dissenting.
What to do when civil litigants invoke the Fifth Amendment’s privilege against self-incrimination during discovery but waive the privilege on the eve of trial? This is an important question, for when used tactically a late waiver of the privilege can wreak havoc on an opposing party and create a fundamentally unfair trial. As the Third Circuit put. it in SEC v. Graystone Nash, Inc., 25 F.3d 187, 191 (3d Cir.1994):
[T]he adverse party — having conducted discovery and prepared the case without the benefit of knowing the content of the privileged matter — would be placed at a disadvantage. The opportunity to combat the newly available testimony might no longer exist, a new investigation could be required, and orderly trial preparation could be disrupted. In such circumstances, the belated waiver of the privilege could be unfair.
When a court is faced with a party who waives the Fifth Amendment privilege close to trial, it must manage the situation through “means which strike! ] a fair balance and accommodate!] both parties”— that is, “both a litigant’s valid Fifth Amendment interests and the opposing parties’ needs in having the litigation conducted fairly.” United States v. 4003-4005 5th Ave., 55 F.3d 78, 84-85 (2d Cir.1995). If it finds that the system has been gamed for unfair advantage, the court should either prevent the party from waiving the privilege and testifying at trial, or, as a lesser sanction, allow the opposing party to *748impeach the formerly silent party with its prior silence. Harris v. City of Chicago, 266 F.3d 750, 754 (7th Cir.2001). None of the decisions that I have cited places the burden on the opposing party to move for a continuance in order to mitigate the prejudice it suffers because of the waiving party’s late decision.
This case presents a straightforward legal question: should a district court bar the testimony or allow the impeachment of a party who waives the Fifth Amendment privilege at the last minute, even if the party has left enough time for some discovery in the final, hectic moments before trial? Or should these sanctions be reserved for the situation in which a formerly silent party waits until trial has actually begun before waiving the privilege, thereby precluding any discovery at all? The majority takes the latter position, implicitly if not explicitly; I take the former.
In reaching its conclusion, the majority relies on Harris, in which the defendant refused to participate in discovery entirely, and yet was allowed to testify at trial. We reversed the district court’s decision to allow the testimony and its refusal to let the plaintiff impeach the defendant with his prior silence, holding that this was “tantamount to allowing [the defendant] to avoid discovery altogether.” 266 F.3d at 754. The majority leaves the impression that Harris is not simply one situation where sanctions are appropriate — it is the only situation where sanctions are appropriate. Some discovery, any discovery— no matter how crammed or last-minute; no matter what tactical advantage it affords the formerly silent party; no matter that it devastates the opposing party’s trial preparation — remedies the prejudice caused by a late waiver. This reading of Harris is too narrow, and is inconsistent with the decision of our sister circuit in Gutierrez-Rodriguez v. Cartagena, 882 F.2d 553, 576-77 (1st Cir.1989) (affirming decision barring party from testifying where party invoked Fifth Amendment for six months during discovery, sought to reverse course one month before trial, and offered to submit to discovery in the days just before trial).
Eight of the defendant police officers invoked their Fifth Amendment rights and refused to participate in discovery for a year and a half. Five weeks before trial and well after the period for fact discovery had closed, seven of them asked to be allowed to waive the privilege and testify at trial. The district court specifically found that the waiver was untimely and caused prejudiced to Evans. However, as the majority notes, the district court appears to have given the defendants an opportunity to attempt to cure this prejudice by submitting to depositions within ten days. The seven depositions did not occur within ten days of the district court’s order; instead, they were completed weeks later, in the nine business days immediately prior to trial. On the first day of trial, after a jury had already been empaneled and just before Evans’s opening statement, the district court ruled that not only could the officers testify, but Evans could not even impeach them with their prior silence.
The district court did not provide any explanation for this ruling — which was a bonanza for the defendants — yet the majority would defer to it as a reasonable exercise of discretion. Although we do not know why the district court barred impeachment, the majority assumes that it must have been because the court believed that the frantic depositions had cured all prejudice to Evans. But the majority does not discuss at any length how the district court could have reasonably reached this conclusion, because it does not weigh the prejudice to Evans against the potential curative power of the late discovery. I do so below, and I believe that the facts show *749that the district court’s decision was unreasonable. I wish to emphasize that while I disagree with this particular ruling, I certainly understand the position of a very busy trial judge who finally brings to trial a contentious, hard-fought case and is faced with multiple motions in limine and discovery disputes — matters that are not always presented to a district court with time to spare. Nevertheless, I would announce a rule that a late-waiving party should be penalized not just when its waiver precludes any discovery at all, but also in a case like this, when its waiver prevents its opponent from a meaningful opportunity to conduct discovery.
Before weighing the prejudice to Evans against the impact of the late discovery, I wish to explain, for context, why the circumstances surrounding the defendants’ waiver suggest that they were milking their Fifth Amendment privilege for as long as possible with the intention of waiving it just before trial. In this regard, the majority is incorrect when it asserts at page 18 that after Katalinic moved to waive his privilege in November 2005, the other “officers followed suit in January of 2006, around the time the special prosecutor was wrapping up his probe.” The other defendants did not move to withdraw their assertion of the privilege in January 2006. Their exact words at that time were: "when and i/the Special Prosecutor concludes its investigation, the defendants who have elected to take the Fifth may decide to waive their privilege and testify” (emphasis added). This assertion could hardly have been more equivocal. The defendants did not actually seek to testify until mid-May 2006, approximately five weeks prior to trial. Moreover, it is open to question whether the special prosecutor was “wrapping up his probe” in January 2006, as the majority states. (His investigation did not end until April, see Op. at 738, n. 1 — and he did not issue his report until July.) The point is, when the officers finally sought to waive the privilege in May, nothing had changed in terms of their potential exposure to criminal liability. The special prosecutor’s report was not yet released, and they are not arguing that they knew the special prosecutor’s findings before he announced them. What apparently had changed was that they began to consider just how bad it would look for the seven officers most intimately involved with the Cabassa investigation to refuse to testify at trial for fear of incriminating themselves.
Now to the prejudice to Evans from the defendants’ late waiver, beginning with tactical disadvantage he suffered. See Harris, 266 F.3d at 755. The defendants sat out discovery for months and months, and then, after seeing Evans’s entire case unfold, they elected to testify. They knew the strengths and weaknesses of his case; they knew where his emphasis lay; they knew what he would ask them about; they had heard testimony from not just all of his fact witnesses, but also all of his expert witnesses. Another aspect of the tactical advantage was the effect of the late depositions on Evans’s trial preparation. This was a massive case, lasting over two years from complaint to jury verdict, with over 150 deposition witnesses and a month-long trial. While the defendants’ lead counsel was busy preparing for that trial in the week before it began, Evans’s lead counsel was personally conducting all seven depositions, which were too critical to Evans’s case to be left to a second-chair lawyer. By allowing the defendants to tie up Evans’s lawyer’s final days before trial in this way, the district court created a serious imbalance between the parties.
Two of the most significant aspects of prejudice Evans suffered involved a semen test in the 1970s and Evans’s star witness, Frank Laverty, who died in late 2006. As to the former matter, one of the defendant *750officers revealed for the first time in a deposition five days before trial began that a semen test had been conducted in the 1970s and had ruled out Evans. Evans contends that he was unable to conduct any follow-up discovery on this matter by seeking verification from other witnesses or demanding the production of test results. The defendants’ efforts to get around this revelation are wholly unconvincing, especially given that Evans was ultimately exonerated in 2002 based on semen testing. The officer “was most likely mistaken that a test was run,” they contend (without citation); they also claim that while semen testing did exist at that time, it was “not reliable.” These self-interested and wishy-washy assurances are no substitute for adversarial probing.
Even more importantly, Evans’s star witness, Frank Laverty, was terminally ill with cancer in 2006. Laverty was a former Area 2 police officer, and he gave a videotaped deposition that formed one of the centerpieces of the trial with its insider account of the police abuse committed in Area 2. But because of Laverty’s health, that deposition was taped before the defendant officers were deposed. Evans’s counsel stated at oral argument that Lav-erty’s health in the summer of 2006 — when the defendants waived their privilege and were deposed — was too poor for another video deposition. That means that in the testimony that the jury heard, Laverty had nothing to say about the defendants’ assertions during their depositions, leaving the impression that there was nothing in those depositions to rebut. The Laverty situation is critical because Evans could not have mitigated this prejudice by seeking a continuance. His star witness was too sick to participate in the trial as scheduled, let alone one at an even later date.
The last-minute depositions did not cure this prejudice. In this regard, I must disagree with the majority that the officers “met the three conditions laid down by the judge” to be allowed to testify. See Op. at 742. The district court stated in June 2006, about five weeks before trial, “If they wish to testify, they have to declare that by Wednesday, provide answers to all outstanding discovery, and appear for a deposition within 10 days.” If by “appear for a deposition within 10 days,” the judge meant that the officers could show up for the depositions and then leave without being deposed, then the officers did indeed meet the condition. But as I read the record, the defendants pushed the depositions to the final days before trial. Shortly after the district court ordered the re-depositions, Evans moved to appoint an additional lawyer, Flint Taylor, for help completing this monumental task in such a short amount of time. The defendants objected to Taylor’s participation, saying that he would “harass and intimidate” the officers, and they walked out of their rede-positions. Significantly, the magistrate judge rejected the defendants’ contention about Taylor and added him as counsel several days later. It’s hard to rule out gamesmanship on either side here, but recall that the defendants created this last-minute situation by deciding to testify so late in the process. If the burden was on any party to complete the depositions quickly, it was on them.
The broader lesson of this discovery dispute is that in a massive, acrimonious lawsuit, deposing seven critical witnesses in the nine days before trial simply cannot go off without a hitch. The district court should have recognized that the defendants had created an impossible situation and sanctioned them by allowing Evans to impeach them with their prior silence.
This leads to the question the majority asks: why didn’t Evans request a continuance? Again, and critically, a continuance would not have alleviated the prejudice *751inherent in the Frank Laverty situation— more time would not have allowed Laverty, who was dying, to respond to the defendants’ last-minute depositions. But assuming that a continuance should still matter, the majority suggests that the time to request one was June 2006, when the district court ruled that the defendants could testify at trial. That misses the point, for at that time Evans thought the depositions would be completed within ten days, as ordered, rather than in the final moments before trial began. In other words, Evans didn’t know just how prejudicial the situation would become. The right question is therefore, why didn’t Evans move for a continuance on the first day of trial, when he learned that he could not even impeach the defendants with their prior silence? For even if it was until then an open question whether Evans could impeach, the case law gave him every indication that he would be allowed to by setting out impeachment as the lesser sanction to barring testimony outright. Indeed, in the weeks before trial he was still fighting the preliminary question of whether the defendants should even be allowed to testify, and Harris suggests that barring their testimony outright was a real option. I’m sure Evans was shocked on the first day of trial when he learned that on top of the earlier ruling allowing the defendants to testify, evidence of their prior silence would be inadmissible too.
As to why Evans did not request a continuance on the first day of trial, the majority states that the district court would have been receptive to such a request. Again I must disagree. The trial had already begun; the jury was empaneled and opening statements were about to get underway. Moreover, the district judge made it quite clear, especially after the original trial date of January 2006 was disrupted, that he would not push back the trial.1 I can understand why: this was a very busy trial judge balancing a full criminal docket who understandably (and fairly) wanted to move on with a month-long case. By my count, the district judge rejected no fewer than four motions to continue the trial, including one a month before trial and one the day before trial. On more than one occasion, the judge told counsel in rejecting a motion for continuance things like, “Now gentlemen, don’t waste the paper to file a motion for another extension. It’s not going to happen.” While discussing trial scheduling matters with Evans’s counsel a month before trial, the district judge said, “You can make whatever objections you want, but that’s going to be the schedule.” The judge also said in the final pretrial conference in May 2006, “I won’t have time to try this case for almost a year, so — because of other matters pending — so I’m not going to move this trial.”
Evans certainly would have made a better record by requesting a continuance, even if only to have it emphatically denied, but the law does not require a futile act. If it is clear that a motion for continuance would be denied, the failure to file one does not constitute waiver. See United States v. Dellinger, 472 F.2d 340, 371-72 (7th Cir.1972); cf. United States v. Fish, 34 F.3d 488, 495 (7th Cir.1994) (failure to seek continuance is not ineffective assistance of counsel where district court’s statements show that any motion would *752have been futile); Moody v. Polk, 408 F.3d 141, 151 (4th Cir.2005) (same).
But as I’ve noted, because of the prejudice in terms of Laverty’s video deposition, the question of a continuance is a red herring. It was also, for Evans, a Catch-22. When he found out, just before his opening statement, that the defendants would not only be allowed to testify after sitting out discovery, but also avoid impeachment with their silence, Evans was put in an unenviable position. Either he proceeded with the scheduled trial at a significant disadvantage, or he moved to postpone for another year a day in court that was 27 years in the making — one that the defendants had already delayed once with a frivolous interlocutory appeal. See Evans v. Katalinic, 445 F.3d 953, 955-56 (7th Cir.2006) (describing defendants’ two arguments as, respectively, “absurd,” and “too ridiculous to merit comment”). Since the defendants created this mess with their late waiver, the defendants — not Evans — should have been the ones to face consequences for it. But the defendants suffered no consequences at all: they effectively dodged discovery, got to testify at trial, and kept the whole thing from the jury. Today’s opinion ratifies that neat maneuver, teaching that when a party waives the privilege so late that its opponent suffers prejudice, it is the opponent, rather than the waiving party, that has to fix the situation. That is inconsistent with Hams, and it is a bad rule — it encourages gamesmanship, puts the district court in a difficult situation, and undercuts the goal of timely and fair discovery.
As the majority notes, “The jury here rejected the view that police misconduct, as opposed to something more benign, was the reason why Evans was wrongfully convicted.” See Op. at 747. That is true, but it begs the question, would the jury still have reached that decision if it had all the pertinent information? After all, Evans presented substantial evidence of disturbing police malfeasance, including that the defendants repeatedly coerced and threatened their only witness to identify him; that they locked the witness’s husband in a room to prevent him from telling prosecutors about the witness’s eyesight and credibility problems; that they lied in saying that the witness reported threats from Michael Evans in the weeks after the murder; and that they performed genital inspections on neighborhood boys as a way to pressure them to implicate Evans (which at least one of the boys did).
The main evidence for the other side was the officers’ denials. Knowing that the officers were so worried about criminal liability that they refused to speak in their own defense for a full year and a half might have made all the difference to the jury in choosing whom to believe. The majority thinks otherwise, implying that since one officer invoked the Fifth at trial and was nevertheless exonerated, the result would be the same if everyone’s silence came out. This argument is best rejected by analogy. If there was a single moment that irrevocably changed the fortunes of the tobacco industry, it was in April 1994 when the American public saw seven — not one — tobacco company executives raise their right hands and tell Congress that nicotine isn’t addictive and that smoking doesn’t cause cancer. See Allan M. Brandt, The Cigarette Century: The Rise, Fall, and Deadly Persistence of the Product that Defined America 366-69 (2007). Similarly here, the power of seven silent Chicago Police officers is not, as the majority suggests, a matter of one plus six. Seven is exponentially greater than one. Regardless, the point is not what I think about the impact of the officers’ prior silence, or what my colleagues in the majority think. What matters is what the jury would have thought, had it been given the *753opportunity to consider this critical evidence.
I agree with the majority that what happened to Michael Evans was a tragedy: he spent 27 years in prison for a crime for which he has been exonerated and pardoned. He deserved justice in his civil trial, but he did not receive it because the trial was fundamentally unfair. Accordingly, I respectfully dissent.
. The majority states at page 746 that the district judge merely denied motions to continue based on “scheduling conflicts.” That is incorrect. To take just two examples, the defendants’ emergency motion to continue the day before trial argued that publicity over the case and the expected release of the special prosecutor’s report would prevent the selection of a fair jury. And the defendants’ motion to continue filed in November 2005 asked for more time to conduct discovery.