Opinion for the Court filed by Circuit Judge WILLIAMS.
Appellants Carlos Saro and Cornelio Cabrera-Baez raise a variety of challenges to their convictions for various drug offenses. Except for the attacks on their sentences, all are insubstantial. Because of apparent error in the calculation of Cabrera-Baez’s “base offense level”, we remand his case for further consideration. We affirm Saro’s conviction in all respects.
In their joint trial, Saro and Cabrera-Baez faced three charges common to both; Saro also was charged with four additional counts by himself. The jury convicted both defendants on two of the joint counts: distributing cocaine base, or “crack”, on or about September 17, 1990, see 21 U.S.C. § 841, and conspiring among themselves and with others to distribute crack between July 1989 (or thereabouts) and September 1990, see id. § 846. It acquitted both on the third joint count, a charge of attempting to distribute crack on or about July 20, 1989. The jury convicted Saro of three other distribution offenses with which he alone was charged; it acquitted him on one count of possession with intent to distribute in connection with 80-odd grams of crack found in an apartment linked to him.
Under the federal Sentencing Guidelines, the “base offense level” for drug crimes varies with the type and amount of drugs in question. U.S.S.G. § 2Dl.l(c) (November 1, 1990). Purporting to follow the guidelines’ standards, the appellants’ pre-sentence reports — whose findings and calculations the district court adopted — held each appellant responsible for 7266.08 grams of crack.1 This sum earned both appellants a base offense level of 40. See id. § 2D1.1(c)(2) (covering amounts between 5 and 15 kilograms of crack). The reports then proposed for each appellant a 3-level increase pursuant to § 3Bl.l(b) of the guidelines, which applies “[i]f the defendant [1] was a manager or supervisor (but not an organizer or leader) and [2] the criminal activity involved five or more participants or was otherwise extensive”.2 Saro, who had taken the stand in his own defense, also got a 2-level increase for obstructing justice by lying under oath. See id. § 3C1.1. Since the guideline “range” for offense levels 43 and above is life imprisonment, id. § 5A, the district court sentenced both appellants to life.
The only substantial question is whether the pre-sentence reports adopted by the district court correctly calculated the appellants’ base offense level. For Saro, in fact, even this question turns out not to matter. As long as his base level is at least 38, his total offense level will be at least 43, and the guidelines will call for him to receive *280a life sentence. Base level 38, in turn, applies as long as at least 1.5 kilograms of crack are attributed to him. See id. § 2Dl.l(c)(3) (covering amounts between 1.5 and 5 kilograms). As about four kilograms that the district court attributed to him are beyond question,3 any error in its treatment of the remaining amounts was harmless. See Fed.Rule Crim.Proe. 52(a).
Cabrera-Baez, by contrast, would benefit if he could reduce his base offense level to 38. Since he did not receive Saro’s 2-level increase for obstruction of justice, such a reduction would lower his total offense level to 41, corresponding to a guideline range of 324 to 405 months (rather than life). U.S.S.G. § 5A. But unlike Saro, Cabrera-Baez did not raise a timely objection to the findings or analysis of his pre-sentence report. See Receipt and Acknowledgement of Presentence Investigation Report (filed July 18, 1991) (attesting that pre-sentence report and worksheet computations contain “no material factual inaccuracies”); Sentencing Transcript (July 18, 1991) at 2-3. As Cabrera-Baez concedes, the district court’s decision to adopt the findings and analysis of that report, see id. at 13, is therefore reviewable only for plain error. See, e.g., United States v. Foster, 988 F.2d 206, 209 (D.C.Cir.1993); Fed.Rule Crim.Proc. 52(b).
One could conceivably argue that 18 U.S.C. § 3742(f)(1), which declares that the courts of appeals “shall” remand eases for resentencing if the sentence “was imposed in violation of law or imposed as a result of an incorrect application of the sentencing guidelines”, mandates reversal even in the absence of objection whenever an appellate court finds any potentially harmful error, whether or not “plain”. Though the Supreme Court did not address the question in Williams v. United States, — U.S. -, 112 S.Ct. 1112, 117 L.Ed.2d 341 (1992), its discussion of § 3742(f)(1) did not rule out the possibility. But the plain-error doctrine was well entrenched as a background legal principle when Congress acted, and we think it fanciful to suppose that Congress intended § 3742(f)(1) to override that doctrine. Cf., e.g., Fogerty v. Fantasy, Inc., — U.S. -, -, 114 S.Ct. 1023, 1033, 127 L.Ed.2d 455 (1994); United States v. Giovannetti, 928 F.2d 225, 226 (7th Cir.1991). In keeping with this understanding, our sister circuits also apply the plain-error doctrine to sentencing.
As its name suggests, “plain error” exists only when the error is “obvious”. United States v. Olano, — U.S. -, -, 113 S.Ct. 1770, 1777, 123 L.Ed.2d 508 (1993). Obviousness is assessed from the perspective of the trial court; the error must be “so ‘plain’ the trial judge and prosecutor were derelict in countenancing it, even absent the defendant’s timely assistance in detecting it”. United States v. Frady, 456 U.S. 152, 163, 102 S.Ct. 1584, 1592, 71 L.Ed.2d 816 (1982); accord United States v. Davis, 974 F.2d 182, 190 (D.C.Cir.1992); see also United States v. Merlos, 8 F.3d 48, 51 (D.C.Cir.1993) (denying rehearing) (saying that obviousness is assessed “under current law at the time of trial”). But cf. Olano, — U.S. at-, 113 S.Ct. at 1777 (reserving judgment on “the special case where the error was unclear at the time of trial but becomes clear on appeal because the applicable law has been clarified”).
In addition to being obvious, the error generally must also have been “prejudicial”. The relevant question for gauging “prejudice” is whether the error “affected *281the outcome of the District Court proceedings”, and, in contrast with the rule that applies when a timely objection was made, it is the defendant who bears the burden of persuasion. Id. at-, 113 S.Ct. at 1778. But the Supreme Court has not specified exactly what a defendant must show in order to carry this burden.
At least in the context of trial errors, some circuits have indicated that defendants must prove it more likely than not that they would have been acquitted but for the errors. See, e.g., United States v. Kessi, 868 F.2d 1097, 1103 (9th Cir.1989); United States v. Silverstein, 732 F.2d 1338, 1349 (7th Cir.1984). If this standard is taken literally, it clearly is too high. The Supreme Court has declared that the showing of “prejudice” necessary under the “cause and prejudice” standard of habeas law “is significantly greater” than that necessary under the plain-error doctrine. Murray v. Carrier, 477 U.S. 478, 494, 106 S.Ct. 2639, 2648, 91 L.Ed.2d 397 (1986). Yet we know of no area of habeas law in which a petitioner would fail to establish “prejudice” if he showed, by a preponderance of the evidence, that the outcome of his trial would have been different but for the errors in question. Cf., e.g., Frady, 456 U.S. at 170, 102 S.Ct. at 1595.
At the other extreme are the standards of harmless error, controlling on direct appeal when a defendant did raise a timely objection to the asserted error. For most constitutional errors, an appellate court is to reverse if it entertains a “reasonable doubt” about whether the error affected the outcome below, Chapman v. California, 386 U.S. 18, 24, 87 S.Ct. 824, 828, 17 L.Ed.2d 705 (1967), while nonconstitutional errors are reviewed under some laxer standard not authoritatively captured in a formula (assuming a formula would genuinely clarify such a probabilistic inquiry), see United States v. Lane, 474 U.S. 438, 446 n. 9, 106 S.Ct. 725, 730 n. 9, 88 L.Ed.2d 814 (1986); cf. Kotteakos v. United States, 328 U.S. 750, 764-65, 66 S.Ct. 1239, 1247-48, 90 L.Ed. 1557 (1946). In any event, unless we are to eradicate the need for a timely objection, in plain violation of Rule 52’s distinctions, we must be materially more reluctant to find “prejudice” in the absence of a timely objection than when the error was properly raised. United States v. Harrison, 931 F.2d 65, 70 (D.C.Cir.1991).
Claims of relief for ineffective assistance of counsel provide another point of comparison. There the defendant, to establish prejudice, “must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Strickland v. Washington, 466 U.S. 668, 694, 104 S.Ct. 2052, 2068, 80 L.Ed.2d 674 (1984). This test does not require defendants to show that the errors “more likely than not altered the outcome in the case”, id. at 693, 104 S.Ct. at 2067, but only that they were “sufficient to undermine confidence in the outcome”, id. at 694, 104 S.Ct. at 2068, a standard the Court clearly intended as easier for defendants to satisfy.
Since reversal for “plain error” is designed largely to protect defendants from the defaults of counsel, there is a natural analogy between the assertion of “plain error” and the assertion of ineffective assistance of counsel. But see United States v. Caputo, 978 F.2d 972, 975 (7th Cir.1992) (suggesting that ineffective-assistance claims require stronger showing of prejudice). It should come as no surprise, then, that the Strickland formulation of “prejudice” comes quite close to what we have required in plain-error cases. For these we have applied a standard somewhere between the very demanding one for habeas cases and the relatively low one of “harmless error”, and have required defendants to show that a correct ruling “might likely have substantially affected the outcome of the trial”. United States v. Blackwell, 694 F.2d 1325, 1342 (D.C.Cir.1982).
In the special context of sentencing errors, moreover, we think that the required showing of “prejudice” should be slightly less exacting than it is in the context of trial errors. The rigor of the plain-error standard is designed to protect the finality of judgments, and with good reason; a broad readiness to excuse contemporaneous objection would invite counsel to withhold claims, looking forward either to an acquittal — which if it occurs will be invulnerable — or, in the event of a conviction, to a replay if the newly *282unveiled claim convinces the appellate court. Moreover, apart from some unfairness to the state in allowing the defendant a second bite at the apple, each bite is costly. See generally Wainwright v. Sykes, 433 U.S. 72, 88-90, 97 S.Ct. 2497, 2507-08, 53 L.Ed.2d 594 (1977) (describing benefits of contemporaneous-objection rule). When an error in sentencing is at issue, however, the problem of finality is lessened, for a resentencing is nowhere near as costly or as chancy an event as a trial. But cf. United States v. Pollard, 959 F.2d 1011, 1029 (D.C.Cir.1992).
The interest in finality is, of course, significant for sentencing as well as trials. Resentencing forces the trial judge to relearn the details of the case, and the defendant’s trip to court interrupts his prison regime and burdens authorities. Nonetheless, we think that for purposes of plain-error review, the burden of persuasion in showing “prejudice” should be somewhat lighter in the sentencing context. While it still seems accurate to .say that the defendant must show a reasonable likelihood that the sentencing court’s obvious errors affected his sentence (the “reasonable likelihood” formulation, after all, is not terribly precise), we are somewhat more willing to find “prejudice”.
With these principles in mind, we turn to Cabrera-Baez’s attacks on the analysis that led the pre-sentence report (and hence the district court) to attribute 7266.08 grams of crack to him. We conclude that the presentence report either used the wrong legal standard or failed to make adequate factual findings in attributing 2298.70 of those grams to Cabrera-Baez.
Cabrera-Baez argues, first, that the report misanalyzed his “relevant conduct”. At the time of his sentencing, the guidelines specified that “relevant conduct” could include three categories of acts (if adequately connected to the crime of conviction, which is not in dispute here): (1) those that the defendant himself committed, (2) those that he “aided and abetted”, and (3) those “for which the defendant would be otherwise accountable”. U.S.S.G. § lB1.3(a)(l). The third category incorporates, among other things, the well-settled principle of conspiracy law that someone who jointly undertakes a criminal activity with others is accountable for their reasonably foreseeable conduct in furtherance of the joint undertaking. See, e.g., Pinkerton v. United States, 328 U.S. 640, 66 S.Ct. 1180, 90 L.Ed. 1489 (1946). But Cabrera-Baez suggests that in analyzing this third category, the pre-sentence report neglected the requirement that the acts be “in furtherance of’ the joint undertaking, holding him responsible for all drug quantities attributable to his co-conspirators as long as he could reasonably foresee the transactions.
Such an approach would clearly be error, even under the version of the guidelines that existed at the time of Cabrera-Baez’s sentencing. (The current version makes the “in furtherance of’ requirement more explicit.) The extent of a defendant’s vicarious liability under conspiracy law is always determined by the scope of his agreement with his co-conspirators. Mere foreseeability is not enough: someone who belongs to a drug conspiracy may well be able to foresee that his co-venturers, in addition to acting in furtherance of his agreement with them, will be conducting drug transactions of their own on the side, but he is not automatically accountable for all of those side deals. See, e.g., United States v. Jenkins, 4 F.3d 1338, 1346-47 (6th Cir.1993); United States v. Evbuomwan, 992 F.2d 70, 73-74 (5th Cir.1993). The guilty verdict against Cabrera-Baez on the conspiracy count hardly ends the analysis; “a general verdict does not establish with whom a defendant conspired or the quantity of drugs encompassed by the conspiracy.” United States v. Edwards, 945 F.2d 1387, 1391 (7th Cir.1991). To calculate the amount of drugs attributable to defendants under the third category of “relevant conduct”, the sentencing court must “determine the scope of the conspiratorial agreement each joined”. United States v. Thompson, 944 F.2d 1331, 1344 (7th Cir. 1991); accord United States v. Goines, 988 F.2d 750, 775 (7th Cir.1993).
In Cabrera-Baez’s case, neither the sentencing court nor the pre-sentence report explicitly undertook such an analysis. Without more, this lack of explicitness would not constitute plain error, but there are substantial signs that Cabrera-Baez’s pre-sentence *283report was actively employing the wrong legal standard. Having concluded that Cabrera-Baez was involved in a large but loose-knit crack-distribution ring, the report’s author seemed to think that Cabrera-Baez was automatically responsible for the “[sjpeeific drug sales and drug related activity conducted by members of the conspiracy”, see PreSentence Report of Comelio Cabrera-Baez (“PSR”) V 33, at least if he knew about and could foresee that activity, id. ¶46. Thus, Cabrera-Baez was held accountable for 1.85 grams of crack found in an apartment at 3300 16th Street, merely because “[t]his apartment was used by various members of the conspiracy.” Id. ¶ 43. Similarly, despite characterizing Saro and Cabrera-Baez as “independent dealers” who “would on occasions act in cooperation with one another” but who also “supervised their own businesses”, id. ¶¶ 32, 44, the report held Saro and Cabrera-Baez responsible for exactly the same amounts of drugs — including the 296.85 grams of crack involved in the one possession and the three distribution offenses with which Saro alone was charged. Id. ¶46.
Some phrases of the PSR — such as its reference to “the total of 7266.08 grams of cocaine base involved in this conspiracy”, id. (emphasis added) — do suggest a finding that each transaction was indeed part of a single conspiracy. Even then, however, the PSR appears to treat the scope of the conspiracy as the same for each participant, without making any effort to define Cabrera-Baez’s own agreement. In some conspiracies, of course, each participant has joined (implicitly or explicitly) in the overall scheme, so that the scope of the conspiracy is identical for each: while many participants in a “hub and spoke” conspiracy are parties only to small sub-conspiracies, see, e.g., Kotteakos v. United States, 328 U.S. 750, 66 S.Ct. 1239, 90 L.Ed. 1557 (1946), all the participants in a “chain” conspiracy are “parties to the larger common plan”, Blumenthal v. United States, 332 U.S. 539, 558, 68 S.Ct. 248, 257, 92 L.Ed. 154 (1947). But the PSR here did not engage in this analysis.
One paragraph in the report can be read to suggest that the participants agreed to help promote each other’s drug businesses, in which case each participant’s “independent’ sales might be thought to be in furtherance of their joint undertaking. See PSR ¶ 32. But we do not think that the report actually adopted this theory. Had it done so, it would have attributed far larger quantities of drugs to Saro and Cabrera-Baez than it actually did; to judge from the report, the drug ring to which Saro and Cabrera-Baez belonged involved at least 26 people. See id. ¶¶ 5-31.
Ultimately, it may well be proper for the sentencing court to hold Cabrera-Baez responsible for much the same quantity of drugs that the PSR attributed to him. Portions of the trial evidence suggested a particularly strong linkage between Saro and Cabrera-Baez, who may have acted as partners with each other even if they should not be held responsible for the “independent” sales of the many other participants in the same drug network. See, e.g.,-Trial Transcript (“Tr.”) (April 18, 1991) at 87 (indicating that Cabrera-Baez “normally” was with Saro for drug transactions); id. at 101 (stating that Cabrera-Baez and Saro shared the same beeper number); id. at 107 (indicating that Cabrera-Baez and Saro may have jointly supplied the crack involved in one of the sales for which Saro alone was charged); Appendix for Cabrera-Baez at 209-11 (transcript of videotaped negotiations in which Saro and Cabrera-Baez jointly arranged to sell at least 1.5 kilograms of crack). And even if Cabrera-Baez and Saro may best be viewed as competing suppliers, their relation with the network might have been such as to make them responsible for its entire distribution activity. See, e.g., Edwards, 945 F.2d at 1393; cf. Blumenthal, 332 U.S. at 559, 68 S.Ct. at 257. But the PSR did not make the findings of fact necessary to support either of these theories.
Our interpretation of the report is colored by the misleading nature of the Application Notes that accompanied the “relevant conduct” provision of the Sentencing Guidelines at the time óf Cabrera-Baez’s sentencing. According to the Notes, “Where it is established that the conduct was neither within the scope of the defendant’s agreement, nor was reasonably foreseeable in connection *284with the criminal activity the defendant agreed to jointly undertake^ such conduct is not included in establishing the defendant’s offense level under this guideline.” U.S.S.G. § IB 1.3, Application Note 1 (Nov. 1, 1990) (emphasis added). The tangle of double negatives leaves open the possibility that a defendant can be held responsible for conduct simply because it was foreseeable to him, without regard to the scope of his agreement. The current version of the guidelines and the Application Notes, echoing standard conspiracy law, makes clear that the third category of “relevant conduct” includes a co-eonspirator’s act only if it was “both in furtherance of, and reasonably foreseeable in connection with, the criminal activity jointly undertaken by the defendant.” U.S.S.G. § 1B1.3, Application Note 2 (emphasis added). But given this confusion in the old Notes, we can readily believe that the probation officer who prepared Cabrera-Baez’s pre-sentence report neglected the “in furtherance of’ requirement.
In our view, it was obvious error for the district court to adopt the report’s calculations without any further comment or analysis, because the legal standard applied by the report seems to conflict with the well-' established principles of conspiracy law. In addition, we think it reasonably likely (though not necessarily more likely than not) that a factfinder applying the proper legal standard would not hold Cabrera-Baez responsible for the 296.85 grams of crack stemming from the offenses with which Saro alone was charged, or for the 1.85 grams that were chalked up to Cabrera-Baez simply because they were found in an apartment “used by various members of the conspiracy”.
We have more difficulty dealing with the 2000 grams involved in an attempted distribution on July 20, 1989. It appears likely to us that the author of the pre-sentence report attributed this quantity to Cabrera-Baez simply because it was attributable to some of his co-conspirators, without regard to whether the attempt was in furtherance of Cabrera-Baez’s agreement with these people. If so, the attribution requires re-examination for the reasons discussed above. But the author may instead have placed this attempt into what we have labeled the second category of “relevant conduct” — i.e., he may have found that Cabrera-Baez personally aided and abetted the attempt.
There was some evidence at trial to this effect. The attempt concerned a deal in which an undercover agent of the Drug Enforcement Administration (“DEA”) was supposed to buy 2 kilos of crack for $41,000; the agent called off the deal when an unexpected number of people showed up at the rendezvous on the sellers’ side. The agent and two of Cabrera-Baez’s co-conspirators testified that Cabrera-Baez was among that number. If one inferred that he was present to help protect the sellers, and that he knew what was going on and intended to assist in the illegal venture, a factfinder might well find his presence culpable, cf., e.g., United States v. Ortiz, 966 F.2d 707 (1st Cir.1992), and hold him as an aider and abetter.
But Cabrera-Baez’s lawyer disputed that he was there at all, see Tr. (May 3, 1991) at 20, 26-27, and on cross-examination the undercover agent conceded that he couldn’t really say whether Cabrera-Baez had been present, Tr. (April 18, 1991) at 23. Neither the pre-sentence report nor the district court made any finding on this issue, and they could hardly piggyback on the findings of the jury, which acquitted both Saro and Cabrera-Baez on this count. While the district court did explicitly state that a preponderance of the evidence showed Saro’s guilt, the evidence of Saro’s involvement was independent of (and substantially stronger than) the evidence of Cabrera-Baez’s involvement. The undercover agent was firm about Saro’s presence at the sale. And the testimony at trial indicated that it was Saro who was behind the negotiations, who supplied the 2 kilos, who caused them to be transported to the distribution site, and who told one of his cronies to try to set up a new rendezvous with the DEA agent after the deal fell apart. Tr. (April 18, 1991) at 108-12. The district court’s findings of Saro’s guilt, then, do not constitute implicit findings of Cabrera-Baez’s.
*285Indeed, the only relevant factual finding that would support holding Cabrera-Baez hable on an aiding-and-abetting theory was clearly erroneous. The report stated (and hence the district court found) that Cabrera-Baez and others “negotiated” the deal. PSR ¶ 36. But this finding has no apparent support, and indeed appears to contradict the DEA agent’s testimony. See, e.g., Tr. (April 18, 1991) at 23.
Some courts have held that factual findings in pre-sentence reports, if not the subject of a timely objection, are immune from plain-error review. See, e.g., United States v. Young, 981 F.2d 180, 188 (5th Cir. 1992); United States v. Caba, 955 F.2d 182, 187 (2d Cir.1992); United States v. Saucedo, 950 F.2d 1508, 1518 (10th Cir.1991); United States v. Pilgrim Market Corp., 944 F.2d 14, 21 (1st Cir.1991). But cf. United States v. Rodriguez, 15 F.3d 408, 416 n. 10 (5th Cir. 1994); United States v. Helmsley, 941 F.2d 71, 98 (2d Cir.1991); United States v. Morales-Diaz, 925 F.2d 535, 540 (1st Cir.1991). We see no warrant for this categorical rule. To be sure, since the obviousness of an error is assessed from the sentencing court’s perspective, factual errors in pre-sentence reports may well tend to survive plain-error review more readily than legal errors. Presentence investigators often have access to a much broader spectrum of information than the trial court itself, and so the fact that the trial record does not support a particular finding does not necessarily mean that the finding is inaccurate. Sentencing courts generally are entitled to rely on the unchallenged findings of pre-sentence reports. But at least when those findings are internally contradictory, wildly implausible, or in direct conflict with the evidence that the sentencing court heard at trial, factual errors can indeed be obvious. We think that the district court’s finding that Cabrera-Baez “negotiated” the July 20 deal fits the last of these descriptions, and hence was obvious error. Since this finding is the sole basis for fitting the 2 kilos into the second category of “relevant conduct”, further factual findings would be necessary to sustain the attribution of this quantity to Cabrera-Baez on an aiding-and-abetting theory.
Of the 7266.08 grams that the PSR attributed to Cabrera-Baez, we are left with 4967.38 grams. This quantity consists of (1) 3.38 grams found in an apartment allegedly rented to Cabrera-Baez under a pseudonym, (2) 1964 grams sold on September 17, 1990, and (3) 3000 grams discussed in negotiations on August 31, 1989. Cabrera-Baez does not contest the attribution of the first of these amounts. As for the 1964 grams sold by Miguel Profeta to the undercover DEA agent on September 17,1990, Cabrera-Baez argues that he should be held accountable only for the 700 grams of this amount that he contributed. But the pre-sentence report (and hence the district court) explicitly found that he “assisted in the ... September 17, 1990, sale”, not only by “supplying several hundred grams of cocaine base to Mr. Profeta” but also by “storing the cocaine, ... obtaining a rental car for Mr. Profeta to use in the proposed two kilogram sale, and instructing his two workers ... to assist Mr. Profeta in the drug sale.” PSR ¶40. The report appears, therefore, to have attributed the entire amount to Cabrera-Baez on an aiding- and-abetting theory, and there is ample support for doing so.
The report also attributed 3000 grams of crack to Cabrera-Baez in connection with negotiations held on August 31, 1989. The attribution of at least 1500 grams to Cabrera-Baez is amply supported. The second 1500 grams depend on testimony of the undercover DEA agent indicating that Cabrera-Baez and Saro agreed to make two deliveries of 1500 grams each, one to take place on August 31 and one at an indefinite later time. While this testimony is not borne out by the transcript of the DEA’s videotape of the negotiations, it would not be wholly unreasonable to rely on the agent’s recollection: the transcript was translated from Spanish to English, and it contains several gaps where the conversation was inaudible. We cannot find plain error in attributing the full 3000 grams to Cabrera-Baez.
Although 4967.38 grams withstand plain error review, this falls 32.62 grams short of what would justify a base offense level of 40. Bearing in mind the standard of “prejudice” discussed above, we think a remand for fur*286ther consideration of Cabrera-Baez’s base offense level is appropriate; While we do not think it “more likely than not” that this remand will result in a reduction of Cabrera-Baez’s base offense level, a reduction seems sufficiently likely to justify a remand. And we see no reasoned basis, in the circumstances of this case, for declining to exercise our discretion to correct the plain error that we have found. Cf. Olano, — U.S. at -, 113 S.Ct. at 1778-79.
Because the parties have devoted considerable ink to the question of whether the appellants were tried within the deadline set by the Speedy Trial Act, we briefly address this issue. Even if the appellants were correct that the district court should not have granted the government a continuance under 18 U.S.C. § 3161(h)(8)(A), we still would find no violation of the applicable time limits.
Under 18 U.S.C. § 3161(c)(1), the appellants’ trial was required to begin within 70 days from October 17, 1990, the date their original indictment was filed. It in fact began on April 8,1991,173 days later. But the 70-day clock stops during certain periods. Section 3161(h)(1)(F) calls for the exclusion of the entire time (whether or not reasonable) between the filing of “any pretrial motion” and the conclusion of a hearing on the motion, or, if no hearing is held, the day the court “receives all the papers it reasonably expects” to help it decide the motion. See Henderson v. United States, 476 U.S. 321, 329-31, 106 S.Ct. 1871, 1876-77, 90 L.Ed.2d 299 (1986); United States v. Wilson, 835 F.2d 1440, 1442 (D.C.Cir.1987). Once the hearing has been held (or, in the case of motions submitted without a hearing, all the relevant papers have been filed), the statute also calls for the exclusion of a period “not to exceed thirty days” during which the court actually holds the motion under advisement. 18 U.S.C. § 3161(h)(l)(J). Finally, we have understood § 3161(h)(7) to mean that “an exclusion applicable to one defendant applies to all codefendants”. United States v. Edwards, 627 F.2d 460, 461 (D.C.Cir.1980).
The appellants raise only one argument that is not explicitly precluded by these principles. They contend that under § 3161(h)(l)(J) the maximum period excludable once a motion is taken under advisement should be considerably less than 30 days when the motions are not complicated or numerous and hence do not require lengthy deliberation. This argument cuts strongly against the tenor of Wilson, but it makes no difference here. Even if one totally ignored § 3161(h)(l)(J) in computing exclusions, and allowed the trial court absolutely no time to decide motions after taking them under advisement, the 70-day clock would have been running for only 65 days.
Finding the appellants’ other arguments meritless, we affirm both appellants’ convictions, and we affirm Saro’s sentence. We remand Cabrera-Baez’s case for further consideration of his base offense level.
So ordered.