OPINION OF THE COURT
Wachtler, J.The defendant was convicted at a jury trial of murdering a woman in her Manhattan apartment. The Appellate Division affirmed the conviction, with one Justice dissenting.
On this appeal the defendant primarily claims that the trial court erred in admitting certain evidence used to identify him as the killer. His main point, and the only one to produce disagreement at the Appellate Division, concerns the introduc*44tion of proof that, within five months of the homicide, the defendant committed, and prior to this trial pleaded guilty to, six rapes in which he employed the same modus operandi as was apparently used in the homicide. The modus operandi, he claims, was not unique and thus the prejudicial effect of revealing these other crimes outweighed their probative value as evidence of the killer’s identity. He also argues that the court should have conducted a pretrial hearing, pursuant to United States v Wade (388 US 218), to consider the admissibility of a voice identification. Finally, he objects to the use of expert testimony to identify hair samples.
The murder victim, Carol Hoffman, was found dead in her apartment on East 29th Street in Manhattan at 9:30 on the evening of October 23, 1973. Less than an hour prior to the discovery of her body, her boyfriend, Vincent St. George, had called the apartment and spoken to her on the phone. She told St. George, who had called earlier and received no answer, that she had just returned home from a night class which had been held over. During the conversation St. George detected an unusual tone in her voice and asked if something was wrong. After some initial hesitancy she finally acknowledged that there was someone else in the apartment — "someone” she said "who was looking for someone in the building.” She was reluctant, however, to discuss this further and told St. George that she would call him back in a few minutes.
Two or three minutes later she returned the call and told St. George "here, you talk to him. Maybe you can talk him out of it.” St. George then spoke to a man who explained that he was looking for someone who had raped his wife and thrown her down a flight of subway stairs two weeks ago; that as a result his wife had a miscarriage and was still hospitalized and that the rapist was in this building, on this floor, and he was now trying to find the apartment. He also told St. George the police had done nothing and that he was pursuing the rapist on his own, although he did not know what he would do if he found him and was, in fact, afraid of him.
After 10 or 15 minutes the conversation ended when St. George apparently convinced the man to go back to the police. St. George then spoke to Carol Hoffman who, still reluctant to discuss the matter, said that she would call him back again in a few minutes. When she did not return the call St. George proceeded to her apartment shortly after 9 o’clock.
At approximately 9:15 a tenant on the floor above heard a *45loud thump in the Hoffman apartment followed by a woman crying "stop” or "no”. When St. George arrived 10 or 15 minutes later there was no answer at the door. With the help of the building superintendent he entered the apartment and found Carol Hoffman lying on the floor with a kitchen knife protruding from her stomach and a pair of panty hose around her neck.
Police officers called to the scene observed that the apartment had not been ransacked, with the exception of the dresser drawers all of which were partially open. But only the lingerie drawer was in disarray with some of the articles "hanging out”. They noticed a strand of hair stuck between the victim’s teeth. An autopsy revealed that she had been manually strangled and then stabbed 11 times. She was fully clothed, wearing a dress, slip and underwear. The perforations in the skirt, however, indicated that it had been folded back on itself and therefore apparently raised over the waist at the time of the stabbing. There was no evidence of a completed rape.
St. George described the voice of the man he had spoken to on the phone as deep, soft spoken, and well articulated, without a trace of an accent. He said the man had spoken in a calm, self-assured manner. Two days after the killing he went to the police station where he heard the defendant’s voice on the telephone and identified him as the man he had spoken to prior to Carol Hoffman’s death.
The police clipped hair samples from the defendant’s head and sent them to the FBI laboratory. Later at the trial an expert from the laboratory stated his opinion that the hair found in the victim’s teeth matched hairs removed from the defendant’s head.
At the trial the People also called six women who testified that they had been raped by the defendant during the five months preceding the homicide. Their testimony describing the defendant’s conduct during these incidents was offered and admitted on the theory that it showed that the defendant had previously employed the same modus operandi evident in the homicide and would thus tend to establish his identity as the killer.
In each of those cases the defendant had followed the woman into a residential building as she returned alone to her apartment. In one case he then used force to enter her apartment. In the others he employed a ruse, either posing as *46a neighbor with a petition or a package or claiming that he was trying to find someone who resided in the building. In each case he grabbed the victim by the throat and threatened to strangle her if she did not do what he said. In four of the cases he also used or threatened the use of a knife.
In each instance the defendant rummaged through the victim’s lingerie and if a slip was found directed her to wear it. In the three most recent cases he used the victim’s panty hose as a rope to bind her before the rape or to aid his escape. During these incidents the defendant often mentioned his wife or fiancée and in two instances he also told an unusual story about them. On the first occasion he said that he had become vengeful because someone had raped his fiancée and stabbed her to death. On the other occasion he said that it was his wife who had been raped and murdered.
Four of the women were asked to describe the defendant’s voice and stated that he had a low or deep voice with no accent and that he articulated his words well or spoke in a calm steady tone. All six of the sexual assaults had occurred in Manhattan, three on the upper West Side and the rest on the lower East Side, within a block or two of Carol Hoffman’s apartment. As noted, it was stipulated that in each case the defendant had pleaded guilty.
The defendant objected to this testimony at the trial and on this appeal claims that the prejudicial effect of the evidence of his prior criminal acts so far outweighed its probative value as to deprive him of a fair trial.
It is fundamental that evidence of uncharged crimes is not admissible if the sole purpose is to show that the defendant was predisposed to commit the crime charged (People v Dales, 309 NY 97, 101; Coleman v The People, 55 NY 81, 90), unless the defendant places this in issue by, for instance, claiming that he was entrapped or coerced into committing a crime to which he was not otherwise disposed (People v Calvano, 30 NY2d 199; People v Mann, 31 NY2d 253; cf. People v Thompson, 212 NY 249). The rule is based on policy and not on logic. It is meant to eliminate the risk that a jury, not fully convinced of the defendant’s guilt of the crime charged may, nevertheless, find against him because his conduct generally merits punishment (People v Molineux, 168 NY 264, 294; People v Zackowitz, 254 NY 192, 198; People v Fiore, 34 NY2d 81, 84-85).
On the other hand, evidence relevant to prove some fact in *47the case, other than the defendant’s criminal propensity, is not rendered inadmissible simply because it may also reveal that the defendant has committed other crimes (People v Molineux, supra, at pp 291-294; People v Fiore, supra, at p 84). Thus, for instance, evidence of other crimes may be admitted to show motive, intent, the absence of mistake or accident, a common scheme or plan or the identity of the guilty party (People v Molineux, supra, at p 293). The list, of course, is not exhaustive (see People v Jackson, 39 NY2d 64).
When evidence of uncharged crimes is relevant to some issue other than the defendant’s criminal disposition, it is generally held to be admissible on the theory that the probative value will outweigh the potential prejudice to the accused (People v Dales, supra, at p 101; People v Schwartzman, 24 NY2d 241, 247-248, cert den 396 US 846). This rule, however, cannot be applied abstractly. If the evidence is actually of slight value when compared to the possible prejudice to the accused, it should not be admitted, even though it might technically relate to some fact to be proven (see, e.g., People v McKinney, 24 NY2d 180, 185; People v Liller, 20 NY2d 727; see, also, People v Fiore, supra, at p 84).
In this case the identity of the killer was the primary issue at trial. The People’s case on this point was entirely circumstantial. There was other evidence, apart from the proof of the prior rapes, which placed the defendant at the scene. But the hair sample and the voice identification did not conclusively identify the defendant as the killer. It is settled that unless identity has been Conclusively established the People are not precluded from attempting to identify the defendant as the culprit by proof of other crimes (People v Condon, 26 NY2d 139, 142).
But in order to identify the defendant in this manner it is not sufficient to show that he has committed similar acts if the method used is not uncommon. Simply categorizing the defendant as one of many criminal specialists would be of little probative value in determining whether he committed the crimes charged, and the prejudice would be obvious: "it is much easier to believe in the guilt of an accused person when it is known or suspected that he has previously committed a similar crime” (People v Molineux, 168 NY 264, 313, supra). There must be some additional factor to set the defendant’s crimes apart from the ordinary so that "the mere proof that the defendant had committed a similar act would be highly *48probative of the fact that he committed the one charged” (People v Condon, supra, at p 144; see, also, People v Kennedy, 27 NY2d 551; McCormick, Evidence, § 190).
Here the circumstances evident at the homicide mirror the defendant’s conduct during the rapes. He urges, however, that his prior actions were not unique and thus cannot serve to identify him as the killer. Nevertheless there are certain oddities which alone, or in combination with the other circumstances, serve to distinguish the defendant’s criminal acts.
The most peculiar feature undoubtedly is the bizarre story the defendant told concerning the rape and assault on his wife or fiancée. He mentioned it to one of his rape victims in August and to another in September. On the last occasion he was only a block from the scene of the homicide. It is essentially the same story told by the man who was in Carol Hoffman’s apartment minutes before her death on October 23. This is too unusual to be mere coincidence.
There is also the defendant’s compelling interest in lingerie. He habitually rummaged through his victim’s lingerie as a prelude to the rapes. He had done it three times within a block or two of the Hoffman apartment in July, August and September. The testimony of the other victims shows that this was a consistent , pattern in all six of the rapes. It was evident also at the scene of the homicide where the lingerie drawer was in disarray. This was remarkable because the rest of the apartment, including the dresser drawers, had not been rummaged. It may be that it is not unusual for sex offenders to display a compulsive interest in lingerie. But it is unlikely that there would be many with this particular interest operating in the same neighborhood, at the same time, all with similar voices who employ identical methods in gaining entry to women’s apartments and, once inside, display a readiness to stab or strangle their victims if they do not get their way. The fact that this was an invariable part of the defendant’s modus operandi was, at least, "sufficiently unique” to be submitted for the jury’s consideration (People v Kennedy, supra).
In sum, the evidence was admissible not because it showed the defendant’s propensity to commit the crime but because it demonstrated a distinctive repetitive pattern, evident at the scene of the homicide, which was characteristic of the defendant’s behavior. In this homicide case, where there were no witnesses and the other evidence was less than conclusive, it *49was highly probative of the murderer’s identity. The trial court properly limited the proof to the surrounding circumstances without dwelling on the details of the sex acts and the prosecutor during his summation exercised similar restraint. In addition, the court instructed the jury, during the voir dire and later in the charge, that the evidence was only to be considered on the issue of identity. Thus prejudice to the defendant’s case, which is always possible whenever evidence of other crimes is admitted, was appropriately minimized by the care exercised by the prosecutor and the Trial Judge and was outweighed by the importance and probative value of this testimony to the central issue of the killer’s identity.
As noted, the defendant also claims that the court should have conducted a pretrial hearing to determine the admissibility of the voice identification made by St. George. He claims that the standards and procedures applicable to visual identifications (United States v Wade, 388 US 218, supra) should be applicable to voice identification and that the voice identification made at the station house was unduly suggestive. In this case, however, there is no need to consider whether, or to what extent, the procedural and substantive standards initiated in the Wade case are applicable to voice identifications. Here it is sufficient to note that at the time the application was made there was no factual support in the record for the contention that the identification procedure was suggestive. There was only speculation. Under these circumstances the court did not err in denying the motion (CPL 710.60, subd 1).
Finally, the defendant claims that the court erred in permitting the People’s expert to offer his opinion that the hair found in the victim’s mouth had come from the defendant’s head. He argues that the results of hair analysis are no more reliable than lie detector evidence which has been held to be inadmissible (People v Forte, 279 NY 204; People v Leone, 25 NY2d 511).
The People’s expert testified that he had microscopically compared the hair samples taken from the defendant’s head with the hair found at the scene of the crime. He stated that the test, like fingerprint analysis, involved comparing a number of characteristics, generally 15 to 20. He conceded that the results would not be as conclusive as fingerprinting, but stated that if a sufficient number of similarities could be found, it could be determined with a reasonable degree of certainty *50that a hair had come from a certain individual. He said that he was able to do that in this case.
Expert testimony is admissible if the analysis involved is beyond the ken of the typical juror and the results would be relevant to an issue in the case (see, e.g., Dougherty v Milliken, 163 NY 527, 533). Generally conclusive results are not required. It is sufficient if the expert can state his opinion with a reasonable degree of certainty (see, e.g., Matter of Miller v National Cabinet Co., 8 NY2d 277). The higher standard is applicable only when the expert is called to assess credibility (People v Williams, 6 NY2d 18; cf. People v Leone, supra).
Here, of course, the expert was not asked to assess credibility, and his testimony satisfied the general requirements applicable to expert proof. We also note that the value of hair analysis, with its limitations, is recognized by legal scholars (see, e.g., Moenssens & Inbau, Scientific Evidence in Criminal Cases, pp 405-414; McCormick, Evidence, § 206, p 503) and is routinely accepted by courts in other jurisdictions (see, e.g., People v Abbott, 47 Cal 2d 362; Padilla v The People, 156 Col 186; People v Kirkwood, 17 Ill 2d 23; Commonwealth v Tarver, 369 Mass 302, 309-311; State v Baldwin, 47 NJ 379; State v Vargus, — RI —, 373 A2d 150, 156-157 [RI]).
Accordingly, the order of the Appellate Division should be affirmed.
Chief Judge Cooke and Judges Jasen, Gabrielli, Jones, Fuchsberg and Meyer concur.
Order affirmed.