(dissenting, with whom Lenk, J., joins). I agree with the court that the evidence was legally sufficient to support the defendant’s convictions. Because I disagree with the court’s analysis of the admission of photographs of items found in the defendant’s apartment, however, I respectfully dissent.
Today, the court purports to announce a new brand of identity evidence, which involves the use of similar, but not distinctive, conduct of a defendant to infer that the defendant acted in conformity with that conduct on another occasion. If this sounds *491like the language often used to describe the type of character inference that fact finders are roundly prohibited from making, that is because it is.
The court sets forth three “relevant, noncharacter purposes” for which it holds the photographs were admissible: (1) identity, (2) state of mind, and (3) motive. I address each in turn, and finally, I assess the prejudicial effect in this case.
1. Identity evidence. I agree with the court that the posters hanging in the defendant’s apartment were not sufficiently similar to the methods by which the victims’ bodies were dismembered to qualify as modus operandi evidence.11 also agree that evidence may be admissible to prove a defendant’s identity, absent such similarity, when the evidence is ultimately relevant because the evidence makes it more likely than it would be without the evidence that the defendant is the individual responsible for the crime. This latter category of “identity” evidence, however, does not permit the use of the defendant’s conduct “to prove [his] character in order to show that on a particular occasion [he] acted in accordance with the character.” Mass G. Evid. § 404(b)(1) (2017). This rule limiting the use of a defendant’s prior conduct applies without regard to the probative strength of the conduct.
“One of the oldest principles of Anglo-American law is that a person ‘should not be judged strenuously by reference to the awesome spectre of his past’ ” (citation omitted). D.P. Leonard, The New Wigmore: A Treatise on Evidence § 1.2, at 2 (2009) (Wigmore). “For nearly two centuries, courts have excluded the evidence not because of its lack of probative value but primarily because of the dangers it is thought to present. Most commonly cited is the danger of unfair prejudice.” Id. at 6. This prohibition on character evidence includes using a defendant’s other, relevant *492conduct to prove his or her “propensity” to commit the charged crime. Id. at 2-3. See Mass. G. Evid. § 404(b). The danger of admitting such character evidence against the defendant is not that it is irrelevant. Rather, the danger is that the jury will overvalue the evidence. Wigmore, supra at 5-7.
To say that other conduct is permissibly probative of “identity,” rather than impermissibly probative of character, merely because a defendant’s character makes him more likely to be guilty, is an exercise in circular logic that renders the prohibition on the character inference inert. Thus, the court’s reasoning today, at best, dilutes the stringent requirements for modus operandi evidence, or, at worst, eviscerates the rule prohibiting use of a defendant’s other conduct to show his propensity to commit the crime charged.
I would classify admissible evidence that is probative of identity into two categories: (1) modus operandi, and (2) what I will call “identity-based evidence.” Unlike modus operandi, identity-based evidence does not require a high level of distinctiveness shared between the defendant’s other conduct and the charged conduct.2 Rather, such other conduct constitutes admissible identity-based evidence when introduced for a nonpropensity purpose, such as, motive, opportunity, knowledge, state of mind, or many other purposes, but the nonpropensity purpose is ultimately relevant to “identify” the defendant as the individual who committed the charged crime. See P.C. Giannelli, Understanding Evidence 174-175 (4th ed. 2013) (Giannelli).
The following example of identity-based evidence is illustrative. If shortly before committing armed robbery, a defendant steals a particular weapon to commit the armed robbery, evidence that the defendant stole the weapon would be admissible to establish that he had the means or opportunity — because he had the particular weapon used to commit the crime. See Giannelli, supra at 174. That the defendant had the means to commit the crime is relevant to his “identity” as the perpetrator of the armed robbery. See id. The judge, however, must still balance the probative value of the theft of the weapon against the danger of undue prejudice that the jury will consider the theft as evidence *493of bad character. See Mass. G. Evid. § 403. Yet, there is minimal danger that a jury would impermissibly consider the theft as indicative of the defendant’s propensity to commit armed robbery, i.e., that the defendant showed his bad character by stealing a firearm, and that it was more likely that he committed the crime due to this bad character. Rather, the more probable and logical inference is that the evidence “identifies” the defendant as the individual who committed the crime because he possessed the weapon used in its commission. The latter conclusion is not based on an impermissible propensity inference.
In this case, the connection between the defendant’s other conduct and the charged conduct is squarely based on an impermissible propensity inference. The other conduct is hanging posters depicting medical amputations. The charged conduct is chopping up three human beings. The logical connection between the two is that the defendant acted in conformity with the character trait demonstrated by displaying images of amputation by brutally chopping up the victims on a subsequent occasion — a stark contrast to the firearm example above, which involves no such impermissible character inference.
The court conditions the admissibility of the drawings on Adam Lee Hall’s statement that “one of the guys really enjoyed” chopping the victims up. The court says that the anatomical drawings are thus probative of identity in the same way that posters of medieval weapons would be admissible identity evidence if Hall had said one of the participants was fascinated by medieval weapons.
The court’s example is not analogous to the present case. Unlike the present case, the court’s example does not implicate a propensity inference, because the medieval weapons referred to in the hypothetical example do not relate to the commission of the crime. Thus, the example does not ask the jury to conclude that, because the defendant had posters of medieval weapons, he is the type of person who would participate in three brutal murders. In the hypothetical example, Hall’s statement serves only to identify a person who has an interest in medieval weapons. The medieval weapons posters are relevant because they show an interest of the accused, and the hypothetical statement identifies an individual who has that interest as a participant in the crime. The posters in no way suggest that the defendant acted in accordance with that interest in killing the victims.
By contrast, in this case Hall stated that one of the participants enjoyed the act of torturing and chopping up people. The ana*494tomical drawings only corroborate this statement if one presumes that the defendant acted in accordance with his interest in anatomical dismemberment on a subsequent occasion by chopping up the victims in a manner that did not meaningfully resemble the dissections depicted in the drawings. Regardless of whether the defendant’s display of the posters makes it more likely that he was the third participant than it would be without such evidence, this is the quintessential, impermissible propensity inference.
2. State of mind. The court concludes that the fact that the dismemberment of the victims “appears to have had no pragmatic purpose and . . . must have taken a considerable amount of time to complete” was an indication of the defendant’s state of mind. Ante at 484. Even setting aside the evidence suggesting that there was in fact a pragmatic purpose for dismembering the bodies,3 this evidence still requires a jury to assume that an individual who is “fascinated by amputation and human dissection,” demonstrated only by display of posters, would “seize the opportunity of these killings to engage in actual amputations and human dissection.” Id.
Under this rubric, the court’s theoretical path of admissibility is “identity-based” evidence: a person who is fascinated with amputation is more likely to engage in the act of physically dismembering people. The court may be correct that displaying the posters is probative of the defendant’s state of mind, which ultimately is relevant to identify him as the perpetrator of the crime. But, we do not allow in evidence simply because it is relevant. See Wigmore, supra at 5. This theory still requires the quintessential prohibited inference, although labeled as “state of mind,” in this application. To be relevant to the defendant’s state of mind, one must conclude that he acted in conformity with his other conduct of hanging the posters on a subsequent occasion by participating in the murders.
*495The Commonwealth itself described the state of mind only as “depraved.” This is a thin veil. It is difficult to imagine an interpretation of this argument that is not a bald assertion that the defendant’s bad or “depraved” character makes him more likely to be guilty of murder. See Commonwealth v. Crayton, 470 Mass. 228, 251-252 (2014) (jury prohibited from inferring that defendant’s interest in child pornography meant he must have been person who accessed child pornography in library). Admitting the photographs as “state of mind” evidence where the photographs reflect only a general character trait of the defendant eviscerates any distinction between evidence of a character trait and that of state of mind.
The admission of this evidence was coupled with testimony from one witness that she observed “a lot of creepy shit everywhere” inside the defendant’s apartment. On this theory of admissibility, the Commonwealth does not attempt to factually or temporally tie this so-called “state of mind” evidence to the crime at issue. Contrast Commonwealth v. Drew, 397 Mass. 65, 78-79 (1986), S.C., 447 Mass. 635 (2006) (defendant’s participation in Satanic rituals relevant to prove involvement in ritualistic killings). Accordingly, I would conclude that the posters were not admissible to prove the defendant’s state of mind.
3. Motive. The court concludes that the posters were independently probative of the defendant’s “motive.” The court relies on cases in which we have allowed the Commonwealth to establish a “context for the killing” when it would otherwise appear to the jury as an “inexplicable act of violence” (citation omitted). Commonwealth v. Marrero, 427 Mass. 65, 68 (1998). The circumstances of this case do not resemble those relied on by the court. See, e.g., id. (Commonwealth allowed to introduce significant detail regarding defendant’s relationship with victim and witnesses involved in drug business connected to murder); Commonwealth v. Bradshaw, 385 Mass. 244, 269-270 (1982) (Commonwealth permitted to introduce evidence of defendant’s activities on day of murder because they were “inextricably intertwined with the description of events on the [day] of the killing”). See also Commonwealth v. Guy, 454 Mass. 440, 443 (2009) (Commonwealth permitted to admit evidence of defendant’s fascination with serial killings in absence of any other evidence of motive).
The court relies primarily on Guy, 454 Mass. at 443-444, which is not analogous. In that case, the Commonwealth had significant *496physical evidence tying the defendant to the crime, but was faced with a peculiar situation of having no explanation for the jury as to why the defendant had committed the crime.4 By contrast, Hall orchestrated David Glasser’s death to prevent his testimony. The other two victims were murdered to eliminate witnesses to Glasser’s killing. The defendant, as the Commonwealth argued at trial, was motivated to assist Hall because the defendant was aspiring to become a member of the Hells Angels motorcycle club. The Commonwealth did not, and does not on appeal, argue that the defendant was motivated to participate in the crime to seize the opportunity to dismember human beings, or that the dismemberment had no practical purpose.
4. Prejudicial effect. For the reasons set forth above, I would conclude that the anatomical drawings were probative only of the defendant’s character and were thus inadmissible. Accordingly, it is unnecessary to weigh the probative value against the danger of undue prejudice, since this evidence has no probative value other than propensity. Instead, the relevant inquiry is whether the error created a “reasonable possibility that. . . might have contributed to the jury’s verdict.” Commonwealth v. Alphas, 430 Mass. 8, 23 (1999). The Commonwealth bears not only the burden to show the lack of error, but also the “risk of doubt when any exists” as to whether the error influenced the jury’s verdict. Id. The Commonwealth does not even argue that the evidence, if erroneously admitted, was not prejudicial. That may well end the inquiry, but there are five factors that I believe enhanced the danger of prejudice in this case.
First, even if the evidence had been admissible, it should have been accompanied by a limiting instruction. No limiting instruction was requested or given at trial, despite the palpable danger of undue prejudice of the evidence. Without a limiting instruction, the photographs were before the jury for all purposes, including as impermissible propensity evidence. This danger created a substantial likelihood of a miscarriage of justice. G. L. c. 278, *497§ 33E. See Crayton, 470 Mass. at 252 (illustrations too prejudicial to justify admission, even with limiting instruction).
Second, in other cases, we have found that the failure to give a limiting instruction did not warrant reversal, when other circumstances mitigated the danger of unfair prejudice. For example, in Guy, 454 Mass. at 443-444 & n.3, in which we did not discuss a limiting instruction, we noted that the prosecutor utilized evidence of the defendant’s interest in books about serial killings — which were not admitted in evidence — in a “technical, analytical manner, without drama or undue emphasis that might have released its potential for unfair prejudice.” Further, the prosecution in that case had compelling physical evidence connecting the defendant to the crime, reducing the probability that the jury would return a guilty verdict based on the defendant’s macabre interest.
This case is distinguishable from Guy. Unlike the books in Guy, the posters themselves, depicting graphic images, were admitted in evidence. Also, the prosecutor was not especially cautious in avoiding drama or character-related implications in his closing argument, to mitigate the danger of unfair prejudice. Rather, the prosecutor made a graphic emotional appeal to the jury, referring to the defendant’s intent “not just... to keep three men from testifying in court but of satisfying some retribution and intent to take apart humanity piece by piece.” The court concludes that this statement, clearly referring to the horrendous nature of the dismemberment and not any pertinent evidentiary point, is comparably “technical and analytic.” I disagree.
Third, although the prosecutor did not explicitly refer to the photographs in his closing argument, defense counsel quite understandably addressed the evidence in his closing argument three times, in an effort to dampen its prejudicial force.
Fourth, the judge gave proper limiting instructions regarding the defendant’s association with the Hells Angels and Hall’s history with Glasser. By informing the jury that there was specific evidence that they should not consider as evidence of bad character, the jury were left to infer that the remainder of the evidence could be considered as evidence of the defendant’s bad character.
Finally, the likelihood that the jury considered the evidence for a prohibited purpose was further enhanced by the entirely cir*498cumstantial nature of the case against the defendant.5 As in Crayton, 470 Mass. at 250, the primary issue at trial was the defendant’s identity. Due to the lack of direct evidence and a limiting instruction, the jury were more apt to use the photographs as character evidence to infer the defendant’s guilt. Contrast Guy, 454 Mass. at 442-444, 447 (evidence of defendant’s interest in serial killings used to establish his identity as killer, but also deoxyribonucleic acid evidence matched defendant). Even with the impermissible character evidence, the issues were difficult enough to resolve that the jury deliberated nearly five full days before reaching verdicts.
I believe that the Commonwealth did not satisfy its burden to demonstrate that there was no “reasonable possibility” that the erroneous admission of these photographs contributed to the jury’s verdicts. Accordingly, I would reverse the defendant’s convictions and grant a new trial.
When asked whether the dismemberment of the victims was consistent with the surgical illustrations, the Commonwealth’s expert said, “They do show amputation of limbs, so that portion is consistent. I can’t say if the exact location on the bone is consistent in some of them. I can see at least one is inconsistent in terms of location. Sometimes I just can’t tell what paid of the bone I’m looking at.” The prosecutor then asked whether the victims’ limbs were dismembered at the site of the joint (as depicted in the illustrations), the witness responded, “They were chopped through sometimes near a joint but they were chopped through mostly right to the bone itself. In terms of the vertebrae, a lot of the chopping was aimed between two bones, so both bones were damaged but they were separated where they normally separate.” Further, on cross-examination, defense counsel asked, “There’s nothing in that diagram that’s consistent with the multiple large chopping injuries which you just discussed, correct?” The witness answered, “That’s correct.”
As the comí notes, ante at 483-484, a similarity that is merely general is a reason to exclude evidence of other conduct. The danger that a jury will consider prior conduct as propensity evidence is at its apex when the prior conduct resembles the charged conduct, but is not sufficiently similar for purposes of modus operandi.
The comí concludes that the dismembered bodies were likely all placed in Hall’s Buick. Ante at 481. Further, after the killings. Rose Dawson and Alexandra Ely, who were not alleged participants in the killings, drove in the Buick to a supermarket. They did not look in the trunk, but they also did not testify that they saw any blood or body parts in the cabin of the automobile. Moreover, David Casey testified that he later observed Hall open the trunk of the Buick and drop a number of plastic garbage bags into the hole Casey had dug. Accordingly, chopping up the bodies may well have been a practical measure for purposes of transporting three bodies in the trunk of the vehicle, while still retaining the vehicle for limited use until the time it could be destroyed. Whether the dismemberment of the bodies had any practical purpose was not an issue at trial and was not argued by the Commonwealth on appeal.
There was also a greater quantum of evidence that the defendant in Guy, 454 Mass. at 443-444 & n.3, had a significant fascination with serial killings, including a large number of books seized from his home (which were not themselves admitted in evidence, but were referenced) and testimony from his coworkers attesting to his ongoing fascination. Here, the only evidence of the defendant’s “fascination” was that he had placed posters on his wall, and the record suggests that the posters had not been displayed on the wall for a long period of time because the defendant was still in the process of moving his belongings from his previous residence into this residence.
Of course, the Commonwealth is entitled to prove its case entirely by circumstantial evidence. Commonwealth v. Woods, 466 Mass. 707, 713, cert. denied, 134 S. Ct. 2855 (2014).