UNITED STATES of America v. Terrell DAVIS, Appellant, 726 F.3d 434

OPINION SMITH, Circuit Judge. Police arrested Terrell Davis after finding him in a Jeep with nearly a kilo of cocaine in the backseat. The arrest led to a conviction for possession with intent to distribute. As evidence that Davis recognized the cocaine in the Jeep, the government proved at trial that he had two prior convictions for possessing cocaine. Yet the government never proved that the cocaine from his past was similar in appearance, quantity, or form. We acknowledge that some of our cases admitting prior criminal acts under Federal Rule of Evidence 404(

b) have been expansive. But our expansiveness is. finite, and this case crosses the line. We will vacate Davis’s conviction and remand. I The events at issue took place on a wintry afternoon over two years ago. Two Philadelphia police officers were patrolling near 5100 Market Street — roughly four miles west of Independence Hall and the Liberty Bell. This is a dangerous part of the city where drug deals and robberies are commonplace. Officer Clifford Gilliam parked his patrol car, and Officer Shawn Witherspoon joined him on foot.

On the opposite side of the street, the officers spotted a black Jeep Grand Cherokee, later determined to be from Enterprise Rent-A-Car. Inside were two men, Terrell Davis and Jamar Blackshear. The Jeep’s engine was running but nothing seemed amiss. After a period of time, Davis and Black-shear began to act suspiciously. They reached toward each other with “body motions [that] were consistent with the exchanging of narcotics in a narcotics transaction.” B.A. 8. The officers exited their patrol car and approached the Jeep.

Upon noticing the officers, Davis and Blackshear had “expressions of shock on their faces,” B.A. 8, and they tossed something into the backseat. They exited the Jeep and quickly walked away — so quickly, in fact, that Blackshear did not bother closing his door. Officer Gilliam stopped Blackshear and patted him down to search for weapons. He instead found a wad of cash in his pocket. In the meantime, Officer Wither-spoon stopped Davis and patted him down. He found a similar amount of cash.

Everything indicated to the officers that this was a drug deal: the suspicious movements, the hurried departures, the wads of cash, and the neighborhood itself. Knowing that guns often accompany drug deals, the officers decided to search the Jeep for weapons — and to see if there were any other occupants. Officer Witherspoon tried to look through the tinted rear window, but it was too dark. So he opened the already-ajar driver’s door and saw a handgun wedged between the driver’s seat and the middle console. At that point, the officers arrested Davis and Blackshear and placed them in the patrol car.

The handgun was not the only item in the Jeep. Officer Witherspoon returned and spotted an opaque shopping bag in the backseat. It was open and contained a white substance. The officers requested a drug-detection dog, which alerted to the presence of drugs. The officers obtained a warrant and recovered ten cell phones, a pair of binoculars, and two shopping bags with roughly 740 grams of cocaine distributed among nine smaller Ziploc bags. The cocaine itself was compressed into the shape of a brick and had a street value over $75,000.

Davis and Blackshear were charged with possessing a controlled substance with intent to distribute under 21 U.S.C. § 841(a)(1) and with possessing a firearm in furtherance of a drug-trafficking crime under 18 U.S.C. § 924(c). They were also charged with aiding and abetting under 18 U.S.C. § 2. Davis and Blackshear filed a motion to suppress all evidence from the Jeep. They argued that because the Jeep’s front driver’s side window was tinted, the officers could not have seen the alleged reaching, gawking, and tossing — and so they could not have had any cause for suspicion in the first place.

The District Court inspected the Jeep and discovered that the window was in fact tinted. The Court nonetheless denied the suppression motion. It credited the testimony of the officers who said that the window had been tint-free on the day of the arrests eight months earlier. It also credited the testimony of an Enterprise employee who said that neither Enterprise nor the manufacturer had tinted the windows and that since the arrests over fifty people had rented the Jeep. The defendants then pursued separate paths. Blackshear pled guilty but reserved the right to appeal the denial of his suppression motion.

He received two consecutive sixty-month sentences plus four years of supervised release. Davis opted for a jury trial. As the trial approached, the government asked permission to introduce Davis’s two prior convictions for possessing cocaine. The District Court consented, stating that the convictions were admissible under Federal Rule of Evidence 404(

b) to show that Davis recognized the drugs in the Jeep. At trial, the jury heard testimony from a range of witnesses, including Officers Gilliam and Wither-spoon; Keith Festus, the owner of a nearby cell-phone store; and a narcotics expert. The jury ultimately found Davis guilty of the drug crime but not guilty of the gun crime. He received a seventy-eight-month sentence plus four years of supervised release. Davis raises four issues on appeal: the denial of his suppression motion, the admission of his prior convictions, and two other evidentiary issues.

II Davis’s first argument is that the officers illegally stopped him after he exited the Jeep. This would make the cocaine inadmissible as the product of an illegal seizure. “Where a motion to suppress has been denied, we review the order for clear error as to the underlying facts, but exercise plenary review as to its legality in the light of the court’s properly found facts.” United States v. Brownlee, 454 F.3d 131, 137 (3d Cir.2006) (quotation marks omitted). The District Court rejected Davis’s constitutional argument, and with good reason.

The Fourth Amendment prohibits “unreasonable searches and seizures.” U.S. Const, amend. IV. The general rule is that a search or seizure is unreasonable if the police lack either probable cause or a warrant — though courts have created several exceptions to the warrant requirement. See Kentucky v. King, — U.S. -, 131 S.Ct. 1849, 1856, 179 L.Ed.2d 865 (2011) (noting that the ultimate touchstone is “reasonableness”).

Over the past few decades, the Supreme Court has created a broad exception to both requirements: “an officer may, consistent with the Fourth Amendment, conduct a brief, investigatory stop when the officer has a reasonable, articulable suspicion that criminal activity is afoot.” Illinois v. Wardlow, 528 U.S. 119, 123, 120 S.Ct. 673, 145 L.Ed.2d 570 (2000) (citing Terry v. Ohio, 392 U.S. 1, 30, 88 S.Ct. 1868, 20 L.Ed.2d 889 (1968)).

This exception also allows officers to search the passenger area of a vehicle without probable cause or a warrant if they conduct a lawful stop and reasonably believe that the suspect is dangerous and has a weapon inside. See Michigan v. Long, 463 U.S. 1032, 1049-50, 103 S.Ct. 3469, 77 L.Ed.2d 1201 (1983). Davis’s constitutional argument turns on whether the officers had a reasonable suspicion when they first stopped him.

And that turns on whether the front driver’s side window was tinted at the time of the arrests — for if it was, the officers could not have seen through it, and they would have lacked any reason to suspect an illicit transaction. Though the District Court inspected the window at the suppression hearing and saw that it was tinted, the Court found that it was not tinted on the day of the arrests eight months earlier. That finding was not clearly erroneous. For one thing, Officers Gilliam and Witherspoon both testified that the front driver’s side window, unlike the rear window, was tint-free when they saw the Jeep.

And an Enterprise employee testified that neither Enterprise nor the manufacturer had tinted the window and that over fifty people had rented the car between the arrests and the hearing. He also said that another renter could have been responsible for the tint. “Anybody could have put it on.” J.A. 325. This testimony supports the District Court’s finding. To be sure, the police took a picture of the Jeep on the day of the arrests, and the front and rear windows appear to have the same tint. But the picture was taken at night in low lighting.

Davis also points to a witness who testified at the suppression hearing that the window “probably was a little tinted.” J.A. 385. But the District Court found that his testimony was not “particularly credible” for “a number of reasons.” B.A. 6 n. 3. The record thus contains no evidence that plainly contradicts the officers’ testimony. And “when the district court’s decision is based on testimony that is coherent and plausible, not internally inconsistent and not contradicted by external evidence, there can almost never be a finding of clear error.” United States v. Igbonwa, 120 F.3d 437, 441 (3d Cir.1997).

As a result, the District Court did not clearly err when it found that the window was tint-free. Nor did the District Court err in concluding that the officers had a reasonable suspicion to stop Davis. The officers observed odd behavior through the front window—an exchange, shocked expressions, and tossing motions. Davis and Blackshear rapidly left the car and began walking away, the latter failing to close the car door. And the activity took place in a high-crime area. The officers thus had a reasonable suspicion that a crime might be afoot.

See Terry, 392 U.S. at 30, 88 S.Ct. 1868; Wardlow, 528 U.S. at 124, 120 S.Ct. 673 (recognizing presence in a “high crime area,” “unprovoked flight,” and “nervous, evasive behavior” as factors supporting a reasonable suspicion); see also United States v. Bonner, 363 F.3d 213, 218 (3d Cir.2004) (concluding that flight from a traffic stop creates a reasonable suspicion). The officers also had authority to search the Jeep. The pat downs revealed large wads of cash, suggesting that Davis and Blackshear were in the middle of a drug deal.

Because drug dealers often carry guns, the officers had “a reasonable belief based on specific and articulable facts” that Davis and Blackshear were dangerous and might have weapons inside the Jeep. Long, 463 U.S. at 1049, 103 S.Ct. 3469 (quotation marks omitted); see also United States v. Arvizu, 534 U.S. 266, 273, 122 S.Ct. 744, 151 L.Ed.2d 740 (2002). That belief allowed the officers to search the Jeep for weapons.

During the search, they found something else—cocaine in the backseat—but they “clearly cannot be required to ignore the contraband” discovered “while conducting a legitimate Terry search of the interior of the automobile.” Long, 463 U.S. at 1050, 103 S.Ct. 3469. For these reasons, we will affirm the denial of Davis’s suppression motion. Ill ? second argument is that the District Court erred in admitting his two prior convictions for possessing cocaine. We review that decision for an abuse of discretion. United States v.

Butch, 256 F.3d 171, 175 (3d Cir.2001) (noting that a decision is an abuse of discretion if “clearly contrary to reason and not justified by the evidence” (quotation marks omitted)). Though we have held that some prior drug convictions are admissible under Federal Rule of Evidence of 404(b), we have never held that a possession conviction is admissible to show knowledge or intent in a distribution trial. We decline to do so today. A American courts have long excluded evidence of a person’s prior bad acts.

This tradition reflects a fear that the jury will place too much weight on past crimes and prior misdeeds. “[I]t is said to weigh too much with the jury and to so overpersuade them as to prejudice one with a bad general record and deny [the accused] a fair opportunity to defend against a particular charge.” Michelson v. United States, 335 U.S. 469, 476, 69 S.Ct. 213, 93 L.Ed. 168 (1948); see also H. Richard Uviller, Evidence of Character to Prove Conduct: Illu sion, Illogic, and Injustice in the Courtroom, 130 U. Pa.

L.Rev. 845, 884 (1982) (“[A]s the special conditions of predictive value coalesce, the potential for prejudice also rises.”). The risk is that jurors will focus on evidence of prior acts, believing that someone with a criminal record cannot change and discounting any evidence to the contrary. Over the past two hundred years, the prior-acts rule has changed much in form but little in function. In the early days of the common law, courts used an inclusionary approach: evidence of prior acts was presumptively admissible unless it was relevant only to the defendant’s propensity to commit a crime.

See Julius Stone, The Rule of Exclusion of Similar Fact Evidence: America, 51 Harv. L.Rev. 988, 989-90 (1938). In the nineteenth century, the rule slowly became exclusionary: such evidence was presumptively inadmissible unless the proponent could show that it was relevant to one of several specific purposes, such as motive or intent. See id. at 990-93 (concluding that American courts applied this rule on the mistaken belief that the exclusionary approach was part of the English common law). But that trend faded, and courts began to use different approaches — some inclusionary, some exclusionary.

See United States v. Long, 574 F.2d 761, 765-66 (3d Cir.1978) (noting the division of authorities). The Federal Rules of Evidence settled the matter in 1975, establishing a uniform inclusionary approach. Id.; United States v. Green, 617 F.3d 233, 244 (3d Cir.2010). Yet this change, “like the nineteenth century switch from the inclusionary to the exclusionary approach, did not give rise to any significant change in the admissibility of such evidence.” Kenneth J. Melilli, The Character Evidence Rule Revisited, 1998 B.Y.U. L.Rev. 1547,1560.

The modern approach is set forth in Federal Rule of Evidence 404(b). “Evidence of a crime, wrong, or other act is not admissible to prove a person’s character in order to show that on a particular occasion the person acted in accordance with the character.” Fed.R.Evid. 404(b)(1). That principle seems strict, but prior-acts evidence “may be admissible for another purpose, such as proving motive, opportunity, intent, preparation, plan, knowledge, identity, absence of mistake, or lack of accident.” Fed.R.Evid. 404(b)(2). Uneontroversial at the time of adoption, Rule 404(

b) has become the most cited evidentiary rule on appeal. See Thomas J. Reed, Admitting the Accused’s Criminal History: The Trouble with Rule k0k(L), 78 Temp. L.Rev. 201, 211 (2005). The text of Rule 404(

b) has led to a four-part test. Prior-acts evidence is admissible only if it is (1) offered for a proper purpose under Rule 404(b)(2); (2) relevant to that purpose; (3) sufficiently probative under the Rule 403 balancing requirement; and (4) accompanied by a limiting instruction, if requested. See Green, 617 F.3d at 249; see also Huddleston v. United States, 485 U.S. 681, 691-92, 108 S.Ct. 1496, 99 L.Ed.2d 771 (1988) (discussing these four requirements). All this really means is that such evidence must have a nonpropensity purpose and satisfy the same relevancy requirements as any other evidence.

And yet the relevancy requirements pose problems of their own in this context. Indeed, the problems are in many cases insurmountable. See Uviller, 130 U. Pa. L.Rev. at 878 (“The test of ordinary relevance is often an insuperable barrier.”). For starters, the prior-acts evidence must be relevant to a proper purpose, and it must be relevant without requiring the factfinder to make a propensity inference. See United States v.

Sampson, 980 F.2d 883, 887 (3d Cir.1992) (“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.”). Consider a defendant who has been convicted of manslaughter. In a later assault prosecution, the government might want to use the conviction, perhaps to prove intent.

But that use is off limits if the only reason the conviction is relevant to intent is the inference that because the defendant has committed manslaughter before, he must have committed assault now. See id. at 887-88. In addition, the conviction must be relevant based on what the factfinder knows about the prior act. So even if the defendant was convicted of intentional manslaughter, the conviction will be relevant to intent only if the jury knows the act was intentional and not reckless or negligent. That is why the use of prior-acts evidence requires care from prosecutors and judges alike.

In proffering such evidence, the government must explain how it fits into a chain of inferences — a chain that connects the evidence to a proper purpose, no link of which is a forbidden propensity inference. Id. at 887. And then the “district court, if it admits the evidence, must in the first instance, rather than the appellate court in retrospect, articulate reasons why the evidence also goes to show something other than character.” Id. at 888. The reasoning should be detailed and on the record; a mere recitation of the purposes in Rule 404(b)(2) is insufficient.

Unfortunately, these requirements are “so often honored in the breach” that they resonate “about as loudly as the proverbial tree that no one heard fall in the forest.” United States v. Givan, 320 F.3d 452, 466 (3d Cir.2003) (McKee, J., dissenting). B With these principles in mind, we conclude that Davis’s convictions for possessing cocaine were inadmissible to prove knowledge or intent in his trial for possessing with intent to distribute. The District Court abused its discretion by admitting this evidence, and we will vacate Davis’s conviction.

Davis was twice convicted of possessing cocaine under Pennsylvania law — once in 2007 and once in 2008. The government filed a motion to introduce these convictions, advancing a pentad of purposes. J.A. 18 (“This evidence is relevant to prove the defendant’s plan to, knowledge of, and intent to distribute and/or possess cocaine, and absence of mistake or accident.”).

To its credit, the able District Court admitted the convictions as relevant to a single purpose: “Clearly, evidence of his prior convictions for possession of crack cocaine makes it more likely than not that Davis knew that the white substance in the plastic bag on the back seat of the Jeep was cocaine.” D.A. 18. The government now argues on appeal that the evidence also was relevant to intent. See Appellee Br. at 45. Knowledge and intent are indeed proper purposes under the first part of our Rule 404(

b) test. And “[t]here is no question that, given a proper purpose and reasoning, drug convictions are admissible in a trial where the defendant is charged with a drug offense.” Sampson, 980 F.2d at 887. We have held, for example, that evidence of past distribution is relevant to prove knowledge of the same or different drug in a later distribution trial. E.g., Givan, 320 F.3d at 461 (“The evidence that Givan had been convicted of distribution of cocaine makes Givan’s knowledge of the presence of the heroin more probable than it would have been without the evidence.”); United States v.

Boone, 279 F.3d 163, 187 (3d Cir.2002) (considering the defendant’s past cocaine-distribution acts as evidence that he was not “an ignorant ‘go-fer’ ”); cf. United States v. Vega, 285 F.3d 256, 263 (3d Cir.2002) (“[E]vidence of Vega’s participation in a prior drug conspiracy is probative of his knowledge of, and relationship with a member of, a later drug conspiracy.”). And we have held that evidence of past distribution is relevant to prove intent to distribute in a later distribution trial. E.g., United States v.

Lee, 573 F.3d 155, 166 (3d Cir.2009) (“Lee’s prior drug trafficking conviction was properly admitted as evidence that Lee intended to distribute any drugs in his possession.”); Givan, 320 F.3d at 461; Boone, 279 F.3d at 187. We have even held that evidence of past distribution is relevant to prove knowledge of a different drug in a later possession trial. United States v. Lopez, 340 F.3d 169, 174 (3d Cir.2003). But we have never held that a possession conviction is relevant to prove either knowledge or intent in a distribution trial, and rightly so. 1. Knowledge.

Possession and distribution are different in ways that matter— something that both the District Court and the government failed to appreciate. As to knowledge, one who possesses a drug might not recognize the same drug when prepared for distribution. The packaging or quantity might be different, and objects in greater quantities often have an appearance or smell of their own. Take water, which is transparent by the drop but blue in the ocean, or powdered sugar, which is floury on a donut but dense in a bag. In this case, the jury knew only that Davis had been twice convicted of possessing cocaine.

See Appellee Br. at 19 n. 3. The jury knew nothing of the packaging or quantity that led to those convictions, so it could not have known whether Davis’s past helped him to recognize the nearly one kilogram of cocaine in the Jeep. Then there is the problem that the cocaine from Davis’s past might have been in a different form. Cocaine is consumable either as a powder or as one of several bases, most often crack. See DePierre v. United States, — U.S. -, 131 S.Ct. 2225, 2228-29, 180 L.Ed.2d 114 (2011). Neither form particularly resembles the other.

As its name suggests, powder cocaine is a powder — specifically, a salt— that can be compressed or loose. See David A. Sklansky, Cocaine, Race, and Equal Protection, 47 Stan. L.Rev. 1283, 1290-91 (1995). On the other hand, crack cocaine is hard and waxy and often resembles small rocks or crystals. See id. This distinction matters, and the jury did not know which form Davis had possessed back in 2007 and 2008. For all the jury knew, the cocaine could have been a dash of powder on a golden tray. It could have been hidden in the lining of a suitcase.

Or it could have been crack cocaine — in crystal form, in liquid form, rolled up in paper, or stuffed in a syringe. In any of those instances, Davis’s past would not have helped him to identify the compressed powder in the backseat. The two prior convictions thus fail the second part of our Rule 404(

b) test, the relevancy requirement. See Fed.R.Evid. 401 (explaining that evidence is relevant if it is probative of a consequential fact). Based on the bare-bones stipulation before it, the jury had no way of knowing whether Davis’s experiences made him any more likely to recognize the cocaine in the backseat. The convictions simply were not probative of Davis’s knowledge. See Givan, 320 F.3d at 466 (McKee, J., dissenting) (noting the difficulty when “there is absolutely nothing on this record that would allow the jury to make any meaningful or relevant comparison” between past and present drugs).

At best, the convictions had such limited probative value that they fail the third part of our test, the balancing requirement. See Fed.R.Evid. 403 (allowing courts to “exclude relevant evidence if its probative value is substantially outweighed by a danger of ... unfair prejudice.”). Either way, the convictions are inadmissible to prove Davis’s knowledge. The government nonetheless urges us to follow Lopez and Givan. In Lopez, we held that the defendant’s participation in a cocaine-distribution conspiracy was admissible in a possession trial to prove knowledge of heroin, a different drug altogether.

Lopez, 340 F.3d at 174 (“[The conviction] was admissible for the purpose of rebutting the defendant’s anticipated claim of innocent association with, and lack of knowledge of, the heroin found near his bunk.”). And in Givan, we held that the defendant’s conviction for distributing cocaine was likewise admissible to prove knowledge and intent in a heroin-distribution trial. Givan, 320 F.3d at 461. These cases are at the outer bounds of admissibility under Rule 404(b).

See David Culberg, Note, The Accused’s Bad Character: Theory and Practice, 84 Notre Dame L.Rev. 1343, 1358-59 & n. 83 (2009) (criticizing Lopez and Givan). At all events, the two cases are distinguishable because the defendants had been convicted of dealing cocaine, and drug dealers presumably have more knowledge of drugs in general. By contrast, a possession conviction does not imply a similar level of knowledge. 2. Intent. Nor does a possession conviction imply an intent to distribute.

Possession and distribution are distinct acts— far more people use drugs than sell them — and these acts have different purposes and risks. A prior conviction for possessing drugs by no means suggests that the defendant intends to distribute them in the future. “Acts related to the personal use of a controlled substance are of a wholly different order than acts involving the distribution of a controlled substance. One activity involves the personal abuse of narcotics.” United States v. Ono, 918 F.2d 1462, 1465 (9th Cir.1990). The other usually involves “the implementation of a commercial activity for profit.” Id.

As a result of these differences, Davis’s convictions again fail the second part of our Rule 404(

b) test. In cases such as this, there is an ever-present danger that jurors will infer that the defendant’s character made him more likely to sell the drugs in his possession. But that is precisely the type of inference that Rule 404(

b) forbids. Any other conclusion would run the risk of unraveling the prior-acts rule: [I]f the act of possessing or using marijuana is to be admissible to prove intent to transport and sell marijuana, or, to go even further, to prove intent to transport and sell a different drug, then there is no reason why participation in any drug-related crime could not be used to prove intent to engage in any other drug-related crime, or why any robbery could not be used to prove the requisite intent with respect to any other robbery. A rule allowing such evidence would eviscerate almost entirely the character evidence rule.

David P. Leonard, The New Wigmore. A Treatise on Evidence: Evidence of Other Misconduct and Similar Events § 7.5.2(d); see also Charles Alan Wright & Kenneth W. Graham, Jr., 22A Federal Practice and Procedure: Evidence § 5242 (2d ed. 2013) (“[T]he routine use of [the intent] exception [under Rule 404(b) ] could easily destroy the exclusionary rule.”). We join other circuits in declaring that a possession conviction is inadmissible to prove intent to distribute.

The Sixth Circuit, for example, held that “possession of a small quantity of crack cocaine for personal use on one occasion ... sheds no light on whether [the defendant] intended to distribute crack cocaine in his possession on another occasion nearly five months earlier.” United States v. Haywood, 280 F.3d 715, 721 (6th Cir.2002). The Seventh and Ninth Circuits have suggested likewise. See United States v.

Santini, 656 F.3d 1075, 1078 (9th Cir.2011) (holding that prior convictions “for simple possession” were “not similar to the importation of marijuana and thus lack[] probative value”); Ono, 918 F.2d at 1465 (distinguishing between possession and distribution in dicta); United States v. Monzon, 869 F.2d 338, 344 (7th Cir.1989) (concluding that evidence of the defendant’s prior marijuana possession was not probative of his intent to distribute cocaine); United States v. Marques, 600 F.2d 742, 751 (9th Cir.1979) (distinguishing between “personal use versus resale”); cf. Enriquez v.

United States, 314 F.2d 703, 717 (9th Cir.1963) (concluding that a trial was unfair because the court had admitted evidence of marijuana possession to show intent to sell heroin). But see United States v. Wash, 231 F.3d 366, 370-71 (7th Cir.2000) (allowing the admission of a possession conviction in a distribution trial because the conviction involved “distribution amounts”). Other circuits have reached the opposite result, but we are not persuaded. See, e.g., United States v. Butler, 102 F.3d 1191, 1196 (11th Cir.1997); United States v. Logan, 121 F.3d 1172, 1178 (8th Cir.1997); United States v.

Gadison, 8 F.3d 186, 192 (5th Cir.1993). We conclude that Davis’s convictions should not have been before the jury — not as evidence of knowledge, not as evidence of intent. And problems remain. The District Court also committed two instruction-related errors. First, it did not provide the requested limiting instruction at the time the evidence was admitted; it did so only in the final jury charge.

Second, the Court concluded that the convictions were admissible to prove knowledge, but the jury charge included a wide list of purposes, allowing the jury to consider the convictions as evidence of “state of mind, knowledge, or intent,” as well as absence of “accident or mistake.” J.A. 125-26; see Sampson, 980 F.2d at 889 (“By simply repeating the entire litany of permissible theories under Rule 404(b), the judge’s instruction gave the jury inadequate guidance.”). While these errors are problematic, we would reverse even in their absence.

No instruction could have eliminated the infirmity at the heart of this case: Davis’s convictions were inadmissible for any purpose. IV Davis’s two remaining arguments are mere makeweight. The first is that the District Court improperly admitted a statement from the government’s expert witness. The second is that the Court improperly refused to admit a witness’s prior statement. We review these decisions for an abuse of discretion. See United States v. Mathis, 264 F.3d 321, 335 (3d Cir.2001) (applying the abuse-of-discretion standard to a decision about the admissibility of expert testimony); United States v.

Frazier, 469 F.3d 85, 87 (3d Cir.2006) (applying the same standard to a decision about the admissibility of a prior consistent statement). Davis argues that the government’s narcotics expert, Kenneth Beilis, violated Federal Rule of Evidence 704(b). This Rule bars experts from testifying that the defendant had the necessary state of mind to commit a crime — whether it be intent, knowledge, or something else. Yet the Rule does not bar experts from testifying about the practices of those in the drug trade. United States v. Watson, 260 F.3d 301, 308 (3d Cir.2001).

In fact, such testimony is admissible even if it supports a conclusion that the defendant had the necessary state of mind. United States v. Bennett, 161 F.3d 171, 183 (3d Cir.1998). The only limitation is that the expert may not draw the ultimate conclusion for the jury or testify in such a way that the ultimate conclusion is inevitable. Id. Davis objects to the following statement from Beilis: Government: One further question. If you had that level of cocaine, if you were a distributor, 740 grams — if you were upper level distributor with 740 grams of cocaine, is it common that you would have persons present?

Davis’s attorney: Objection. The Court: Objection overruled. Government: Is it common that you would have persons in the presence of that cocaine who did not have a connection to that cocaine? Davis’s attorney: Objection. The Court: Overruled. Beilis: Not in my opinion, no. J.A. 50-51. According to Davis, these questions elicited Beilis’s opinion on whether Davis intended to distribute the cocaine in the backseat. The major flaw in this argument is that Beilis merely spoke about common practices. He did not in any way connect those practices to Davis.

This means that Beilis did not draw the ultimate conclusion for the jury, nor did the conclusion inevitably follow from his testimony. United States v. Price, 458 F.3d 202, 212 (3d Cir.2006) (allowing an expert to say “that in his opinion ... drug dealers are very likely to carry guns, and drug buyers almost never do” because the expert “said not a word about [the defendant’s] mental state” but rather spoke about “common practices”). Davis’s Rule 704(

b) argument is meritless. The same is true of his final argument. Davis asserts that the District Court improperly refused to admit a prior statement from Festus. The statement called into question the police officers’ accounts, but the Court concluded that it was inadmissible hearsay. See Fed.R.Evid. 802. Federal Rule of Evidence 801(d)(1)(

B) defines a statement as nonhearsay if “[t]he declarant testifies and is subject to cross-examination about a prior statement, and the statement ... is consistent with the declarant’s testimony and is offered to rebut an express or implied charge that the declarant recently fabricated it or acted from a recent improper influence or motive in so testifying.” Festus’s various accounts are inconsistent at best. A few months after the arrests, he told Blackshear’s attorney in a written statement that the patrol car pulled up behind the Jeep with its lights flashing.

Two months later, Festus testified in the suppression hearing that the lights were not flashing. At trial, he returned to his former statement that the lights were flashing. The government impeached Festus’s trial testimony by introducing his inconsistent suppression testimony. On redirect, Davis referred Festus to his written statement to Black-shear’s attorney. See Fed.R.Evid. 612. Davis also asked the District Court to admit Festus’s written statement as a prior consistent statement. The Court denied that request.

As required by Rule 801(d)(1)(B), Festus was subject to cross-examination and his prior statement — at least the one he gave to Blackshear’s attorney — was consistent with his trial testimony. But Davis runs into problems with the requirement that the statement “rebut an express or implied charge that the declarant recently fabricated it or acted from a recent improper influence.” Fed.R.Evid. 801(d)(1)(B). The government never so much as suggested that Festus’s trial narrative was a recent fabrication. It merely pointed out the inconsistency between his suppression testimony and his trial testimony.

Inconsistency alone is not a charge of recent fabrication; we also require a suggestion of “conscious alteration.” Frazier, 469 F.3d at 89 (“The line between challenging credibility or memory and alleging conscious alteration can be drawn when a district court determines whether the cross-examiner’s questions reasonably imply intent on the part of the .witness to fabricate.”); see also Tome v. United States, 513 U.S. 150, 157, 115 S.Ct. 696, 130 L.Ed.2d 574 (1995). Absent that suggestion, Davis’s final argument must fail. The District Court correctly denied Davis’s suppression motion.

We cannot say the same about its decision to admit Davis’s possession convictions, which were inadmissible to prove knowledge or intent in a trial for possession with intent to distribute. We will vacate Davis’s conviction and remand to the District Court. . The officers did not explain how the “body motions” were inconsistent with lawful behavior, such as sharing a meal or exchanging gifts. . The District Court had jurisdiction under 18 U.S.C. § 3231, and we have final-decision jurisdiction under 28 U.S.C. § 1291. .

Davis has standing to challenge the search of the Jeep — even though he was a mere passenger — because his seizure gave the officers a reasonable suspicion to search inside the Jeep. See United States v. Mosley, 454 F.3d 249, 253 (3d Cir.2006). . Rule 404(

b) excludes only extrinsic evidence, or evidence of acts that are not the basis of the current prosecution. It does not exclude intrinsic evidence, which either "directly proves” or ‘‘facilitate^]” the charged offense. Green, 617 F.3d at 248-49 (quotation marks omitted). . We have affirmed even when a district court's analysis was somewhat flimsy — but only when the government had already estabIished a valid chain of inferences. See, e.g., United States v. Lopez, 340 F.3d 169, 173-74 (3d Cir.2003). . Some circuits require prior acts under Rule 404(

b) to "meet a threshold level of similarity in order to be admissible to prove intent” to commit the charged offense. United States v. Long, 328 F.3d 655, 661 (D.C.Cir.2003); see also United States v. Foskey, 636 F.2d 517, 524 (D.C.Cir.1980) (citing cases in Second, Fifth, and Ninth Circuits for the idea that "[w]hen a prior criminal act is relied upon to prove intent or knowledge, similarity between the two events must be shown” (alteration in original and quotation marks omitted)). We need not adopt that requirement in our Circuit or decide whether cocaine possession and distribution are sufficiently similar.

After all, a past intent to possess drugs simply is not probative of a future intent to distribute. . The government did not argue the issue of harmless error in its brief. Such silence usually means that harmless error is waived, but we may still consider the issue depending on "the length and complexity of the record, whether the harmlessness of the error or errors found is certain or debatable, and whether a reversal will result in protracted, costly, and ultimately futile proceedings in the district court.” United States v. McLaughlin, 126 F.3d 130, 135 (3d Cir.1997) (quoting United States v.

Giovannetti, 928 F.2d 225, 227 (7th Cir.1991)). Here, we refuse to do so because we see no reason "that a reversal will [ ] lead to drawn out proceedings,” and we do not "know with sufficient certainty that the error was harmless.” United States v. Faulks, 201 F.3d 208, 213 (3d Cir.2000). . Rule 704(

b) states, "[i]n a criminal case, an expert witness must not state an opinion about whether the defendant did or did not have a mental state or condition that constitutes an element of the crime charged or of a defense. Those matters are for the trier of fact alone.” . Whether the lights were flashing is not relevant to the crimes or the Fourth Amendment analysis. Yet this dispute had the potential to call into question the officers' testimonies.