary rules provide a defendant with ample opportunity to examine a victim’s past, particularly given the questionable value of character evidence and the unfairness of today’s new rule.
The primary reason of the court for adopting this new rule of evidence is that it wishes to follow the “overwhelming trend” of other jurisdictions toward admitting character evidence of the victim. See ante at 654-657. Yet while purporting to adopt the modem “trend” and its rationale, the court adopts an approach that differs significantly from that trend, rendering its primary rationale unpersuasive. The trend in other jurisdictions, with which obviously I disagree, is to admit testimony of the victim’s reputation as proof relevant to who may have been the first aggressor, regardless of whether that reputation was known to the defendant at the time of the incident in question. See ante at 654-655 & n.8, and cases cited. However, the court does not follow this trend. Rather, it adopts a novel approach of excluding reputation evidence offered for the purpose of proving the first aggressor, but allowing the admission of evidence of *675specific acts of violence initiated by the victim, unknown to the defendant, to demonstrate the victim’s propensity for violence. See ante at 664.
The court cites forty-five jurisdictions that have held “some form of [victim character] evidence is properly admissible on the first aggressor issue (emphasis added).” Ante at 655. However, of these forty-five jurisdictions, none has taken the approach crafted by the court today (admission of specific act evidence, unknown to the defendant, to prove first aggression by the victim). To the contrary, the majority of jurisdictions cited by the court that have considered the issue of victim character evidence do not permit the use of specific act evidence unknown to the defendant to prove who was the first aggressor, but instead permit only reputation or opinion evidence to show a victim’s violent character.5 Moreover, many of those States that do allow some form of specific act evidence to demonstrate a victim’s propensity for violence expressly exclude direct evidence of specific acts and only permit inquiry into specific acts on cross-examination.6 Several other jurisdictions cited by the court as embracing victim character evidence limit specific *676act evidence to prior convictions.7 Even the few jurisdictions that allow a broader range of specific act evidence to be admitted for propensity purposes tend also to permit reputation or opinion evidence concerning the victim that was unknown to the defendant.8 None takes the novel approach adopted by the court today of limiting such victim character evidence to those specific acts of violence initiated by the victim and unknown to the defendant. See ante at 664-665.
The court thus cites no authority for adopting its approach of admitting “concrete and relevant evidence of specific acts [of violence initiated by a victim but not] more general evidence of the victim’s reputation for violence.”9 Ante at 665. The court in fact acknowledges that the “dominant interpretation” in other jurisdictions is not the one it adopts today, but rather one that excludes specific acts of violence by victims when those acts are unknown to the defendant. See ante at 661 n.15. See, e.g., Allen v. State, 945 P.2d 1233, 1240-1241 (Alaska Ct. App. 1997), and cases cited (prohibiting specific act evidence of victim’s violent character and limiting proof to reputation and opinion evidence, and noting courts construing Federal Rules of Evidence and State courts construing their counterparts to Federal rules have consistently reached same conclusion); State v. Jacoby, 260 N.W.2d 828, 838 (Iowa 1977) (adopting “majority” rule that homicide victim’s violent character may not be *677established by proof of specific acts).10 I could have been persuaded of the need for today’s new rule of evidence if provided with a convincing reason for adopting it. The distinguishable law of other jurisdictions is not sufficiently convincing.
The court attempts to rein in its new rule through judicial discretion. See ante at 663 (“sound discretion of trial judges to exclude marginally relevant or grossly prejudicial evidence can prevent the undue exploration of collateral issues”). However, the court’s hope that judges will exercise their discretion to exclude specific act evidence after today’s sea change in our rules of evidence is unrealistic. Judges are no longer able to rule the evidence per se inadmissible, theoretically leaving open for “discretion” the issues of relevancy and probative value versus prejudice. But the court has already made these determinations as well for judges, in effect leaving little discretion to be exercised by judges who do not wish to contradict this court’s holdings. For the court has concluded that this specific act evidence is relevant, ante at 657, has “substantial probative value,” ante at 656, and that the risk of prejudice to the defendant is the “greater danger,” ante at 658. While the court’s reliance on judicial discretion sounds reasonable, it offers little more than a false hope. On the rare occasions when trial judges risk reversal in order to exclude specific act evidence, the inadequate guidance provided to judges by the court today, see ante at 662-664, is certain to lead to inconsistent application of the new rule, protracted proceedings, and lengthy appeals. Because our deferential standards of review require us to uphold judges’ evidentiary decisions absent abuses of discretion, see, e.g., Commonwealth v. Arroyo, 442 Mass. 135 (2004), our case law will soon reflect the uncertainty and inconsistency of the case-by-case assessments encouraged by the court today.
Contrary to the court’s intentions, today’s endorsement of *678character evidence will not foster more accurate jury findings; nor will it enhance justice for victims or defendants. Because I believe we should affirm our existing law that evidence of a victim’s violent character is only admissible if the defendant knew of that violent character at the time of the alleged crime, see Commonwealth v. Graham, 431 Mass. 282, 291 (2000), I would affirm the defendant’s conviction.
Although today’s rule applies only where “the identity of the first aggressor is in dispute,” ante at 664, it would not be surprising if, after today, we find an increasing number of trials where defendants assert such disputes.
The court downplays the vast differences between the “known” evidence currently admitted under Commonwealth v. Fontes, 396 Mass. 733 (1986), and Commonwealth v. Edmonds, 365 Mass. 496 (1974), and the character evidence of which the court now approves. See ante at 654, 662-663. The former type of evidence helps a jury to consider whether a defendant reasonably feared for his or her life during the incident in question. The latter invites a jury to consider, “Was the victim the sort of person who might have been a first aggressor?”
Presumably aware of these risks, the court encourages judges to “mitigate the dangers of prejudice and confusion inherent in introducing evidence of the victim’s specific acts of violence by delineating the precise purpose for which the evidence is offered.” Ante at 664. I do not believe limiting instructions will be adequate on this subject.
In a recent decision, we determined that a third-party suspect’s history of sexual aggression was relevant to a theory that he, rather than the defendant, may have assaulted and killed the victim. See Commonwealth v. Conkey, ante 60, 66-70 (2004). The evidence deemed relevant in the Conkey case is wholly distinct from the evidence at issue here. There the evidence was probative of whether the defendant, as opposed to someone else, committed the crime. The trial judge had excluded the evidence on relevance grounds. In reversing, we found the evidence was relevant and did not change any evidentiary rules to do so. Here, the court must change our common-law rules of evidence in order to admit the evidence in question.
See, e.g., Higginbotham v. State, 262 Ala. 236, 240 (1955); Allen v. State, 945 P.2d 1233, 1240 (Alaska Ct. App. 1997); State v. Santanna, 153 Ariz. 147, 149 (1987); State v. Zamora, 140 Ariz. 338, 340-341 (Ct. App. 1984); McClellan v. State, 264 Ark. 223, 227 (1978); People v. Ferguson, 43 P.3d 705, 710 (Colo. Ct. App. 2001); Smith v. State, 606 So. 2d 641, 642-643 (Fla. Dist. Ct. App. 1992); State v. Custodio, 136 Idaho 197, 204 (Ct. App. 2001); State v. Jacoby, 260 N.W.2d 828, 838 (Iowa 1977); Commonwealth v. Davis, 14 S.W.3d 9, 14 (Ky. 1999); State v. Edwards, 420 So. 2d 663, 671 (La. 1982); People v. Cellura, 288 Mich. 54, 68 (1939); State v. Irby, 368 N.W.2d 19, 23 (Minn. Ct. App. 1985); State v. Sattler, 288 Mont. 79, 95-96 (1998); State v. Newell, 141 N.H. 199, 201-202 (1996); State v. Barbour, 295 N.C. 66, 74 (1978); State v. Barnes, 94 Ohio St. 3d 21, 24 (2002); Harris v. State, 400 P.2d 64, 70 (Okla. Crim. App. 1965); State v. Dellay, 687 A.2d 435, 438 (R.I. 1996); Tate v. State, 981 S.W.2d 189, 192-193 & n.5 (Tex. Crim. App. 1998); State v. Roy, 151 Vt. 17, 31 (1989); State v. Hutchinson, 135 Wash. 2d 863, 886-887 (1998), cert. denied, 525 U.S. 1157 (1999); State v. Boggess, 204 W. Va. 267, 275 n.11 (1998); Werner v. State, 66 Wis. 2d 736, 744 n.6 (1975).
See, e.g., Brooks v. State, 683 N.E.2d 574, 576-577 (Ind. 1997); Daniel v. State, 78 P.3d 890, 901 (Nev. 2003), cert. denied, 541 U.S. 1045 (2004); State v. McIntyre, 488 N.W.2d 612, 617-618 (N.D. 1992); State v. Botches, 331 Or. 455, 489 (2000), cert. denied, 534 U.S. 833 (2001); State v. Blem, 610 N.W.2d 803, 812 (S.D. 2000); State vs. Hammock, No. M2000-00334-CCA-R3-CD 824 (Tenn. Crim. App. Oct. 12, 2001).
See, e.g., State v. Smith, 222 Conn. 1, 17-18, cert. denied, 506 U.S. 942 (1992); State v. Deavers, 252 Kan. 149, 156-157 (1992), cert. denied, 508 U.S. 978 (1993); State v. Aguiar, 322 N.J. Super. 175, 183 (App. Div. 1999); Commonwealth v. Beck, 485 Pa. 475, 478 (1979); State v. Howell, 649 P.2d 91, 96 (Utah 1982).
See, e.g., People v. Wright, 39 Cal. 3d 576, 587 (1985); Rawls v. United States, 539 A.2d 1087, 1089 (D.C. 1988); Chandler v. State, 261 Ga. 402, 407 (1991); Bennett v. State, 254 Ga. 162, 164 (1985); State v. Basque, 66 Haw. 510, 513-514 (1983); State v. Lui, 61 Haw. 328, 329-330 (1979); People v. Florey, 153 Ill. App. 3d 530, 538-539 (1987); Newsom v. State, 629 So. 2d 611, 613-614 (Miss. 1993); State v. Lewchuk, 4 Neb. App. 165, 175 (1995); Edwards v. State, 973 P.2d 41, 45-46 (Wyo. 1999).
I do not read the court’s decision as altering our traditional law permitting admission of evidence of a victim’s reputation where it is known to a defendant claiming self-defense. See Commonwealth v. Edmonds, 365 Mass. 496, 502 (1974). Nor do I read it as altering the long-standing rule that an accused may introduce reputation evidence of his or her own good character. See Commonwealth v. Belton, 352 Mass. 263, 268, cert. denied, 389 U.S. 872 (1967).
In critiquing the court’s decision to allow the admission of specific act evidence while prohibiting reputation evidence, I do not wish to suggest my approval of reputation evidence to prove the identity of the first aggressor where such reputation was unknown to the defendant. I find both forms of character evidence equally lacking in probative value where unknown to the defendant. Instead, my point is that the approach taken by the court today comports with neither logic nor any modem trend.