STEPHEN F. WILLIAMS, Senior Circuit Judge:

A jury convicted Gary Bailey of conspiracy to distribute 500 grams or more of cocaine, in violation of 21 U.S.C. §§ 846, 841(a)(1), 841(b)(1)(B)(ii) (2000). The court sentenced Bailey to 130 months' imprisonment and to eight years of supervised release. In this appeal, Bailey challenges three evidentiary rulings by the district court. We affirm the district court, finding that while two of Bailey's challenges have merit, the errors were harmless.

According to the government's evidence, Bailey supplied five kilograms of powder cocaine to Daniel Clayton on April 15, 1999 for use in a drug deal near the Watergate Hotel. Bailey owned the cocaine, and Clayton (who pled guilty and testified for the government) served as a "broker" for its sale. Clayton drove the cocaine from New York to Washington in a Honda with a secret compartment (a "trap") installed by a third defendant, Darryl Simmons, who flew down to D.C. separately. Bailey followed Clayton in an Acura. Bailey and Clayton later picked up Simmons from the airport in D.C., and the two cars went on to the Watergate area. Though the deal was originally for five kilos of crack cocaine (apparently Simmons was going to "cook" the powder cocaine in D.C. to make it crack once it was clear that the sale would go through), the quantity was for some reason later reduced, and the police were given four kilos of powder. Clayton and another defendant gave the cocaine to the "buyer," Kevin Goode, a Drug Enforcement Administration "cooperator." After inspecting the contents, Goode gave the signal, and officers arrested Clayton, Bailey and another defendant. Simmons escaped initially but was eventually arrested in New York.

Several items of evidence linked Bailey to the drug transaction. When Bailey was arrested, he was in the Acura, which several officers had seen driving in tandem with the Honda — arriving at the Watergate area, proceeding to the airport to pick up Simmons (who knew how to get the cocaine out of the trap), and returning to the Watergate area. Throughout these peregrinations the Acura that Bailey drove stuck with the Honda that contained the cocaine; observing officers saw the vehicles making U-turns together, indeed virtually "bumper locked." In the Acura was a piece of paper with Clayton's cell phone number, and Bailey's cell phone records showed five calls to Clayton's cell phone on April 15, and two on the day before. Clayton testified extensively about Bailey's involvement. Simmons testified that he let Clayton use his Honda with the secret compartment to carry the drugs, and that while Clayton was in the Honda in New York, Simmons saw Clayton transfer the drugs from a car that Bailey was driving and put them into the Honda.

The government also introduced evidence of other drug deals of Bailey's. First, Clayton testified that before the April 15 deal, he had made between seven and twelve powder cocaine deals with Bailey over a nine-month period, ranging from three ounces to three quarters of a kilogram. There was also testimony of two specific drug episodes in New York. First, an officer testified about finding two small bags of cocaine on Bailey, totaling about 50 grams, on February 2, 2000. Second, several witnesses testified to a cocaine transaction in the Bronx on June 12, 1998, in which 225 grams of cocaine were found in Bailey's car. Both Clayton's vague reference (the seven to twelve deals) and the two New York transactions were admitted under Rule 404(b) of the Federal Rules of Evidence to show knowledge, motive, opportunity, intent, and plan. There is no dispute about those admissions or the court's instruction allowing that use. But there was also an instruction — hotly contested — allowing the jury to consider the two New York transactions to corroborate Clayton's testimony about his prior drug transactions with Bailey.

Finally, the government produced an expert witness, Detective Tyrone Thomas, who testified as to the modus operandi of drug dealers. The testimony tended to explain the maneuvers of Bailey and his colleagues in terms of the purposes and problems facing participants in major drug deals.

The defense was skimpy. Bailey did not testify at all. The core of his defense appeared to rest on evidence that a number of ladies' clothes, including undergarments, were found in the Acura. He was apparently trying to show that he had come to the District to have a tryst with a woman — though he offered no evidence that such a project was mutually exclusive with a drug deal. In cross-examination he severely attacked the credibility of Clayton.

Bailey raises three challenges. First, he argues that it was error to prevent him from eliciting testimony that he had not yet been tried on his two New York offenses. Second, he argues that the jury should not have been instructed that it could use the evidence of his two prior New York offenses to corroborate Clayton's testimony. Finally, he argues that Detective Thomas's testimony violated Rule 704(b) by suggesting that the witness had some special insight into Bailey's mental processes.

* * * Admissibility of the Legal Status of Bailey's Past Acts

Bailey first argues that he should have been allowed to elicit testimony that he had not yet been tried for the two New York offenses that were introduced under Federal Rule of Evidence 404(b). We review this claim for abuse of discretion. See United States v. Clarke, 24 F.3d 257, 267 (D.C. Cir. 1994).

Since the government has given us no affirmative reason for excluding the evidence of the legal status of defendant's other acts, the question is whether that evidence is relevant under Federal Rule of Evidence 401. It defines relevant evidence as "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." Fed.R.Evid. 401. At first glance, the relevance of the fact that Bailey had not yet been tried for the other crimes appears feeble to non-existent. While an affirmative government decision not to prosecute may suggest weakness in the government's evidence, the simple fact that the prosecution will likely take place in the future supports no such inference.

But Bailey argues that the jury would speculate about whether or not he had been convicted of the 404(b) crimes, and would likely infer conviction from silence. The government's evidence of the crimes took the form of the arresting officers' testimony about the narcotics transactions and the ensuing arrests. It seems plausible that not a few jurors would have speculated that conviction followed. We can find no case directly holding that evidence is relevant solely to refute a likely mistaken jury inference (though there are a few where that appears to be the dominant purpose, see, e.g., United States v. Powell, 226 F.3d 1181, 1199 (10th Cir. 2000); United States v. Johnson, 802 F.2d 1459, 1463-64 (D.C. Cir. 1986)); one case, United States v. Jones, 808 F.2d 561 (7th Cir. 1986), mentions the possible use of an acquittal for the specific purpose of correcting jury speculation that the defendant had been convicted, but reaches no conclusion on relevance, id. at 566-67. Our own closest case, United States v. Thomas, 114 F.3d 228 (D.C. Cir. 1997), also considers the admissibility of an acquittal, but not its possible relevance to rebut an inference of conviction, and in any event ultimately goes off on narrow factual grounds. Id. at 249-50. We think that if the jury inference is plausible, evidence to rebut that inference is relevant. A similar theory may underlie the Supreme Court's positive emphasis, in Dowling v. United States, 493 U.S. 342, 110 S.Ct. 668, 107 L.Ed.2d 708 (1990), on the trial court's having twice instructed the jury that the defendant had been acquitted on the 404(b) crime, id. at 345-46, 353, 110 S.Ct. at 670-71, 674-75.

Jones, mentioned above as entertaining the possible relevance of acquittal for refuting mistaken jury speculation, in the end upheld exclusion on the grounds of possible prejudice, 808 F.2d at 566-67, presumably the risk that the jury might over-read acquittal to signify innocence rather than merely failure of the government to show guilt beyond a reasonable doubt. Plainly the risk of prejudice (which under Rule 403 would support exclusion) is far less in the case of mere evidence that the defendant hasn't yet been tried; indeed, as we said earlier, the government never adduced any affirmative reason to exclude the evidence.

Evidence that no trial has occurred need not raise the sort of hearsay issues posed by convictions or acquittals. Where either of those is offered to show commission or noncommission of the acts in question, they are hearsay: in effect they simply quote the assertion of 12 jurors (who did not themselves perceive the acts charged) that the person did or didn't do the acts. But convictions come into evidence all the time — thanks to the Rules' explicit provision of a hearsay exception, Federal Rule of Evidence 803(22). No comparable exception exists for acquittals, though other exceptions, such as the one for public records, see Rule 803(8), may allow some uses of acquittals. But evidence that a defendant has not been tried on a charge, where offered simply as proof that he has not been tried and convicted, is not hearsay. Obviously the defendant knows that directly; non-hearsay reports as to the absence of relevant court records provide an alternative means. In most cases, we should expect, the parties could resolve it by stipulation.

Having found that the evidence was relevant and that there was no affirmative reason for exclusion, we believe the exclusion was error — though we hesitate to find abuse of discretion, given the narrow purpose the evidence would have served. In any event any error was harmless.

For nonconstitutional error, an appellate court must determine with "fair assurance . . . that the judgment was not substantially swayed by the error." Kotteakos v. United States, 328 U.S. 750, 765, 66 S.Ct. 1239, 1248, 90 L.Ed. 1557 (1946). The judge here expressed his concern that there might be an "improper [jury] inference that Mr. Bailey was convicted of these charges which are now still pending," 2 Trial Tr. at 250, and sought to cure the problem by instructing the jury, "You may not consider or speculate about the status of any charges against the defendant in that case." 5 Trial Tr. at 234. We think this considerably reduced the risk that the jury might leap to the conclusion that the defendant had been convicted.

Beyond the judge's instruction, the government's evidence was quite strong, both on the whole and as to these past incidents. Several law enforcement officers testified to the June 12, 1998 offense. They had been surveilling one Robert Hunt, who had agreed to sell cocaine to an undercover police officer. They saw Bailey conferring with Hunt immediately before Hunt made the sale. As the police were arresting the participants and searching the apartment where the sale occurred, Bailey looked in the direction of the police, threw his keys on the ground, and walked away. Inside the car that matched the keys, the police found 225 grams of cocaine, and inside Bailey's wallet they found a registration for the car in Bailey's mother's name. For the February 2, 2000 incident, the police found two bags of cocaine in Bailey's pocket. It is highly likely that the jury would have found these incidents to be true even if it had been told that Bailey hadn't yet been tried. Further, the government's case as a whole was strong, consisting of both the direct testimony of Clayton and Simmons, the strong circumstantial evidence, and the expert testimony of Detective Thomas. For these reasons, we have a fair assurance that the judgment was not substantially swayed by the error.

Use of Rule 404(b) Evidence for Corroboration

The two New York incidents are the source of another argument — that the court erred in instructing the jury that it could use the incidents to corroborate Clayton's testimony that he had engaged in prior cocaine deals with Bailey, and thus his story about the crime actually charged. The parties agree that the two incidents were properly admitted under Rule 404(b) to show knowledge, motive, opportunity, intent and plan, and dispute only the corroboration instruction. We review for abuse of discretion. Cf. United States v. Bowie, 232 F.3d 923, 926-27 (D.C. Cir. 2000) (stating standard of review for admission of 404(b) evidence).

Bailey argues that before evidence is used under Rule 404(b) for corroboration it should have to pass two extra screens imposed by two other circuits — that the corroboration be direct and the corroborated matter be significant. See United States v. Pitts, 6 F.3d 1366, 1370-71 (9th Cir. 1993); United States v. Everett, 825 F.2d 658, 660 (2d Cir. 1987). But we rejected the approach of these circuits in Bowie, 232 F.3d at 933 n. 7, reasoning that any special concerns were properly addressed in balancing probative value against prejudice, etc. under Rule 403. Bailey attempts to distinguish Bowie on the ground that there the corroboration was of the act actually charged in the indictment, while in this case, the corroboration is only of other acts admitted under 404(b) (Clayton's prior seven to twelve cocaine deals with Bailey). But Bowie itself makes no such distinction, and no language in Rule 404(b) supports it.

Nonetheless, use of 404(b) evidence for corroboration does have inherent complications. Corroboration, in and of itself, is not a separate purpose belonging in the open class of permissible purposes referred to in Rule 404(b)'s second sentence. If it were, evidence could slide past the rule against improper character evidence. To decide if Rule 404(b) evidence is admissible for corroboration, the court must determine what is being corroborated and how. If similar past acts were corroborative only because they showed the defendant's character and the likelihood of "action in conformity therewith," plainly the rule would call for exclusion. On the other hand, evidence might corroborate a witness's testimony by showing plan, purpose, intent, etc. and therefore be admissible under 404(b). The label "corroboration" thus merely invites a closer look at exactly how the evidence may be probative.

Clayton, it will be recalled, was not himself a part of Bailey's two New York drug deals. The only legitimate way for that evidence to corroborate Clayton's testimony was to help show Bailey's knowledge, intent, etc. as to the seven to twelve Rule 404(b) drug deals that Clayton mentioned. In other words, they were 404(b) evidence for Clayton's 404(b) evidence.

As to Bailey's intent or purpose in the seven to twelve incidents, it is hard to see how the New York deals could add. Clayton had summarily described the deals, making clear that they involved cocaine and that Bailey was selling to him, see 4 Trial Tr. 7-11, but Clayton had never in any way put Bailey's state of mind in issue. Bailey could have been a zombie throughout the seven to twelve deals, and the fact would not have contradicted Clayton's testimony in the slightest. But the incidents do help to show that Bailey had opportunities to acquire cocaine, and at least the 1998 incident (involving some semblance of a drug deal rather than mere possession) seems relevant to show knowledge of how to deal in cocaine.

In the weighing required by Rule 403, however, the evidence had little probative value for permissible corroborative purposes. Given the vague nature of Clayton's testimony about the seven to twelve past deals with Bailey, jury interest in the sort of opportunity and knowledge described was likely minimal. On the other side, by contrast, there were risks of unfair prejudice, confusion and misleading the jury. The permissible link is subtle, the impermissible one obvious and likely to be salient. The likelihood of confusion and prejudice seems overwhelming, at least in the absence of a clear and emphatic instruction, the framing of which strikes us as a daunting task. In any event, no such instruction was even attempted here, rendering the corroboration instruction error.

But this error was harmless. The incremental effect of the corroboration instruction was likely very small. It seems improbable that it did much to strengthen the jury's belief in Clayton, especially as the evidence of the New York incidents had already been properly admitted. And the case against Bailey was generally strong. We do not think the jury's judgment was substantially swayed by the error, even in combination with the error discussed above.

Testimony of Detective Thomas

Bailey's last objection is to the allowance of testimony that he says violated Rule 704(b)'s ban on an expert witness's giving an opinion as to "whether the defendant did or did not have the mental state or condition constituting an element of the crime charged or of a defense thereto." Fed.R.Evid. 704(b). Here he points to the evidence of Detective Thomas. The closest that testimony came to suggesting that the expert knew what was in the defendant's head was his statement that "no one would want to tag along [on a large-scale drug transaction] who is an innocent person. . . ." 6 Trial Tr. at 16 (emphasis added).

There having been no objection to this testimony, we review for plain error. United States v. Olano, 507 U.S. 725, 732-35, 113 S.Ct. 1770, 1776-78, 123 L.Ed.2d 508 (1993) (requiring error that is plain and that affected substantial rights). Because we find no error at all, the extra elements for plain error need not detain us.

In United States v. Smart, 98 F.3d 1379 (D.C. Cir. 1996), we said that testimony should not be excluded under Rule 704(b) as long as it is clear that the expert is testifying on the basis of his knowledge of general criminal practices and not on some special knowledge of the defendant's mental processes. Id. at 1388 (also involving testimony by Detective Thomas). We said that this inquiry required consideration of

(1) the language used by the questioner and/or the expert, including use of the actual word "intent"; and (2) whether the context of the testimony makes clear to the jury that the opinion is based on knowledge of general criminal practices, rather than "some special knowledge of the defendant's mental processes."

Id.

Here the phrase quoted earlier — that no one would tag along who was "an innocent person" — is certainly a close equivalent of what Rule 704(b) excludes, namely stating "the mental state or condition constituting an element of the crime charged or of a defense thereto." Clearly prosecutors should take care not to elicit, and experts not to make, statements that fall so close to the line. But Smart directs us to consider the context of the testimony, and here the context would have told the jury that Thomas was testifying solely from "general criminal practices."

Thomas's testimony centered on how drugs were distributed and the roles that workers in the drug world played in their distribution. He explained that drug dealers at the kilo level took great precautions against getting caught, using intermediaries, for example, such as brokers and people called "mules" who help transport the drugs. He also testified as to the cities into which drugs are trafficked generally, the way in which drugs are usually packaged, the difference between crack and powder cocaine and how transformation into crack increases the number of highs per unit of cocaine, the amount of cocaine a person buys at one time for personal consumption, how much money will be gained by the sale of a certain quantity of cocaine on the street level, the size and nature of traps in vehicles used to store drugs, money, or guns, and the use of informants as opposed to undercover officers. Thomas's testimony was far broader than the specific circumstances of Bailey's activities. Additionally, many of the specifics to which he testified had little to do with Bailey's mental state. His testimony was clearly intended to give the jury information about the way high-level drug sales worked.

Further, we said in Smart that if it is made clear to the jury that the expert was not qualified to testify to the ultimate issue of intent, there is no violation of Rule 704(b). Smart, 98 F.3d at 1389. That was made clear here. Detective Thomas frequently specified the source of his testimony. For instance, in discussion about whether drug dealers dealing with kilo-size weights normally arm themselves, Thomas said that he had been involved in a couple of hundred investigations where kilograms of cocaine were being sold, and that in only 15% of them were any sort of weapons involved. In discussing how many doses of crack five kilos of powder cocaine would yield, Thomas prefaced his answer by saying that it was what he knew from personal experience in his past work for the police department and drug enforcement.

And at a number of points Thomas explicitly noted his limited connection to the facts of the Bailey case. In answering the defense counsel's question on cross examination of whether he was "talking about the facts in this case," Thomas replied by saying, "No, I don't know the facts, all the facts in this case. I wasn't a part of the investigation." 6 Trial Tr. at 18. He said he had never met Bailey "that I can recall, and I know nothing about him, that I know of. No." Id. at 33. And in virtually the final impression that Thomas left with the jury, on redirect, he and government counsel had the following exchange:

Q: Were you involved in the investigation of Mr. Bailey?

A: No, I wasn't.

Q: Did you do surveillance?

A: No, I didn't.

Q: Did you have any involvement at all in that investigation?

A: None, whatsoever.

Q: Is your testimony yesterday and today about your general experience?

A: That's what I'm testifying, my general experience that I've accumulated over twenty-nine years of investigating narcotics[,] as what I've learned as how narcotic deals go.

Id. at 50. These statements clear up any suggestion that Thomas might have purported to rest the "innocent" comment on any special knowledge of the defendant's mental processes. Admission of his testimony was not error, and therefore we need not reach the other elements of plain error.

The judgment of conviction and sentence are

Affirmed.