I
Douglas L. Crawford was convicted following a jury trial of possession with intent to deliver 50 to 225 grams of cocaine. MCL 333.7401(2)(a)(iii); MSA 14.15(7401)(2)(a)(iii). He claims error in the admission of evidence of his prior conviction for delivery of 225 to 650 grams and conspiracy to commit the same offense. Specifically, he contends that the evidence of his prior conviction should have been excluded under MRE 404(b) as improper evidence of his character or propensity to commit the charged offense. We agree that the defendant’s prior conviction was improperly admitted in this case and that his conviction, therefore, should be reversed.
*379n
On September 22, 1992, Oak Park police officer Rene Gobeyn was patrolling Eight Mile Road. He noticed a man, later identified as the defendant, placing something in the trunk of a car in the parking lot of the Embassy Motel. The motel’s reputation apparently caused Officer Gobeyn to suspect that criminal activity might be afoot.
Officer Gobeyn did not immediately confront the defendant, however. Instead, he watched as the car left the motel parking lot and headed west on Eight Mile Road. Gobeyn estimated that the car drove off at a speed in excess of the posted speed limit, and he tried to catch up in his patrol car. Before he could do so, he saw the defendant turn left at a “turn around lane” and head south on an intersecting street. There was a stop sign in the turn around lane, but the defendant did not heed it. Officer Gobeyn decided to stop the defendant for this violation.
On request, the defendant presented his driver’s license, registration, and proof of insurance. Gobeyn returned to his patrol car to verify those documents. As Gobeyn sat in his patrol car, he observed the defendant leaning over to the passenger side of his car. All he could see was the top left portion of the defendant’s shoulder. Fearing that the defendant might be reaching for a gun, Gobeyn radioed for backup assistance.
When a second officer arrived, the two ordered the defendant to get out of the car with his hands in plain view. A patdown search produced only a beeper. Gobeyn asked the defendant if there were any drugs or weapons in the vehicle. The defendant responded, “No, go ahead and look.” While the second officer *380detained the defendant outside, Gobeyn looked around the car’s front-seat area, and found a jacket in which he found a plastic baggie that contained several smaller baggies. Gobeyn thought, and this was later confirmed by stipulation at trial, that the large baggie contained cocaine residue. Further, as Gobeyn was leaving the car, he noticed a box beneath the driver’s seat. Inside the box he found a digital scale.
Gobeyn arrested the defendant on the basis of the discovery of the suspected cocaine residue in the baggie. Defendant was taken to the Oak Park police station and his car was towed and impounded there. An Oak Park police officer searched the defendant’s car early the next morning and found a yellow baggie that contained more than one hundred grams of cocaine hidden in the dashboard adjacent to the glove compartment. A subsequent, more careful search, found another baggie with a smaller amount of cocaine in the same area.
Defendant was charged with possession with intent to deliver 50 to 225 grams of cocaine. Before trial, the prosecutor advised defense counsel by letter of his intent to introduce evidence of a prior drug crime on an MRE 404(b) “other bad acts” theory.1 Defendant moved to suppress the evidence, citing People v Golochowicz, 413 Mich 298; 319 NW2d 518 (1982),2 *381arguing that it was irrelevant and that, if found to be relevant, it was unfairly prejudicial. In response, the prosecutor argued that the prior conviction was relevant to show defendant’s knowledge of the presence of the cocaine and his intent to deliver it, and that the probative value of the evidence was not substantially outweighed by the danger of unfair prejudice. The trial judge agreed and ruled that the prosecution could introduce evidence of defendant’s prior conviction, and announced his intention to read a limiting instruction to cushion the prejudicial effect on the defendant.
A four-day jury trial was held on May 17-21, 1993. Officer Gobeyn described the circumstances surrounding the defendant’s arrest. The officer who located the cocaine in the dashboard of the car described how he made that discovery.
Over defense objection, a third officer testified in great detail about the defendant’s 1988 crime.3 He testified that he had waited with a codefendant in that earlier case until the defendant and a codefendant appeared on the scene. He said that the defendant and the third man got out of their car and entered an apartment building. The defendant was carrying a distinctive plastic bag. After a few moments, the officer was invited into the apartment. Cocaine was taken from the bag and handed to the officer. After field testing it, the officer handed $5,000 to the defendant. He then gave a prearranged signal that brought in *382other officers to arrest all the participants. The defendant was discovered hiding in a bedroom closet.
The defendant did not testify, but his wife did. She said that both she and the defendant were employed. They were engaged but not yet married when the defendant was arrested. The defendant had been living in the Embassy Motel temporarily, after being evicted by a grandparent with whom he had been living. Further, she provided the testimonial basis for the defense theory that the defendant had been unaware of the cocaine hidden in the car’s dashboard. She testified that she had never seen the defendant use or sell drugs. He had purchased the old car in which he was arrested just five to ten days before his arrest and had loaned the car to others during that time. She had ridden in the car, but had never noticed anything hanging down from behind the dashboard.
The only other defense witness was an Oak Park police officer who searched the glove compartment and trunk of the defendant’s car while it was impounded. He found no incriminating evidence. The trunk contained several garbage bags full of men’s clothing. His testimony, therefore, was consistent with the defense theory that the defendant had been residing at the Embassy Motel temporarily.
The jury found the defendant guilty as charged of possessing with intent to deliver 50 to 225 grams of cocaine. The trial judge imposed a mandatory life sentence without the possibility of parole on June 17, 1993. The Court of Appeals affirmed in an unpublished memorandum opinion, holding merely that under People v Mouat, 194 Mich App 482; 487 NW2d 494 (1992), “the trial court did not abuse its discretion in allowing the prosecution to present evidence of *383prior convictions involving cocaine delivery where defendant’s intent to deliver cocaine in the instant case was an issue.”4 The defendant has appealed to this Court. We reverse.
m
In this case we once again consider the admissibility of other acts evidence under MRE 404(b). The general rule is more easily stated than applied: evidence of other crimes, wrongs, or acts of an individual is inadmissible to prove a propensity to commit such acts. MRE 404(b). Such evidence may be admissible, however, for other purposes under MRE 404(b)(1), which provides:
Evidence of other crimes, wrongs, or acts is not admissible to prove the character of a person in order to show action in conformity therewith. It may, however, be admissible for other purposes, such as proof of motive, opportunity, intent, preparation, scheme, plan, or system in doing an act, knowledge, identity, or absence of mistake or accident when the same is material, whether such other crimes, wrongs, or acts are contemporaneous with, or prior or subsequent to the conduct at issue in the case.
The decision whether such evidence is admissible is within the trial court’s discretion and will only be reversed where there has been a clear abuse of discretion. People v Bahoda, 448 Mich 261; 531 NW2d 659 (1995).
The character evidence prohibition is deeply rooted in our jurisprudence. Far from being a mere technicality, the rule “reflects and gives meaning to the central precept of our system of criminal justice, the pre*384sumption of innocence.” United States v Daniels, 248 US App DC 198, 205; 770 F2d 1111 (1985). Underlying the rule is the fear that a jury will convict the defendant inferentially on the basis of his bad character rather than because he is guilty beyond a reasonable doubt of the crime charged. Evidence of extrinsic bad acts thus carries the risk of prejudice, for it is antithetical to the precept that “a defendant starts his life afresh when he stands before a jury . . . .” People v Zackowitz, 254 NY 192, 197; 172 NE 466 (1930). As the United States Supreme Court recently noted in Old Chief v United States, 519 US 172, 181; 117 S Ct 644; 136 L Ed 2d 574 (1997), the problem with character evidence generally and prior bad acts evidence in particular is not that it is irrelevant, but, to the contrary, that using bad acts evidence can “ ‘weigh too much with the jury and ... so overpersuade them as to prejudge one with a bad general record and deny him a fair opportunity to defend against a particular charge.’ ” Quoting Michelson v United States, 335 US 469, 476; 69 S Ct 213; 93 L Ed 168 (1948). The fundamental principle of exclusion, codified by MRE 404(b), is woven into the fabric of Michigan jurisprudence:
There can be little doubt that an individual with a substantial criminal history is more likely to have committed a crime than is an individual free of past criminal activity. Nevertheless, in our system of jurisprudence, we try cases, rather than persons, and thus a jury may look only to the evidence of the events in question, not defendant’s prior acts in reaching its verdict. See United States v Mitchell, 2 US (2 Dall) 348, 357; 1 L Ed 410 (1795). [People v Allen, 429 Mich 558, 566-567; 420 NW2d 499 (1988).]
*385This Court’s most recent formulation of the MRE 404(b) test appears in People v VanderVliet, 444 Mich 52; 508 NW2d 114 (1993). In that case, we rejected a rigid, bright-line approach to other acts evidence and directed the bench and bar to employ the evidentiary safeguards already present in the Rules of Evidence, as identified by the United States Supreme Court in Huddleston:5
First, that the evidence be offered for a proper purpose under Rule 404(b); second, that it be relevant under Rule 402 as enforced through Rule 104(b); third, that the probative value of the evidence is not substantially outweighed by unfair prejudice; fourth, that the trial court may, upon request, provide a limiting instruction to the jury. [VanderVliet, supra at 55.]
Under this formulation, the prosecution bears the initial burden of establishing relevance of the evidence to prove a fact within one of the exceptions to the general exclusionary rule of MRE 404(b). Where the only relevance is to character or the defendant’s propensity to commit the crime, the evidence must be excluded. Where, however, the evidence also tends to prove some fact other than character, admissibility depends upon whether its probative value outweighs its prejudicial effect, taking into account the efficacy of a limiting instruction in cushioning the prejudicial effect of the evidence.
Thus, the first question that must be addressed is whether the prosecutor has articulated a proper noncharacter purpose for admission of the defend*386ant’s prior drug conviction.6 The prosecutor offers *387two theories of admissibility to support the admission of the defendant’s 1988 drug conviction. The first is to show that the defendant knew the cocaine was hidden in the dashboard of his car. The second is to show that he intended to deliver drugs in 1992. “Knowledge” and “intent” are indeed included among MRE 404(b)’s laundry list of proper purposes. However, a common pitfall in MRE 404(b) cases is the trial courts’ tendency to admit the prior misconduct evidence merely because it has been “offered” for one of the rule’s enumerated proper purposes. Mechanical recitation of “knowledge, intent, absence of mistake, etc.,” without explaining how the evidence relates to the recited purposes, is insufficient to justify admission under MRE 404(b). If it were, the prosecutor could routinely admit character evidence by simply calling it something else. Relevance is not an inherent characteristic, Huddleston, supra at 689, nor are prior bad acts intrinsically relevant to “motive, opportunity, intent, preparation, plan,” etc. Relevance is a relationship between the evidence and a material fact at issue that must be demonstrated by reasonable inferences that make a material fact at issue more probable or less probable than it would be without the evidence. United States v Sampson, 980 F2d 883, 888 (CA 3, *3881992). In order to ensure the defendant’s right to a fair trial, courts must vigilantly weed out character evidence that is disguised as something else. The logical relationship between the proffered evidence and the ultimate fact sought to be proven must be closely scrutinized.7
Logical relevance, the “touchstone” of the admissibility of prior acts evidence, is determined by the application of Rules 401 and 402.
MRE 401 provides:
“Relevant evidence” means evidence having any tendency to make the existence of any fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.
MRE 402 provides:
All relevant evidence is admissible, except as otherwise provided by the Constitution of the United States, the Constitution of the State of Michigan, these rules, or other rules adopted by the Supreme Court. Evidence which is not relevant is not admissible.
Pursuant to MRE 401, evidence is relevant if two components are present, materiality and probative value. Materiality is the requirement that the proffered evidence be related to “any fact that is of consequence” to the action. “In other words, is the fact to be proven truly in issue?” Wade & Strom, Michigan Courtroom Evidence (rev ed), Rule 401, p 71. A fact *389that is “of consequence” to the action is a material fact. People v McKinney, 410 Mich 413; 301 NW2d 824 (1981). “Materiality looks to the relation between the propositions for which the evidence is offered and the issues in the case. If the evidence is offered to help prove a proposition which is not a matter in issue, the evidence is immaterial.” 1 McCormick, Evidence (4th ed), § 185, p 773.
It is well established in Michigan that all elements of a criminal offense are “in issue” when a defendant enters a plea of not guilty. People v Mills, 450 Mich 61, 69; 537 NW2d 909 (1995). Because the prosecution must carry the burden of proving every element beyond a reasonable doubt, regardless of whether the defendant specifically disputes or offers to stipulate any of the elements, the elements of the offense are always “in issue” and, thus, material. See Old Chief, supra. The elements of the charge of possession with intent to deliver cocaine, as applied to this case, are as follows: (1) the defendant knowingly possessed a controlled substance; (2) the defendant intended to deliver this substance to someone else; (3) the substance possessed was cocaine and the defendant knew it was cocaine; and (4) the substance was in a mixture that weighed between 50 and 225 grams. See also CJI2d 12.3.
Because the prosecution was obligated under the statute to prove that the defendant knowingly possessed cocaine and that he did so with the specific intent of distributing it, knowledge and intent were “in issue.” The first step on the ladder of relevance, materiality, is thus reached.
The probative force inquiry asks whether the proffered evidence tends “to make the existence of any *390fact that is of consequence to the determination of the action more probable or less probable than it would be without the evidence.” The threshold is minimal: “any” tendency is sufficient probative force. MRE 401. See Beaubien v Cicotte, 12 Mich 459, 484 (1864), and Collins v Beecher & Marquette & Pacific Rolling Mill Co, 45 Mich 436, 438; 8 NW 97 (1881). In the context of prior acts evidence, however, MRE 404(b) stands as a sentinel at the gate: the proffered evidence truly must be probative of something other than the defendant’s propensity to commit the crime. If the prosecutor fails to weave a logical thread linking the prior act to the ultimate inference, the evidence must be excluded, notwithstanding its logical relevance to character.8
*391Turning to the present case, the question becomes whether the prosecutor carried its burden of demonstrating that the defendant’s prior conviction establishes some intermediate inference, other than the improper inference of character, which in turn is probative of the ultimate issues in this case, the defendant’s knowledge of the presence of cocaine and his intent to deliver it.9 We agree with the defendant that *392no such intermediate inference has been established.10
To establish the probativeness of the evidence, the prosecutor invokes the “doctrine of chances,” also known as the “doctrine of objective improbability.” 11
*393This theory, which is attributed to Professor Wig-more, is widely accepted, although its application varies with the issue for which it is offered. Where material to the issue of mens rea, as here, it rests on the premise that “the more often the defendant commits an actus reus, the less is the likelihood that the defendant acted accidentally or innocently.” Imwinkelried, Uncharged Misconduct Evidence, § 3:11, p 45. Consequently, the forbidden intermediate inference to defendant’s subjective character is not implicated:
[T]his theory of logical relevance does not depend on a character inference. The proponent is not asking the trier of fact to infer the defendant’s conduct (entertaining a particular mens rea) from the defendant’s personal, subjective character. The intermediate inference is an objective likelihood under the doctrine of chances rather than a subjective probability based on the defendant’s character. [Id., § 5:05, P 12.]
*394However, Imwinkelried cautions against the routine admission of prior misconduct evidence under the doctrine of chances because the theory is prone to abuse and may result in the admission of character evidence in disguise:
[The doctrine of chances theory] can easily be abused. . . . [I]ntent is an essential element of every true crime. Whenever the prosecutor has evidence of an uncharged crime similar to the charged offense, the prosecutor can attempt to invoke Wigmore’s doctrine of chances; the prosecutor can always argue that a similar uncharged crime triggers the doctrine of chances and is, therefore, logically relevant on a noncharacter theory both to disprove accident and thereby to prove mens rea. If the courts accept these arguments uncritically, the prosecutor may be able to introduce bad character evidence in disguise. ... To counter this tendency, the courts should clearly enunciate and rigorously enforce the foundational requirements applicable when the prosecutor relies on the doctrine of chances to establish mens rea. [Imwinkelried, The use of evidence of an accused’s uncharged misconduct to prove mens rea: The doctrines which threaten to engulf the character evidence prohibition, 51 Ohio St L J 575, 595 (1990).]
Elaborating on the foundational requirements for triggering the doctrine of chances to prove mens rea, Imwinkelried explains that the prosecutor must “make persuasive showings that each uncharged incident is similar to the charged offense and that the accused has been involved in such incidents more frequently than the typical person.”12 Id. at 602. We find *395this reasoning to be sound. The applicability of the doctrine of chances depends on the similarity between the defendant’s prior conviction and the crime for which he stands charged.13
We conclude that there is an insufficient factual nexus between the prior conviction and the present charged offense to warrant admission of the evidence *396under the doctrine of chances. The arresting officer from the 1988 offense testified at length about how he had waited with a codefendant in that earlier case until the defendant and a codefendant appeared on the scene. He said that the defendant and the third man got out of their car and entered an apartment building. The defendant was carrying a distinctive plastic bag. After a few moments, the officer was invited into the apartment. Cocaine was taken from the bag and handed to the officer. After field testing it, the officer handed $5,000 to the defendant. He then gave a prearranged signal that brought in other officers to arrest all the participants.
In this case, however, the defendant was not caught in the act of selling drugs. Rather, he was stopped for a routine traffic violation, which ultimately led to the discovery of cocaine hidden in the dashboard of his car. There was evidence at trial that the defendant had purchased the car just five to ten days before his arrest, and that the car had been in the possession of others during that time, lending support to the defense theoiy that the prior owner or someone else left the drugs in the car, unwittingly or in an attempt to frame the defendant. The plausibility of this defense was to be determined by the jury on the basis of its assessment of the credibility of the witnesses. However, the factual relationship between the 1988 crime and the charged offense was simply too remote for the jury to draw a permissible intermediate inference of the defendant’s mens rea in the present case. The facts of the 1988 drug offense simply do not bear out the prosecutor’s contention that the defendant “obviously knew” the drugs were in his dashboard and that he intended to deliver them. The *397prior conviction only demonstrates that the defendant has been around drugs in the past and, thus, is the kind of person who would knowingly possess and intend to deliver large amounts of cocaine. To the extent that the 1988 conviction is logically relevant to show that the defendant was also a drug dealer in 1992, we believe it does so solely by way of the forbidden intermediate inference of bad character that is specifically prohibited by MRE 404(b). Thus, the defendant’s prior conviction was mere character evidence masquerading as evidence of “knowledge” and “intent.” Because MRE 404(b) expressly prohibits the use of prior bad acts to demonstrate a defendant’s propensity to form a certain mens rea, we hold that the trial court abused its discretion in admitting evidence of the defendant’s prior conviction and reverse and remand the case for a new trial.14
IV
Even if we were to find that the evidence of the defendant’s prior conviction had some logical relevance distinct from the impermissible character inference, we would nevertheless conclude that it should have been excluded by MRE 403 because the danger of unfair prejudice substantially outweighed whatever *398marginal probative value it might have had. Rule 403 does not prohibit prejudicial evidence; only evidence that is unfairly so. Evidence is unfairly prejudicial when there exists a danger that marginally probative evidence will be given undue or preemptive weight by the jury. In the context of prior bad acts, that danger is prevalent. When a juror learns that a defendant has previously committed the same crime as that for which he is on trial, the risk is severe that the juror will use the evidence precisely for the purpose that it may not be considered, that is, as suggesting that the defendant is a bad person, a convicted criminal, and that if he “did it before he probably did it again.” People v Johnson, 27 F3d 1186, 1193 (CA 6, 1994). Because prior acts evidence carries with it a high risk of confusion and misuse, there is a heightened need for the careful application of the principles set forth in MRE 403.15 Id.
Applying the important principles of MRE 403 to the present case, we cannot escape the conclusion that the most powerful, if not the only, inference that the jury was likely to make from the prior conviction is the forbidden one: that because the defendant was convicted of selling cocaine in 1988, he must be guilty here. Thus, the specter of impermissible character evidence is likely to have significantly overshadowed any legitimate probative value. To use Justice Car*399dozo’s expression, we believe the “reverberating clang” of the evidence that the defendant sold drugs in 1988 drowned the “weaker sound” of the other evidence properly before the jury, leaving the jury to hear only the inference that if the defendant did it before, he probably did it again. Shepard v United States, 290 US 96, 104; 54 S Ct 22; 78 L Ed 196 (1933); United States v Merriweather, 78 F3d 1070, 1077 (CA 6, 1996). We hold that the evidence was substantially more prejudicial than probative and should not have been admitted.16
v
Finally, we consider whether the introduction of the defendant’s prior conviction constituted harmless error. Error requires reversal only if it is prejudicial. People v Mateo, 453 Mich 203, 215; 551 NW2d 891 (1996). The prejudice inquiry “focuses on the nature of the error and assesses its effect in light of the *400weight and strength of the untainted evidence.” Id. Our analysis in the foregoing section leads us to the inescapable conclusion that admission of the defendant’s prior conviction was not harmless.17 People v Gearns, 457 Mich 170; 577 NW2d 422 (1998).
CONCLUSION
We hold that on the facts of this case the trial court abused its discretion by admitting evidence of the defendant’s prior conviction because the prosecution failed to establish a proper purpose under MRE 404(b). We further hold that the error was not harmless. We therefore reverse the defendant’s conviction and remand the case to the trial court for further proceedings consistent with this opinion.
Mallett, C.J., and Cavanagh and Kelly, JJ., concurred with Brickley, J.In 1988, events that predate this 1992 case, the defendant participated in the delivery of a pound of cocaine to an undercover Oakland County police officer. The defendant ultimately pleaded guilty of delivering 225 to 650 grams of cocaine and conspiracy to commit the same offense. These were his first adult or juvenile offenses. His arrest in the present case occurred ten months after he was paroled on November 12, 1991.
The defendant’s trial in this case predated People v VanderVliet, 444 Mich 52; 508 NW2d 114 (1993), in which we articulated a clarified standard for determining the admission of prior acts evidence under MRE 404(b).
In an effort to preclude detailed testimony about the prior conviction, defense counsel offered at trial to stipulate to the prior conviction, as well as to the amounts and substance. The prosecutor persuaded the trial judge that the jury needed to hear the details of the prior crime in order to utilize MRE 404(b) properly. The suggested stipulation about the prior conviction also was read to the jury.
Issued October 10, 1995 (Docket No. 165956).
Huddleston v United States, 485 US 681; 108 S Ct 1496; 99 L Ed 2d 771 (1988).
The dissent vociferously objects to the proposition that the prosecutor must bear the burden of articulating a proper noncharacter purpose for the admission of prior acts evidence under MRE 404(b). However, the principle that the proponent of evidence bears the burden of establishing relevance and admissibility is a matter of basic hornbook law. See 22 Wright & Graham, Federal Practice & Procedure, § 5166, pp 65-76. Instead of fashioning an argument why the prosecutor should be relieved of this burden in MRE 404(b) cases, the dissent accuses the majority of somehow shifting or heightening the prosecutor’s burden under MRE 404(b). Our requirement that the prosecutor articulate a proper noncharacter purpose comports not only with the plain language of MRE 404(b), but also with the approach utilized by the majority of federal circuits. In United States v Sampson, 980 F2d 883 (CA 3, 1992), for example, the Third Circuit reversed the defendant’s conviction and remanded the case for a new trial because the government had failed to meet its burden of establishing a sufficient noncharacter purpose for admission of the defendant’s prior drug convictions. After expressing its concern that, although the proponents of Rule 404(b) evidence “will hardly admit it, the reasons proffered to admit prior act evidence may often be [P]otemkin village, because the motive, we suspect, is often mixed between an urge to show some other consequential fact as well as to impugn the defendant’s character,” Sampson at 886, the court described the contours of the government’s burden when seeking the admission of prior crimes:
If the government offers prior offense evidence, it must clearly articulate how that evidence fits into a chain of logical inferences, no link of which can be the inference that because the defendant committed drug offenses before, he therefore is more likely to have committed this one. . . . Thus, the burden on the government is not onerous. All that is needed is some showing of a proper relevance. Whereupon the trial court must judge the government’s proffered reason, the potential for confusion and abuse, and the significance of the evidence, and decide whether its probative value outweighs its prejudicial effect. [Id. at 887-888.]
Federal circuits similarly requiring the government to articulate, in a clear and logical manner, at least one proper noncharacter purpose for admission of prior acts evidence include: United States v Rackstraw, 7 F3d 1476, 1478-1479 (CA 10, 1993) (“When offering 404(b) evidence, the government ‘must articulate precisely the evidential hypothesis by which a fact of consequence may be inferred from the evidence of other acts.’ . . . Concomitantly, the trial court must identify specifically the purpose for which the evidence is admitted. . . . There must be a clear and logical connection between the ‘other acts’ evidence and the case being tried”); United States v Mayans, 17 F3d 1174, 1181 (CA 9, 1994) (“the government ‘must articulate precisely the evidential hypothesis by *387which a fact of consequence may be inferred from the other acts evidence’ ”); United States v Yeagin, 927 F2d 798, 803 (CA 5, 1991) (“A trial judge faced with the problem of admissibility of other crimes evidence should exercise caution and should require the government to explain why the evidence is relevant and necessary on a specific element that the government must prove”); United States v Arias-Montoya, 967 F2d 708, 713 (CA 1, 1992) (admission of prior acts evidence “ ‘is by no means a routine exercise and should not be accepted unless the government articulates with suitable precision the “special” ground for doing so’ ”); United States v Mothershed, 859 F2d 585, 589 (CA 8, 1988); United States v Zelinka, 862 F2d 92, 99 (CA 6, 1988).
Weinstein notes that although “courts on occasion have admitted other-acts evidence almost automatically, without any real analysis, if they find it fits within one of the categories specified in Rule 404(b). . . . Rule 404(b) does not authorize automatic admission, and the proponent of the evidence must demonstrate its relevance.” 2 Weinstein, Federal Evidence, § 404.20[3], pp 404-41 to 404-42.
We do not quarrel with the dissent’s assertion that VanderVliet embraced the “inclusionary” approach to prior misconduct evidence. However, the term “inclusionary” can be deceptive. The distinction between MRE 404(b) as a rule of “inclusion” as opposed to a rule of “exclusion” does not signify a shift to a more liberal policy toward the admission of prior act evidence. The “inclusionary” theory merely recognizes that similar acts can be admissible despite the inference to character so long as there is at least one proper noncharacter-based inference linking the prior act to the ultimate inference, and that, contrary to the position taken at common law, Rule 404(b)’s list of proper purposes is nonexclusive. As Wright and Graham explain in their influential treatise on the Federal Rules of Evidence:
The common law rule has been described as “an unctuous, but easily circumvented rule of exclusion.” Wigmore stated it this way: “The doing of another criminal act, not a part of the issue, is . . . not admissible as evidence of the doing of the criminal act charged, except when offered for the specific purpose of evidencing Design, Plan, Motive, Identity, Intent, or other relevant fact . . . distinct from Moral Character.” While the general rule of exclusion is often applauded- — and occasionally enforced — it is the exceptions that are of most practical significance. Courts have found it difficult to distinguish the rule from the exceptions and Professor Stone, in two influential articles, argued that the true rule was the reverse of the one stated by Wigmore; i.e., evidence of other crimes is admissible except when it proves nothing but the *391propensity of the defendant to engage in criminal conduct. Professor Stone’s version of the rule — -usually referred to as the “inclusionary” or “positive” rule as against Wigmore’s “exclusionary” or “negative” formulation — has been favored by many commentators and adopted by some courts, but the majority rule at common law took the form stated by Wigmore. [22 Wright & Graham, Federal Practice & Procedure, § 5239, pp 428-432.]
The “inclusionary” theory thus recognizes the rule’s restrictive application to evidence offered solely to prove criminal propensity, a point that was settled before VanderVliet, but does not require courts to err on the side of admission, or to admit evidence that would not have been admissible before VanderVliet. See United States v Figueroa, 618 F2d 934, 939, n 2 (CA 2, 1980) (“The exclusionary approach to similar act evidence obliges the trial court to determine whether the issue sought to be proved is among the traditional exceptions to the rule barring prior act evidence; the inclusionary approach permits the evidence to be used to prove any issue other than propensity, but the trial court is still obliged to ask, ‘Is the evidence in any way relevant to a fact in issue otherwise than by merely showing propensity?’ ” Stone, The rule of exclusion of similar fact evidence: America, 51 Harv L R 988, 1004 [1938]).
Imwinkelried, whom the dissent, post at 417, ironically accuses the majority of citing selectively, expressly rejects, as contrary to the plain language of the rule, the dissent’s contention that “an intermediate noncharacter inference is required only where the prior act is offered to prove conduct”:
The second sentence of Rule 404(b) refers to “identity” as a permissible use of uncharged misconduct evidence. However, no one would suggest that Rule 404(b) permits the prosecutor to invite the jury to reach the conclusion of identity through an intermediate inference of character. The second sentence also alludes to “intent.” By parity urging of reasoning, the rule should be construed as prohibiting the prosecutor from urging the jury to reach that conclusion through an intermediate inference of the defendant’s disposition of a certain mens rea. The better-reasoned cases *392require the prosecutor to reach the ultimate inference of mens rea without relying on an intermediate inference of character. [Imwinkelried, Uncharged Misconduct Evidence, § 5:01.50, p 4.]
Under the dissent’s view, prior convictions offered to prove a defendant’s state of mind are admissible as a matter of course, subject only to MRE 403 balancing. Given that MRE 403 already applies to all offers of evidence, however, it strains logic to suggest that MRE 404(b) erects no barrier to the admission of such evidence.
The rationale for the exclusion of character evidence to establish mens rea, as described by Imwinkelried, is twofold:
One is that if the jury must consciously focus on the issue of the defendant’s character, there is a grave risk that at a subconscious level, they will be tempted to punish the defendant for his or her uncharged misconduct. That risk is certainly present when the jury addresses the question of whether the defendant has a propensity for forming a mens rea. The jury might be repulsed by the defendant’s “criminal mind.” The other probative danger underpinning the character evidence prohibition is the risk that the jurors will overestimate the probative value of character as a predictor of conduct. That risk is also applicable when the jury reasons from the defendant’s propensity for a mens rea to the conclusion that the defendant had a guilty state of mind on the charged occasion. The cognitive and volitional aspects of conduct account in large part for the unpredictability of human behavior. [Imwinkelried, n 9 supra, § 5:01.50, pp 3-4. See People v Schweitzer, 23 Mich 301, 304 (1871).]
We infer the prosecution’s reliance on the doctrine of chances from his opening and closing statements. In his opening statement, the prosecutor explained the relevance of the defendant’s prior conviction as follows:
An additional piece of evidence that you’re going to hear, ladies and gentlemen, will be testimony from other officers, other officers who came into contact with the Defendant in 1988. And at that time, you will hear testimony from Oakland County Narcotics Enforcement Team officers who purchased one-half kilogram or 500 grams of cocaine from the Defendant in June of 1988. . . . [T]hat evidence is going to show that the Defendant obviously knew that the cocaine was in his car, that somebody didn’t leave this valuable amount, this much cocaine in his car without his *393knowing it. And it will show that the defendant, at the time he possessed this cocaine, intended to deliver it, that he intended to deliver the cocaine in 1988.
And in closing argument:
[D]id [Defendant] know the cocaine was there? In 1988, he delivered a half kilogram of cocaine to Deputy Moore at the Diplomat Towers Apartments in Southfield, Michigan. It’s common sense, ladies and gentlemen. Based upon these facts that are admitted into evidence, they clearly show that the Defendant knew the cocaine was there.
And later:
[T]he fact that he was convicted in 1988 for delivery of a half kilogram of cocaine, 500 grams of cocaine, inescapably leads to the conclusion that he knew the cocaine was there.
Weinstein states:
In each individual case, the trial judge must decide whether the proffered evidence tends to make the consequential fact more or less probable. If the connection between the other crime and the *395charged crime is strong, admission may be appropriate. If the connection is weak, exclusion is generally sound. If enough time has passed, the other crime may be of such attenuated probative value as to warrant exclusion. [2 Weinstein, n 7 supra at § 404.21(2)(c), pp 404-54 to 404-55.]
A simple analogy will prove the point. If the prosecutor were “offering” evidence of a prior arson conviction to prove that the defendant knowingly possessed cocaine with the intent to deliver, even the dissent would likely concede that the offer would fail the initial test of relevancy on the ground that the two acts were too dissimilar. Under this scenario, the evidence is inadmissible even though it is “offered” for a “proper purpose” under Rule 404(b), that is, to prove knowledge and intent. If, however, defendant’s prior crime involved the concealment of drugs in the dashboard of his car, that evidence would likely be admissible under the doctrine of chances because of the stark similarity of the two crimes. There is, then, a continuum upon which each proffered prior act must be placed; the more similar the prior act to the charged crime, the closer the evidence to the admissibility threshold. Cases cited by the prosecution and the dissent that have admitted prior crimes to prove knowledge and intent have similarly required a close factual nexus between the prior and charged crimes. See, e.g., United States v Rackstraw, n 6 supra at 1479 (“We have long recognized the relevance of prior crimes in the context of narcotics violations where the uncharged misconduct is close in time and similar in method to the charged scheme and where knowledge . . . was at issue”); United States v Garcia, 983 F2d 1160, 1173 (CA 1, 1993) (“In this case, there is a close nexus between the past act and the current charges”); United States v Hernandez, 84 F3d 931, 935 (CA 7, 1996) (“we agree that Hernandez’ prior conviction satisfied the similarity/proximity requirements”); United States v Adrian, 978 F2d 486, 493 (CA 9, 1992) (“It does not appear from the record whether the prior convictions were similar enough and proximate enough in time to the present offenses to make them highly relevant to appellee’s intent”); see also United States v Hernandez-Miranda, 601 F2d 1104, 1108 (CA 9, 1979) (“When a prior criminal act is relied upon to prove intent or knowledge, similarity between the two events must be shown to establish the threshold requirement of relevance”).
Rather than attempting to explain the specific logical progression that makes either knowledge or intent more likely in light of the defendant’s prior crime, the dissent simply argues that the involvement of cocaine in defendant’s 1988 conviction alone “makes [it] more probable than not that defendant knowingly possessed drugs with the intent to distribute them” in the present case. Post at 414. This is a poorly disguised propensity argument, which is precisely what Rule 404(b) expressly forbids. United States v Betts, 16 F3d 748, 759 (CA 7, 1994); People v Beasley, 809 F2d 1273 (CA 7, 1987); Arias-Montoya, n 6 supra; United States v Lynn, 856 F2d 430 (CA 1, 1988); United States v Mehrmanesh, 689 F2d 822 (CA 9, 1982).
The dissent urges us to defer to the trial court’s balancing of prejudicial effect versus probative force because our “rejection of the trial court’s balancing under Rule 403” is simply a substitution of our “judgment for that of the trial court.” Post at 439. However, the trial court’s misapplication of MRE 404(b) caused it to overvalue the probativeness of the evidence. Having held that the evidence had no relevance apart from the impermissible character inference, we refuse to abdicate our responsibility to review the circuit court’s balancing under MRE 403 and to reverse that ruling where necessary to prevent injustice.
The prosecutor and dissent argue that any undue prejudice was offset by the following cautionary instruction given by the judge at the conclusion of the trial:
You have heard evidence that was introduced to show that the Defendant committed a crime for which he is not on trial. If you believe this evidence, you must be very careful only to consider it for certain purposes. In this case, you may only think about whether this evidence tends to show that the Defendant specifically meant to or intended to commit the crime as charged, which I will give you the elements of in a moment. You must not consider this evidence for any other purpose.
While a limiting instruction will often suffice to enable the jury to compartmentalize evidence and consider it only for its proper purpose, we note the absence of a proper purpose in this case to which the jury could limit its use of the evidence. Although the prior conviction was purportedly “offered” to prove knowledge and intent, the only real relevance of the prior conviction was to depict the defendant as a drug dealer. Against this, the limiting instruction was not limiting at all. See United States v Wright, 901 F2d 68, 70 (CA 7, 1990).
The dissent’s statement that “[a]ny error in admitting the prior acts evidence was slight or negligible” ignores the fact that the prosecution was allowed to parade the defendant’s prior conviction before the jury constantly throughout this brief trial. Post at 443. The jury heard detailed information about the prior arrest in the prosecution’s opening and closing statements, the direct examination of the arresting officer from 1988, and in the cross-examination of the defendant’s wife. This repetition greatly enhanced the danger of unfair prejudice arising from the admission of defendant’s prior conviction. The decision of the First Circuit in Arias-Montoya, n 6 supra, is instructive on this point. In a case that is remarkably similar to the case at bar, the court ruled that evidence of the defendant’s prior drug conviction was improperly admitted, but that the admission of the evidence constituted harmless error. The court found “particularly significant” the fact that, “aside from eliciting two lines of testimony about the 1983 conviction, the prosecutor made no further reference to it, either at trial or ... in his closing.” Id. at 714.