Opinion for the Court filed by Senior Circuit Judge STEPHEN F. WILLIAMS.
STEPHEN F. WILLIAMS, Senior Circuit Judge:Johnny St. Valentine Brown for years testified as an expert witness for the government in narcotics cases. But it later developed that Brown was something of a con man himself, so much so that he was charged with and pleaded guilty to having committed perjury about his educational background. Among the trials at which he testified, was that of Robert Gale, who was convicted of possession of marijuana and possession of heroin with intent to distribute. Thus we again consider the effects of Brown’s testimony on the adequacy of a trial. Compare, e.g., United States v. Williams, 233 F.3d 592 (D.C.Cir.2000).
Gale challenges his convictions under the federal habeas corpus statute, 28 U.S.C. § 2255, arguing that the prosecution violated Brady v. Maryland, 373 U.S. *283, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), by (1) knowingly offering perjured testimony by Brown at the trial itself, thereby fatally tainting the trial, and (2) failing to disclose Brown’s past perjuries. We will assume in Gale’s favor that the government’s connections to Brown were such that it could be said to have the requisite knowledge to trigger the precedents Gale invokes. But Gale has offered no reason to think that Brown’s testimony at his trial was perjurious. And, again assuming that the government could be said to have had the kind of knowledge or notice of Brown’s past perjuries to create an obligation to disclose them to Gale, the non-disclosure was irrelevant because there is no “reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.” United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 3383, 87 L.Ed.2d 481 (1985) (opinion of Blackmun, J.).
The facts of Gale’s case have been set out at length in his direct appeal, United States v. Gale, 136 F.3d 192 (D.C.Cir.1998), so we will repeat them here only as relevant to his § 2255 action. He was arrested in an apartment where the police found drug-related paraphernalia, including bagging equipment, ziplock bags, white powder, cutting agents, razor blades, measuring equipment, electronic scales, plastic gloves, cellular phones, beepers, and nine Amtrak tickets in various names. Although the police recovered no drugs from Gale himself, his co-defendant, who was with Gale in the apartment at the time, possessed both cocaine base and 118 zi-plock bags of a mixture of cocaine and heroin. The apartment also contained cocaine powder on a paper plate, more cocaine in a box in the kitchen, and marijuana in the living room and kitchen.
Brown, who had no role in the arrest or other events leading to the prosecution, testified as an expert witness on narcotics. He said that he was “presently a narcotics consultant to the Metropolitan Police Department [ (“MPD”) ] in the District of Columbia.” As the prosecutor sought to establish Brown’s qualifications as a narcotics expert, defense counsel said, “I have no objection if he’s seeking to qualify Detective Brown. If he wants to go on, that’s fine, too.” The prosecutor then continued briefly, establishing that Brown had previously been an active member of the MPD for 26 years, during which, Brown said, he “worked homicide, checking fraud, robbery, prostitution, gambling, and of course my last assignment, which lasted for 22 years, was as a narcotics investigator with the Narcotics and Special Investigations Division [ (“NSID”) ].”
Brown testified that in his 22 years at NSID, he probably had worked on 1,500 narcotics cases and had become familiar with how heroin and cocaine are packaged, sold, and used in the District of Columbia. Brown did not testify about his educational background (the subject of his false testimony that ultimately led to his perjury convictions) nor about being qualified as an expert in other cases. Without objection from the defense, the court allowed Brown to testify as an expert.
Brown explained generally how cocaine and heroin are packaged and sold in the city, identifying how various items found at the apartment are used. As is relevant to this appeal, Brown described “what we call a ‘pev’ in the pharmacy world,” which he said was “used to. crush items that are in a rock-hard form. You pulverize it, you break it down into a crystalline or powder form.” He also explained that rubber gloves can be used in the drug preparation process to prevent transferring residue from hand to mouth. And he testified as *3to the street values of various substances and explained the chain-of-custody procedures used by MPD and the Drug Enforcement Administration to safeguard narcotics evidence.
In addition, Brown testified generally about the risks associated with the drug business, including “being ripped off, stuck up or robbed, or the possibility of the substances being seized by the police.” Because of these risks, he testified, a drug trafficker “would never allow anyone that’s not involved in the business to even be in any way associated, especially if the place where those substances are being prepared, what we call a 'bag-up house,’ is being prepared for distribution purposes.”
The jury found Gale guilty, and he was sentenced to 121 months of incarceration followed by three years of supervised release. We upheld the conviction and sentence. See United States v. Gale, 136 F.3d 192.
In his petition under § 2255 Gale argued that the government had violated the Brady rule in two ways. First, he said, it had introduced testimony that it knew or should have known was perjured, specifically various aspects of Brown’s self-identification as an expert. Second, he said the government violated Brady by failing to disclose that Brown had committed perjury in other cases and had lied in a prior application for re-employment at the MPD. Gale also sought discovery “to determine the nature and extent of Brown’s perjury.”
The district court rejected Gale’s petition and denied his request for discovery. We affirm for reasons similar but not identical to those given by the district court.
Gale first contends that his trial was tainted because Brown committed perjury at that trial. Gale cites three instances: while Brown used the phrase “we in the pharmacy world” (when referring to “what we in the pharmacy world call a ‘pev’"1), in fact he had no pharmacy degree; while Brown said that he had “worked homicide,” in fact he was never specifically assigned to the homicide division or as a homicide investigator; and while he identified himself as a “narcotics consultant” to the MPD, he was at the time not officially employed by the MPD.
Under United States v. Agurs, 427 U.S. 97, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976), “a conviction obtained by the knowing use of perjured testimony is fundamentally unfair, and must be set aside if there is any reasonable likelihood that the false testimony could have affected the judgment of the jury.” Id. at 103, 96 S.Ct. at 2397 (footnote omitted). We assume arguendo that the government’s use of the challenged testimony, to the extent it was false, could—because of Brown’s various professional links to the government—be viewed as “knowing use” within the meaning of Agurs. But as the district court found, Gale has failed to show that Brown’s statements were false.
While Brown did not possess a pharmacy degree, he didn’t claim to: rather, he merely implied an affiliation with “the pharmacy world,” something he undoubtedly had in light of his years of investigating and testifying about the process of making, packaging, and distributing drugs, albeit illegal ones. Nor can we say that *4his statement about having “worked homicide” was perjury. Brown testified that he “worked homicide, checking fraud, robbery, prostitution, gambling, and of course my last assignment, which lasted for 22 years, was as a narcotics investigator.” Gale has presented no argument -that Brown did not work on homicide cases, only that he was never expressly assigned to a homicide division. That is not enough to make Brown’s general statement perjury. Finally, Gale presented no reason to doubt that Brown was a “narcotics consultant” to the government, as he was an expert appearing in numerous cases on the government’s behalf. Nor is there any “reasonable probability” that any gap between Brown’s statements and perfect truth on these trivial and peripheral issues could have affected the judgment of the jury. Thus there can be no Agurs violation.
Second, Gale argues that the government committed a Brady violation by failing to advise him of Brown’s prior perjuries and the incomplete information on his job application. Again we assume in Gale’s favor that it is possible to attribute Brown’s knowledge of his past perjury to the prosecutors for Brady purposes. Compare United States v. Brooks, 966 F.2d 1500, 1502-05 (D.C.Cir.1992) (discussing scope of prosecutor’s obligation to search for potentially exculpatory materials). Gale’s claim fails, nonetheless, because the non-disclosure was immaterial.
Whereas the prosecution’s knowing use of false testimony entails a veritable hair trigger for setting aside the conviction (“any reasonable likelihood that the false testimony could have affected the judgment of the jury,” see Agurs, 427 U.S. at 103, 96 S.Ct. at 2397), non-disclosure of exculpatory evidence (including impeachment evidence) is governed by a more general standard: “[Fjavorable evidence is material, and constitutional error results from its suppression by the government, ‘if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different.’” Kyles v. Whitley, 514 U.S. 419, 433-34, 115 S.Ct. 1555, 1565-66, 131 L.Ed.2d 490 (1995) (quoting United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 3383, 87 L.Ed.2d 481 (1985)) (opinion of Blackmun, J.); Bagley, 473 U.S. at 685, 105 S.Ct. at 3385 (White, J., concurring in part and concurring in the judgment and in the formula above). The defendant bears the burden of showing a reasonable probability of a different outcome. See Strickler v. Greene, 527 U.S. 263, 291, 119 S.Ct. 1936, 1953, 144 L.Ed.2d 286 (1999). In applying this test, the court must consider the non-disclosure dynamically, taking into account the range of predictable impacts on trial strategy. See, e.g., United States v. Bowie, 198 F.3d 905, 909-12 (D.C.Cir.1999).
Gale’s arguments do not meet this burden. He has offered no reason to believe that, had the impeachment evidence in question “been disclosed to the defense,” the government would have foolishly charged ahead, blindly offering Brown and exposing itself to his inevitable demolition on cross. Why would it have done so, rather than simply offering another expert? Brown’s expertise was drawn not from his command of some arcane field but from an experience that is widely-shared in urban police forces: he had investigated narcotics cases for many years. Moreover, while the government has argued that it had several narcotics experts available from MPD (and identified one in particular, Sergeant Brennan, who is a 25-year veteran narcotics investigator), Gale has offered nothing suggesting that Brown could not have been replaced with a similarly qualified witness. This case is thus *5unlike the ones cited by Gale, Kyles v. Whitley and Strickler v. Greene, 527 U.S. 263, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999), in both of which the impeached witness was a fact witness who could not be readily replaced. By contrast, in our case of United States v. Williams, 233 F.3d 592 (D.C.Cir.2000), although applying a different substantive standard to a Rule 33 motion for a new trial, we noted the availability of other experts to offer exactly the same evidence as Brown. Id. at 595. For other cases involving readily replaceable witnesses, see United States v. Matthews, 168 F.3d 1234, 1242-43 (11th Cir.1999); United States v. Garcia, 13 F.3d 1464, 1472 (11th Cir.1994).
Gale raises two objections. First, he cites hearsay from news articles suggesting that Brown was a particularly charismatic expert witness, and argues that a replacement would have been less effective. We see no reason to credit such hearsay — other defendants are convicted both in the District of Columbia and across the country when experts other than Brown testify. Even if Brown had been the “best” expert witness, we have no reason (and no evidence) to believe that the “second best” witness would have been materially inferior. Furthermore, the “charisma” argument on which Gale relies is precisely the type of demeanor evidence that disappears at the time of trial. See, e.g., United States v. Zeigler, 994 F.2d 845, 849 (D.C.Cir.1993); Dyer v. MacDougall, 201 F.2d 265, 269 (2nd Cir.1952). Thus the evidence Gale offers — hearsay regarding evanescent evidence of charisma — does not create a reasonable probability that government use of a different expert with similar experience in narcotics cases would have changed the result.
Second, Gale argues that Brown’s testimony that persons not involved in the drug trade would not be in an apartment used for packaging drugs was particularly damaging to Gale’s defensive claim that he just happened to be present. Gale goes on to suggest — but not affirmatively argue — that only Brown would have been willing to offer such testimony:
His expert claim that only those “involved” in the drug business would be present in an apartment containing drugs — which was devastating to Mr. Gale’s defense in this case — seems particularly suspect. The government did not submit any declaration from any other expert expressing agreement with Brown’s testimony in Mr. Gale’s case.
To the extent that Gale is simply arguing that the testimony in question may have been overstated, we assume that to be true. Surely a non-participant might be present in an apartment containing drugs because, for example, the drugs and drug paraphernalia were hidden, or because the non-participant’s presence was momentary and accidental. But Gale never articulates such an argument, likely because any such overstatement (besides being easily torpedoed on cross) was wholly irrelevant to him — who at the time officers entered had been sleeping in the apartment, which was positively littered with drugs and drug paraphernalia in plain view. See United States v. Gale, 136 F.3d at 193-94. In fact, until pressed at oral argument, Gale did not even claim that Brown’s statement in this regard was false but merely that there are “questions about the truth of everything to which [Brown] testified.”
Further, any suggestion that, as a matter of substance, only Brown would give such testimony is plainly untrue. Indeed, in a case now on the court’s docket, a Detective Tyrone Thomas from the MPD provided the following testimony:
Everybody who is a part of that has a role in that drug operation, maybe one *6of those roles that I mentioned earlier, and that’s to oversee those operations. I mean, it’s — nor would anyone want to go around and just be hanging around a scenario where some large quantities of narcotics are gonna be sold because they’re not gonna want to risk being caught up in a situation like that unless they have some method or role involved in the drug operation.
Just like somebody going to rob a bank, they’re not gonna take a Mend along just for the ride; nor is that friend gonna want to be going to where somebody is gonna rob a bank.
United States v. Bailey, No. 99-164-4, Tr. Vol. VI, 1/26/01, p. 5. (We of course express no opinion as to the permissibility of testimony so formulated.) In light of other experts who have offered substantively similar testimony and Gale’s failure to even allege before argument that the cited testimony was false, we cannot say that Gale has shown a “reasonable probability” of a different result had the government disclosed the potential impeachment evidence against Brown.
Finally, we reject Gale’s argument that the trial court erred in denying his request for further discovery about Brown’s alleged perjury in this and other trials, and materials “reflecting knowledge” of Brown’s perjuries within the government. We review the district court’s denial of this request for abuse of discretion. Bracy v. Gramley, 520 U.S. 899, 909, 117 S.Ct. 1793, 1799, 138 L.Ed.2d 97 (1997). We find no such abuse, as there was no real chance that discovery could have turned up information altering the outcome. As we have seen, there is no serious claim of perjury in Gale’s trial. And with respect to knowing use of a perjurer, additional discovery would be of no use because we have assumed in Gale’s favor government knowledge of Brown’s perjurious inclinations but found no Brady violation: were that assumption true and the information disclosed to the defense, the government would have replaced Brown with a different expert.
The judgment of the district court is
Affirmed.
. Gale does not contend that the use of the term "pev” constituted perjury. At trial, Brown stated that a "pev” is a pharmaceutical tool "used to crush items that are in rock-hard form.” While not characterizing this statement as peijury, Gale notes in his reply brief that he has been unable to find a definition of the term. Reply Br. at 9. Our searches have been equally unavailing.