(dissenting).
I start my dissent with a question that troubles me. Why did this deliberate brutal beating of Officer Cox take place? The government describes the beating in its brief, which accurately summarizes the record:
As Cox was preparing to climb the fence in pursuit of Brown, Cox was struck from behind with a blunt object by police officers who apparently mistook him for a suspect. The officers beat and kicked Cox repeatedly in the head, back, face, and mouth. Cox then heard an officer shout, “Stop, Stop, he’s a cop. He’s a cop,” and the officers all fled. With no one left to assist him, Cox was forced to use the bumper of a police car to pull himself off the ground. Cox, bleeding and seriously injured, was later taken in an ambulance to a hospital for treatment.
The majority opinion gives a terse summary of the event that led to this case: “As Cox stood at the fence, he was brutally attacked and beaten by other arriving police officers who thought he was one of the suspects (and also thought, wrongly, that a policeman had been shot). The assaulting officers discovered their mistake and dispersed, leaving Cox badly injured.”
Even if we accept the hypothesis that the beating resulted from mistaken identity, my question still remains. Surely it is not normal or standard police procedure to administer a brutal beating to a fleeing felon where no resistance is made. There is nothing in the record to suggest that Cox fought back. All he did was try to protect himself from repeated kicks to the head. It is troubling to me that the police acted as barbarians rather than police officers sworn to uphold the law.
But, of course, this has nothing to do with the issues before us. In our quest for justice, however, I think it advisable to not only address the specific issues in a case but the framework which gave rise to the ease.
My dissent is based on two grounds. The first is that I think the majority opinion has applied the wrong law. I think the case should have been decided on the “law-of-the-ease” doctrine. Rather than engage in an analysis of that doctrine again, I am attaching to this a copy of my panel decision setting forth in detail why this doctrine should be applied here.
My second reason for dissenting is the majority’s decision to remand. This probably means another appeal and further consideration by the court of appeals. I think that the case ought to be decided by *17the en banc court. The majority sets forth a number of reasons for the remand:
The evidence underlying the Brady claim has never been systematically assembled and fully analyzed in any court in any of the various decisions to date. There are fragments of discussion in several of the district court decisions, Conley IV, 164 F.Supp.2d at 221-23; Conley II, 103 F.Supp.2d at 51-58; but nowhere are all of the items of supposed new evidence set forth, their significance considered, them status assessed under Brady, and a judgment made (based on a review of all the evidence in the entire trial) as to whether the Brady-qualifying items of new evidence (taken together) undermine confidence in the verdict.
Save in the most straightforward case, an en banc court is not the best initial forum for this kind of undertaking. Some Brady issues can require fact-finding as to what the government knew and when it knew it. And, where (as here) the case was circumstantial and much of the evidence concerns arguable impeachment of key witnesses, en banc review is generally assisted by having a well-worked-out district court assessment before the appellate court attempts its own evaluation.
With respect, none of the stated reasons are persuasive. The majority says that the new evidence has never been fully collected and analyzed. Quite the contrary is true. There is an extensive record in this case which is familiar to the en banc court. The parties themselves agree on which materials constitute the newly discovered evidence. All that is left is an analysis under Brady v. Maryland, 373 U.S. 83, 87, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963).
The majority also suggests that additional fact-finding is necessary to determine “what the government knew and when it knew it.” This is true as a general matter, but is simply not applicable here. What the government knew is only relevant to the first prong of the Brady determination, namely whether evidence favorable to the accused was suppressed by the government, either willfully or inadvertently. See Strickler v. Greene, 527 U.S. 263, 282, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999). This case has never hinged on whether the government suppressed evidence; in fact, the government concedes in its brief that the evidence was “inadvertently overlooked ... in providing discovery.”
Instead, the focus of this ease is on the final prong of Brady, which rests on whether there is a reasonable probability that, had the newly discovered evidence been disclosed to the defense, the result of the proceeding would have been different. See Kyles v. Whitley, 514 U.S. 419, 433-34, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995). Deciding whether the outcome of a trial would be different is not simply a fact-bound issue, it is interlaced with a constitutional determination. See Brady, 373 U.S. at 86, 83 S.Ct. 1194 (explaining that the overarching issue is whether the government’s conduct violates the Due Process Clause of the Fourteenth Amendment). Thus, the critical question in this case is a matter of law, not fact, and it is a question that we can decide. See Strickler, 527 U.S. at 289, 119 S.Ct. 1936 (giving no deference to analysis by district court or court of appeals when examining habeas corpus petition in which defendant claims Brady material creates a reasonable probability of a different outcome).
This brings me to the majority’s final reason to remand. The majority says that it makes sense to have a district court make the Brady determination because the government’s case was circumstantial and all the newly discovered evidence relates *18to impeachment of key witnesses. On its face, this justification seems persuasive. After all, the district court judge who listened to the witnesses’ testimony, saw their demeanor on the stand, and observed the jury’s reaction seems best positioned to determine whether new impeachment evidence would change the result of the trial. But the majority’s decision to remand this case to a different district court judge eviscerates any benefit that might be gained from a remand.
We can determine whether the outcome of the trial would be different using circumstantial and impeachment evidence just as well as a new district court judge, and considering the appellate record here, with which we are all familiar, in a significantly shorter time. Review of circumstantial evidence is part of the gristmill for appellate courts. In addition, circumstantial evidence is just as reliable as testimony and, at times, more reliable because it does not depend on the memory or judgment of what a witness saw and remembered and it is not subject to the biases and prejudices that are part of all human beings.
With respect, I consider a remand an unnecessary waste of time and judicial resources. It is going to take the new district court untold hours to attain a working knowledge of the record, which this court already has. I also point out that five of the seven members of the en banc court have had considerable experience as trial court judges. This court is well versed in looking at a record and deciding whether an outcome of a trial would be different in light of new evidence.
It is clear from the appellate record that Kenneth Conley has had more than his “day” in court. After being convicted by a jury, Conley appealed. We upheld the conviction unanimously. United States v. Conley, 186 F.3d 7 (1st Cir.1999). Certiorari was denied. Conley v. United States, 529 U.S. 1017, 120 S.Ct. 1417, 146 L.Ed.2d 310 (2000). Conley appealed again, this time on the grounds of newly discovered evidence. We unanimously reversed the district court because of erroneous legal rulings. United States v. Conley, 249 F.3d 38 (1st Cir.2001). We ended our opinion by stating: “The sentence of the district court, which we affirmed in our prior opinion, shall be executed.” Id. at 47. The sentence was not executed. Conley filed a 28 U.S.C. § 2255 (2000) motion based on essentially the same newly discovered evidence alleged in his prior case. On appeal for a third time, we applied the “law-of-the-case” doctrine in a two to one opinion. Conley’s motion for an en banc was granted and the majority opinion was recalled.
I can think of no good reason why we should revisit this case, let alone remand it to a different district court judge. This court has all it needs to decide the Brady issue itself. We have the entire record and we are familiar with it. Moreover, placing the case in the hands of another district court judge can be construed to infer an abdication of our duty of review. Our duty is to review the decision of the district court, not to look to it for help in deciding a difficult legal issue. Judge Tor-ruella’s treatment of the Brady issue is both thorough and correct. Because I believe this court should decide the Brady issue itself, and because I agree with the substance of Judge Torruella’s analysis, I concur in his dissent.
My final comment is a direct reply to the majority’s discussion of justice. Simply put, “Justice delayed is Justice denied.” I think this court has a duty to decide this case and that a remand is unnecessary.
ATTACHMENT
BOWNES, Senior Circuit Judge.This is the third appeal arising from defendant-appellant Kenneth Conley’s jury *19conviction of perjury in violation of 18 U.S.C. § 1628 and obstruction of a grand jury investigation in violation of 18 U.S.C. § 1503. The conviction followed Conley’s testimony before a grand jury, which was investigating the alleged beating of plainclothes police officer Michael Cox by other police officers.
This case first came before us on direct appeal after Conley’s conviction. We affirmed the conviction and the sentence of thirty-four months, ruling explicitly that the evidence was sufficient to support the conviction. United States v. Conley, 186 F.3d 7, 20 (1st Cir.1999), cert. denied, 529 U.S. 1017, 120 S.Ct. 1417, 146 L.Ed.2d 310 (2000) (Conley I). Conley then moved for a new trial based on newly discovered evidence, violations of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), and jury misconduct. The district court granted the motion, finding that a new trial was warranted “in the interésts of justice.” United States v. Conley, 103 F.Supp.2d 45, 57-58 (D.Mass.2000) (Conley II). We reversed, ruling that the district court did not apply the correct legal test. United States v. Conley, 249 F.3d 38, 46-47 (1st Cir.2001) (Conley III).
The present appeal arises from a petition under 28 U.S.C. § 2255, which Conley filed shortly after our opinion in Conley III had issued. The district court set aside Conley’s conviction:
The Judgment of Conviction under which Petitioner is presently restrained was obtained in violation of the Due Process Clause of the Fifth Amendment to the Constitution of the United States in that exculpatory evidence was withheld from Petitioner during trial, which resulted in a verdict not worthy of confidence.
Conley v. United States, 164 F.Supp.2d 216, 217 (D.Mass.2001) (Conley IV). We reverse the court below.
BACKGROUND
We set forth the factual background and much of the procedural history of this case in Conley I, 186 F.3d at 11-15, and Conley III, 249 F.3d at 40-44, and need not reiterate it. Only the following points bear emphasis:
In his motion for a new trial pursuant to Fed.R.Crim.P. 33, Conley focused on three pieces of evidence: the government’s failure to disclose Charles Bullard’s grand jury testimony; the government’s knowing reliance on Richard Brown’s perjured testimony that the Boston Police Department had brought drug charges against him in retaliation for the testimony he gave at the civil trial; and the government’s failure to disclose the transcript of an interview of Officer Richard Walker by the Internal Affairs Division (IAD) of the Boston Police Department, in which Walker made a tentative photo identification of the tall white officer who chased and arrested Brown as an officer other than Conley.
The district court’s opinion on this motion analyzed these three pieces of evidence. After setting forth the standard for a new trial based on newly discovered evidence, United States v. Wright, 625 F.2d 1017, 1019 (1st Cir.1980), and a prosecutor’s obligation to disclose exculpatory evidence, Brady, 373 U.S. at 87, 83 S.Ct. 1194, it considered the separate and cumulative effect of the evidence. Conley II, 103 F.Supp.2d at 51-55. The district court determined that the government’s failure to disclose Bullard’s grand jury testimony did not violate Brady, and that Officer Walker’s IAD interview transcript was “ 'inconclusive’ as to the government’s duty of disclosure and defense counsel’s diligence.” Conley III, 249 F.3d at 44. The *20court went on to frame two questions for its consideration:
[Question One:] Did the prosecution have and withhold information from defense counsel that would have led a reasonable person to expect that a civil trial would occur, similar to the civil trial that did in fact occur after the criminal conviction and sentence in this case, and that the testimony at the civil trial would be substantially as we now know it was in fact?
[Question Two:] If so, were defense counsel so severely impeded in their preparation of an overall defense strategy and in the performance of the function of cross-examination of those particular witnesses, out of the larger number of police officers, including both uniformed and undercover officers, who were in the vicinity of the brutal beating of Michael Cox, an undercover Boston police officer, by a uniformed Boston police officer, that in the interests of justice a new trial should be allowed?
Conley II, 103 F.Supp.2d at 57-58.
The district court answered the first question in the affirmative. The second question, it stated,
cannot be determined as a matter of law, under the applicable legal standard explained in Part III of this opinion [discussing Wright and Brady, inter aha]. Instead, in the unique circumstances of this case, I conclude that the determination to allow or not to allow a new trial is one committed to an exercise of discretion by the court to which the legal system assigns responsibility for making the determination.
Id. at 58 (emphasis added). The district court then used the discretion it had given itself to order a new trial “in the interests of justice.” Id.
In our review of the district court’s opinion on Conley’s motion for a new trial, we discussed the requirements for a new trial based on newly discovered evidence and/or violation of Brady. Conley III, 249 F.3d at 44-45. We concluded that the district court erred in allowing a new trial “in the interests of justice” instead of applying either the Wright or Brady standards:
As we explained supra, a new trial may be ordered in this case only if the standards set forth in Wright and/or Brady are satisfied. Both Wright and Brady require a showing that the evidence was material and that the defendant was prejudiced to some degree. We must defer to the district court’s explicit findings as to the Bullard and Brown testimony, as well as to its statement that prejudice could not be determined upon a consideration of the evidence as a whole. Therefore, there is no basis for remanding this matter, and we REVERSE the district court’s order.
Id. at 47 (internal citations omitted). We directed that the district court’s sentence, which we had affirmed in Conley I, be executed. Id.
Only one week after we issued Conley III, Conley brought a motion to set aside conviction under 28 U.S.C. § 2255. Again, the district court held that Conley was entitled to relief from judgment. Conley IV, 164 F.Supp.2d at 223-24.
In Conley IV, on appeal before us now, the district court made the following findings:
(1) Richard Walker testified at the trial of defendant Conley that Walker went through a hole in the fence and made his way to the bottom of an incline;
(2) when at the bottom of the incline he saw two men standing in the street, one tall (six feet or six feet two) and the other shorter (about ... five nine);
*21(3) the prosecutor used this evidence in his closing to suggest that the taller was Conley;
(4) during testimony of March 27, 1995 to the Internal Affairs Division (IAD) of the Boston Police Department, Walker was shown a group of photographs and with some uncertainty selected two as persons at the bottom of the incline, neither of whom was Conley;
(5) defendant Conley claims that this evidence was unavailable to Conley and his attorney at the criminal trial;
(6) this and other material evidence came out in open court in the later trial of a civil action brought by Michael Cox against several officers, including the defendant Conley, in December 1998, some .six months after the defendant was convicted in the criminal trial; and
(7) even though the testimony given at the December 1998 civil trial could not have been withheld by the prosecutor in the criminal trial, because it did not then exist, the defendant Conley argued in his motion for new trial and continues to press the argument now that the prosecutor had withheld, during and before the criminal trial, information that the prosecutor then had about the IAD proceedings, and if the prosecutor had made that information available to Conley and his attorney in time for use during the criminal trial, it would have made a material difference in the defense strategy, including cross-examination.
Id. at 221-22.9 The court made no findings concerning the other evidence at issue in its earlier opinion, i.e., the testimony of Bullard and Brown.
The district court held that the correct legal standard was found in Wright, 625 F.2d at 1019. Conley IV, 164 F.Supp.2d at 222. It then restated the same two questions it had formulated in Conley II, set forth supra at page 4. It held:
The newly discovered evidence, taken together with all the conflicts presented in evidence known to the defense before and during the criminal trial, presents a dramatically more compelling basis for finding that defense counsel’s opportunity to present a creditable challenge to the government’s case as a whole and to cross-examine effectively particular witnesses was severely impeded.
Id. at 223.
Based on the factual findings recited supra, the court concluded that Conley satisfied the elements of the Wright test.10
I conclude that the newly discovered evidence is highly probative and neither immaterial nor cumulative in nature. Instead, it is evidence that strikes at the heart of the prosecution’s case, one which is largely based on the credibility of its witnesses. Indeed, the strength of this evidence leads me to find that the defendant has met his burden of show*22ing a probable result of acquittal in a new trial.
Id. The government appeals.
DISCUSSION
Under 28 U.S.C. § 2255, a prisoner in federal custody may petition the sentencing court to vacate, set aside or correct the sentence on the ground that the sentence was imposed in violation of the Constitution or laws of the United States. Brackett v. United States, 270 F.3d 60, 63 (1st Cir.2001). Claims that previously have been addressed on direct review, however, may not be readjudicated collaterally under § 2255 absent equitable considerations, such as actual innocence or cause and prejudice. Withrow v. Williams, 507 U.S. 680, 721, 113 S.Ct. 1745, 123 L.Ed.2d 407 (1993) (Scalia, J., concurring); see also United States v. Michaud, 901 F.2d 5, 6 (1st Cir.1990); Tracey v. United States, 739 F.2d 679, 682 (1st Cir.1984)(“[a]bsent an intervening change in the applicable law, issues that have been raised and decided on a motion for a new trial cannot be reconsidered in a subsequent collateral attack” (quoting United States v. Sanders, 723 F.2d 34, 36 (8th Cir.1983))); Dirring v. United States, 370 F.2d 862, 864 (1st Cir.1967).
In a related vein, the doctrine of “law of the case” is a prudential principle that “precludes relitigation of the legal issues presented in successive stages of a single case once those issues have been decided.” Field v. Mans, 157 F.3d 35, 40 (1st Cir.1998) (quoting Cohen v. Brown Univ., 101 F.3d 155, 167 (1st Cir.1996)). “For a bar to exist, an issue must have been actually considered and decided by the appellate court, or a decision on the issue must be necessarily inferred from the disposition on appeal.” Id. (internal quotation marks omitted) (quoting Commercial Union Ins. Co. v. Walbrook Ins. Co., Ltd., 41 F.3d 764, 770 (1st Cir.1994)).
The law of the case doctrine prohibits a litigant from resurrecting an issue decided by the trial court that either has not been challenged on appeal or has been decided on appeal. Id. at 40; United States v. Rosen, 929 F.2d 839, 842 n. 5 (1st Cir.1991). Another aspect of the doctrine is the “mandate” rule, which requires the district court to follow the ruling of the court of appeals. Field, 157 F.3d at 40. We review an application of the law of the case de novo. Id.
The district court’s conclusion in Conley ZF that the newly discovered evidence pertaining to Walker satisfied the Wright criteria is foreclosed by these doctrines. First, the district court essentially reversed course as to whether the Walker evidence fulfilled the requirement of prejudice. In its previous decision, it held that the question of prejudice “cannot be determined as a matter of law” under the standards set forth in Wright and Brady. Conley II, 103 F.Supp.2d at 58. That same evidence, it now says, is sufficient to warrant a new trial under Wright. Conley IV, 164 F.Supp.2d at 222.11
Moreover, the district court’s most recent treatment of the Walker evidence is *23at odds with its previous ruling as to at least one other key element. Earlier, it called Walker’s IAD testimony “inconclusive” as to the government’s duty of disclosure and defense counsel’s diligence. Conley II, 103 F.Supp.2d at 55. In Conley III, however, the court held that the Walker evidence satisfied all of the elements of Wright, including, by inference, the requirement that Conley’s failure to learn of the evidence was not attributable to his lack of diligence. 164 F.Supp.2d at 223 (discussing Wright, 625 F.2d at 1019).
In Conley III, we discussed and affirmed the district court’s factual findings as to these elements of the Wright test, and instructed the court to carry out Conley’s sentence. 249 F.3d at 46-47. That decision binds the district court. See Withrow, 507 U.S. at 721, 113 S.Ct. 1745; Field, 157 F.3d at 40. Accordingly, we reverse the district court’s order setting aside Conley’s conviction.
We know of no additional evidence that would explain the district court’s new and contrary conclusions as to prejudice and defense counsel’s diligence. While the district court mentions in passing an “enlarged record,” it makes no specific findings that support its about-face. Furthermore, the legal claims, although recast in the form of a § 2255 petition, did not differ from the claims before the district court and before us in Conley III. See Tracey, 739 F.2d at 682 (a petitioner is still barred from relitigating the subject matter of claims on collateral review even if he recasts them in different nomenclature). Nor was there any material change in the controlling law following Conley III. See id. Hence, we see no reason not to preclude the district court’s readjudication of settled issues.
For these reasons, we REVERSE the decision of the district court. We remand this case to the Chief Judge of the District Court for the District of Massachusetts with instructions that Conley’s thirty-four-month sentence be executed.
TORRUELLA, Circuit Judge.Although I fully agree with Judge Bownes’s dissenting opinion, I write separately to address the majority’s decision to remand this case. In our attempt to find and apply the proper legal framework with which to analyze Conley’s new evidence claims, this Court, together with the district court, has now heard the same arguments made regarding the same body of evidence on six separate occasions. Now, straining such jurisprudential norms as finality and repose to their breaking point, the majority elects to remand Conley’s new evidence claims — claims with which this Court is all too familiar — to a new district judge for fresh consideration. Respectfully, I disagree with the majority’s decision to remand this case, as well as the substantive reasons for doing so.
Besides regard for finality, there are three reasons why this en banc Court would have been the best forum to hear Conley’s Brady claim. First, we have developed an intimate familiarity with the record, a familiarity matched only by Judge Keeton and the parties themselves. Second, the record before us is complete: no additional fact-finding is required to assess the merits of Conley’s Brady claim. Both parties invited us to decide this matter and have had full opportunity to collect all of the evidence relevant to the consideration of Conley’s claims. Moreover, the parties agree upon the scope of the evidence to be considered under Brady and there are no facts in dispute regarding the government’s knowledge or suppression of evidence. Finally, once the heavy-lifting is done — once all the new evidence in question has been properly catalogued, labeled and evaluated — it is evident that under *24Brady, there is simply no basis for remand.
Therefore, although I am troubled by the majority’s somewhat tortured efforts to construct an exception to the law of the case to apply to this case, I do not focus my dissent on the majority’s handling of that doctrine. Instead, I will concentrate on the merits of Conley’s Brady claim. Assuming arguendo that the law of the case does not now bar Conley’s claims, and agreeing with the majority that Brady, and not Wright, applies to new evidence claims made in a section 2255 motion, I dissent today because there is no purpose to be served by remanding this case to the district court, and no plausible basis upon which, on remand, the district court might grant a new trial.
Assessing the merits of Conley’s Brady claim requires reconstruction of a good deal of trial testimony, as well as accounting for all of the items of new evidence that might come under the Brady umbrella. Thus, to establish that Conley’s Brady evidence does not create a “reasonable probability that the suppressed evidence would have produced a different verdict” Strickler v. Greene, 527 U.S. 263, 280-81, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999), I will briefly summarize the evidence offered against Conley at trial and assess any impact the new evidence might have produced.
I. The Trial Testimony12
In his grand jury testimony, Conley testified that he was the first officer in pursuit of Robert Brown. He stated that he was forty feet behind Brown and that he climbed over the chain link fence “seconds after” Brown did. He further testified that he did not see Michael Cox or any other plainclothes officer chasing Brown ahead of him.
The testimony of Michael Cox, Robert Brown, and Richard Walker conflicts with Conley’s account. As we stated in our original opinion on the sufficiency of the evidence:
Both Cox and Walker placed Cox at the exact same place where Conley claims to have climbed over the fence. According to their testimony ... Cox was “right behind” Brown, approximately three feet behind him, as Brown approached the fence. When Brown reached the fence, Cox was even closer. At that point, Cox was close enough to make contact with Brown and attempt to pull him back over the fence. Brown corroborated this version of events when he testified that a black man wearing a “black hoody” was behind him as he ran toward the fence and had just started to come over the fence after him when he observed the black man being struck on the head by a police officer. Brown confirmed that the person behind him was close enough to make contact with his foot as he scaled the fence.
Conley’s testimony that he scaled the fence “within seconds” of seeing Brown go over the fence, and that he scaled the fence in the same location as Brown does not square with the testimony of Cox, Walker and Brown. Conley’s version of events provides for no reasonable gap in time during which he could have missed observing Cox at the fence. Indeed Conley concedes that if the Cox/Walker/Brown version is true, he would have seen Cox at the fence. In reaching its verdict, the jury apparently found the Cox/Walker/Brown version *25more credible.13
Conley I, 186 F.3d at 20.
The jury found Conley guilty of perjury for denying that he saw Cox chase, pursue, and grab hold of Brown as Brown ran toward and climbed the fence at Woodruff Way. The jury acquitted him of another count of perjury, finding that his denial of witnessing Boston police officers beating Cox was not perjurious.
This Court affirmed the sufficiency of the evidence supporting the jury’s findings, concluding that the government had “produced ample circumstantial evidence from which a rational jury could conclude that Conley’s statements were false beyond a reasonable doubt.”14 Conley I, 186 F.3d at 19.
II. The New Evidence
Conley moved for a new trial on March 24, 2000. All of Conley’s new evidence can be sorted into three categories: (1) Charles Bullard’s grand jury and civil trial testimony [“Bullard Testimony”]; (2) the allegedly perjured testimony of Robert Brown [“Brown Testimony”]; and (3) the statements and testimony of Richard Walker [“Walker Evidence”]. As Conley’s Brady claim has come into sharper focus, it is now apparent that his hopes for a new trial under Brady hinge entirely upon the impeachment value of this last category of exhibits, the Walker Evidence.15 However, to provide a comprehensive account of all Conley’s new evidence, I will briefly dispatch the Bullard and Brown testimony before appraising the weight of the Walker Evidence.
A. The Bullard Testimony
Charles Bullard was a Boston security guard who was riding in the same vehicle as Officer Cox when Cox’s car came to a stop at the end of Woodruff Way. Bullard testified before the grand jury but was not called as a witness at trial. Six months after Conley was convicted, Bullard testified in a civil suit brought by Michael Cox against several officers.
In the civil trial, Bullard testified that Cox exited the passenger side of the un*26marked police cruiser and went forward to the front of the Lexus. Bullard testified that he “lost track of [Cox] after that,” and that he never saw Cox run toward the fence. According to Conley (who has argued that Cox did not run directly to the fence), Bullard’s testimony that Cox went to the front of the Lexus, and not directly to the fence, contradicted Cox’s testimony and supported Conley’s theory of events.
Although Bullard’s subsequent civil trial testimony is obviously not a matter which the government might have been required to disclose under Brady, Conley has argued that Bullard’s civil trial testimony was sufficiently similar to his grand jury testimony, such that the grand jury testimony should have been disclosed to the defense. Conley’s argument is based upon an affidavit by Bullard, in which Bullard stated that to the best of his recollection, his testimony before the grand jury was the same as the testimony he gave before the civil trial.
The district court concluded that the Bullard Testimony was not evidence that the government was required to disclose under Brady. United States v. Conley, 103 F.Supp.2d 45, 51-52 (D.Mass.2000) (“Conley II ”). The court found that Bul-lard’s civil trial testimony differed from his testimony before the grand jury in material respects. Specifically, the court noted that in his grand jury testimony, Bullard said nothing about Cox going forward to the front of the Lexus; instead, he testified that “Cox jumped out and went toward the right.” Id. Because the Bullard Testimony is in no way exculpatory, it does not come within the rubric of Brady. Strickler, 527 U.S. at 281-82, 119 S.Ct. 1936.
B. The Brown Testimony
Conley claimed that Robert Brown perjured himself at trial when he stated that state drug charges pending against him were manufactured by Boston police in retaliation for his testimony against Conley. According to the defense, the prosecution elicited this testimony in order to lend credibility to Brown by portraying him as having suffered for his testimony. See Conley II, 103 F.Supp.2d at 52-54 (setting forth the perjury allegation in considerable detail). Conley alleged that the prosecutor solicited perjurious statements from Brown in direct examination and then relied on those statements in his closing argument. Conley has claimed that the government knew that Brown’s statements were perjurious because of a federal narcotics investigation against Brown.
Although it is difficult to see how the Brown Testimony might be brought under the ambit of Brady, that is not an issue that needs to be addressed now because the district court determined that Conley failed to show that the Brown Testimony was perjurious. Id. at 54. The court found no evidence that Brown perjured himself, and additionally, expressed skepticism that evidence of a pending federal drug investigation against Brown would have been admissible under Fed.R.Evid. 403, 404(b) and 608(b). Because the court concluded as a matter of law that there was no perjury, there was no exculpatory evidence for the government to disclose under Brady.
C. The Walker Evidence
At oral argument, the parties made two helpful and important concessions regarding the scope and significance of the Walker Evidence for Brady analysis purposes. For its part, the government conceded that the Walker Evidence was in its possession and should have been turned over to the defense under Brady and its progeny. Consequently, we are not faced with *27any dispute regarding what the government knew, or when it knew it.
Conley’s attorney also narrowed the scope of our Brady inquiry by providing us with what was described as a “complete inventory” of the possible Brady material pertinent to the testimony of Richard Walker. Conley has identified four documents relating to the Walker’s testimony which he claims should have been disclosed under Brady. These are: (1) Walker’s March 27, 1995 testimony to the Internal Affairs Division (“IAD”) of the Boston Police Department (the “IAD Testimony”); (2) an internal FBI memorandum concerning a proposed polygraph examination of Walker (“FBI Memorandum”); (3) a January 30, 1995 memorandum by Lieutenant Kevin Foley of the BPD Anti-Gang Unit (“Foley Memorandum”); and (4) a skeletal BPD report prepared by Walker on February 20,1995 (the “Skeletal Report”).
(1) Walker’s IAD Testimony
On March 27, 1995, IAD agents interviewed Walker about his observations on the night of the Cox assault. During the interview, Walker was shown a group of photographs in order to identify the plainclothes officers Walker observed at the bottom of the incline.
When asked to describe the tall white male he saw at the bottom of the hill, Walker testified, “All I know is he was a very tall white male. He was medium build. I forget what he was wearing.” Walker was then given a group of photographs of officers from the Anti-Gang Unit, a unit that Conley was not in, and asked him to try to identify the officers he saw. During the interview, Walker identified the photographs of two officers, Joseph Teahan and Michael DeStefano. He was notably tentative in his identification, stating, “Like I said, I’m not sure about these photographs, but it looks like the tall officer.”
Conley claims that he was prejudiced by the suppression of the IAD Testimony because access to the testimony would have enabled him to make an issue of Walker’s credibility at trial. The testimony would have allowed him to impeach Walker’s memory of events, because Walker’s subsequent inability to identify the tall white officer conflicts with the inference made at trial that Conley was the tall white officer at the bottom of the hill.
In response, the government offers a number of arguments why the IAD Testimony offers no significant impeachment value, and further, why Walker’s unim-peached testimony helped, rather than hurt, the defense’s theory of the case.
First, the government argues that neither Walker nor the government ever suggested that Walker could identify Conley as the tall, white officer he observed. Instead, it was circumstantial evidence, corroborated by Conley’s own testimony, that established that the tall, white officer was the same one he chased after and to whom he returned the dropped radio.
Second, the “misidentification” evidence would have actually weakened the defense offered at trial. That the tall, white officer described by Walker’s testimony was in fact Conley was a point defense counsel diligently tried to prove at trial. Placing Conley at the bottom of the hill ahead of Walker helped the defense by allowing it to argue that Conley would not have had time to witness Cox’s beating before Conley scaled the fence. Defense counsel’s closing argument shows that Conley went to great lengths to use the Walker testimony to prove that Walker observed Conley at the bottom of the incline. Discussing Walker’s testimony, defense counsel stated, “I believe from the evidence it is fair to say that there can be no doubt in your *28mind that the tall white officer in plainclothes was Ken Conley.” The defense alluded to the ample circumstantial evidence tending to show that the tall, white officer was indeed Conley.
During the chase, the tall white officer dropped his radio. Officer Walker says he picked it up. And when the tall white officer apprehended the suspect, put the cuffs on him, with Walker assisting, he gave him back his radio, you dropped this. Walker cannot identify Ken Conley as being that officer. He did not know him, he had not seen him before, and, if you will note, he did not identify him in this courtroom as being that officer. But if you take that testimony in conjunction with [Conley’s] Grand Jury testimony that was read to you, you will see that Ken Conley, when he testified before the Grand Jury, said that during the pursuit, I dropped my radio and, a black uniformed officer picked it up and gave it to me. There is no doubt that Ken Conley was the tall white officer who apprehended Brown.
Tr. IV:51-52 (emphasis added). According to the government, whatever impeachment value the IAD Testimony might have in the abstract, it was never in Conley’s interest to impeach Walker’s ability to recall events at Woodruff Way.
Finally, the government argues that the core of Walker’s testimony was that he observed Cox running closely behind Brown, chasing him to the fence and reaching for him. This testimony was corroborated by both Cox and Brown and stood in stark contrast to Conley’s statement that he followed Brown to the fence and did not observe anyone chasing after Brown. Thus, although Walker’s IAD identifications may have sewn confusion about who was at the bottom of the hill, it was immaterial to the essence of Walker’s corroborated testimony.
(2) The FBI Memorandum
The second document in Conley’s Brady claim is an FBI memorandum concerning a proposed polygraph examination of Walker regarding inconsistent statements he made in his accounts of the assault. At Walker’s first interview with IAD on February 23, 1995 (properly disclosed to the defense), Walker stated three times that he saw someone he believed to be a police officer running behind Cox as Cox chased the suspect to the fence. However, in subsequent testimony (also disclosed), Walker made no mention of seeing anyone running behind Cox.
When the federal investigation began and the government reviewed Walker’s IAD interview on April 9, 1997, investigators met with Walker and questioned him about the discrepancy. After that interview, investigators prepared a memorandum requesting permission to polygraph Walker regarding the inconsistencies. Walker agreed to take a polygraph examination. However, an attorney representing Walker later notified the FBI that Walker would decline to take the polygraph.
The FBI Memorandum, which was not turned over to the defense, related: (1) Walker’s initial willingness to submit to a polygraph examination; (2) his present belief that he did not see anyone running behind Cox; (3) that when investigators confronted him about the inconsistency between his present belief and his prior statements, he explained that because of his friendship with Cox, he must have “convinced himself that he actually saw someone or something” when in fact he did not; and (4) that Walker “suggested that perhaps if he was hypnotised [sic] he might truly recall what was going on versus what he indicates was tunnel vision.”
*29Most of these facts ((l)-(S)) were thoroughly elicited in Walker’s testimony before the federal grand jury on June 5, 1997. There, Walker testified that he did not see anyone running behind Cox. During his grand jury testimony, Walker’s IAD interview was read back to him at length, and Walker was asked to explain the discrepancy between his grand jury testimony and his prior inconsistent statements. Walker acknowledged the inconsistency and explained his prior statements by asserting that he was probably motivated by a feeling of guilt “for not seeing more than [he] saw.”
Because Conley had access to the IAD interviews, as well as Walker’s grand jury testimony, the defense was in possession of nearly all the information contained in the FBI Memorandum. It already possessed substantial information regarding Walker’s inconsistent statements and his explanation of the discrepancy between those statements. The only information Conley would have gleaned from the disclosure of the FBI memorandum that he was not already in possession of was (1) that Walker consented to submit to a polygraph examination (which he later refused to take); and (2) the FBI’s notes regarding Walker’s cryptic statement that if hypnotized, he might be better able to explain the discrepancy between his inconsistent statements.
(3) The Foley Memorandum
Conley also alleges that a January 30, 1995 memorandum by Lieutenant Kevin Foley of the BPD Anti-Gang Unit is Brady material that must be considered in his challenge. The three-page memo, prepared less than a week after the Cox assault, summarizes facts that Foley believed justified launching an investigation into the assault.
According to Conley, the Foley Memorandum states that Officers Ryan and Teahan came upon Cox lying on a patch of ice as they were returning to their police cars- — not when they arrived on Woodruff way. This statement would support the defense view that it was actually Ryan and Teahan who Walker encountered at the bottom of the hill, not Conley.
The government argues that the Foley Memorandum proves very little. First, the government notes that nowhere in the report, written five days after the incident, does Foley indicate that he spoke with members of the Gang Unit, much less Teahan or Ryan in particular. Thus, whether Teahan and Ryan even gave a conflicting account of their actions, or to whom, is not shown by the memorandum. Second, because Officers Teahan and Ryan gave sworn testimony at trial and to previous grand juries contending that they drove down the cul-de-sac and spotted Cox lying on the ground, it is immaterial whether Lieutenant Foley believes they were actually returning to their vehicle.
(4) Walker’s Skeletal Report
The final piece of evidence in Conley’s Brady claim is Walker’s February 20,1995 report of events, which he was ordered to write by his commanding officer. According to Conley, the report is significant because it was prepared by Walker almost contemporaneously with the assault, yet it lacks any of the facts he recounted at trial. The report is indeed skeletal, containing only a few sentences of text explaining the events on Woodruff Way. Conley argues that its omissions could have been usefully exploited during cross examination.
The government contends that although the skeletal report does not contain any details, its usefulness for impeaching Walker would be greatly diminished by the fact that in the same memo in which Walk*30er was instructed to write the report, he was ordered to appear for an interview with IAD three days later. At that interview, Walker gave a highly detailed accounting of his observations.
III. Discussion
Assuming that the four pieces of Walker Evidence should have been disclosed under Brady, the withholding of evidence does not warrant reversal except upon a showing of prejudice, that is, unless there is a “reasonable probability that the suppressed evidence would have produced a different verdict.” Strickler, 527 U.S. at 281, 119 S.Ct. 1936. In this context, “reasonable probability” means a probability sufficient to undermine confidence in the outcome. See Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985).
Failure to provide impeachment evidence, if the evidence is powerful enough, can be prejudicial and grounds for a new trial. United States v. Martinez-Medina, 279 F.3d 105, 126 (1st Cir.2002). This is particularly true where the evidence is highly impeaching or when the witness’s testimony is uncorroborated and essential to the conviction. Id.; see Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972). On the other hand, impeachment evidence that is merely cumulative or collateral is insufficient to establish prejudice under Brady. See United States v. Dumas, 207 F.3d 11, 16 (1st Cir.2000).
Considered collectively, the Walker Evidence could possibly have been employed at trial to impeach Walker’s ability to accurately recall the events on Woodruff Way by showing: (1) that he was unable to identify the officers at the bottom of the hill; (2) that he made prior inconsistent statements about someone chasing behind Cox; (3) that perhaps if he was “hypnotized,” he might be able to recall events more clearly; and (4) that he prepared an early report which was bereft of any details of the events from the night of January 25, 1995. Additionally, Walker’s refusal to submit to a polygraph examination might, if admissible, have impeached his overall reliability.
The question is whether this evidence “can reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.” Kyles, 514 U.S. at 435, 115 S.Ct. 1555; United States v. Josleyn, 206 F.3d 144, 156 (1st Cir.2000). The undisclosed impeachment evidence is, for the most part, minor and cumulative. It is undisputed that Conley was in possession of Walker’s prior inconsistent statements regarding seeing someone chasing behind Cox.16 Moreover, Conley was fully aware of the fact that Walker was unable to identify the tall, white officer at the bottom of the hill. Brady does not protect a defendant who is aware of essential facts that would allow him to take' advantage of the exculpatory evidence at issue. See United States v. Hicks, 848 F.2d 1, 4 (1st Cir.1988); Coleman v. Mitchell, 268 F.3d 417, 438 (6th Cir.2001); United States v. Grossman, 843 F.2d 78, 85 (2d Cir.1988); Lugo v. Muñoz, 682 F.2d 7, 9-10 (1st Cir.1982). Setting aside information in the Walker Evidence that Conley already had knowledge of, Conley is only left with Walker’s cryptic hypnotism statement, his skeletal police report, and his refusal to submit to a polygraph exami*31nation.17
Such weak impeachment evidence is not sufficient to warrant the drastic remedy of a new trial. While Walker’s statement that if hypnotized, “he might truly recall what was going on,” does support the notion that he was uncertain about his memory of the events on Woodruff Way; this does not undercut Walker’s testimony (corroborated by Cox and Brown) that he saw Cox chasing Brown to the fence and that as Brown scaled the fence, Cox tried to reach for him. Moreover, in the face of Walker’s many detailed statements (one of which was given three days after the memorandum was prepared) the fact that his February 20, 1995 memorandum is skeletal is immaterial.
Since it was never in Conley’s interest to impeach the credibility of Richard Walker, the exculpatory value of the hypnotism statement and skeletal report is de minimus. During cross examination and closing arguments, the defense attempted to demonstrate circumstantially (through evidence of the dropped and recovered radio) that Conley was the officer Walker saw at the bottom of the hill. This is not a strategy that Conley was stuck with for lack of better evidence: Conley possessed — and elected not to use — ample impeachment evidence undercutting Walker’s testimony. Conley had evidence that: (1) Walker and Cox were friends; (2) that Walker felt guilty about not seeing more; (3) that this guilt led to Walker making prior inconsistent statements; and (4) that Walker believed that an “Officer Ryan” (rather than Conley) was the officer who arrested Brown. Instead of using this evidence, Conley chose to rely on Walker’s testimony to establish that Conley was down the hill, and therefore could not have observed the beating, and further, that like Walker, Conley had “tunnel vision” so he was not lying when he said he did not see Cox chasing Brown. Therefore, Conley’s argument that he would have used the Walker Evidence to undercut Walker’s testimony is unavailing.
IV. Conclusion
Conley’s Brady claim should not be remanded to the district court. Both parties have requested that this Court make the final Brady assessment, and the straightforward Brady materiality issue presented here — whether the existence of withheld Walker documents shakes confidence in the jury verdict — is highly amenable to resolution by this en banc Court.
I cannot conceive how any court could find that the nondisclosure of the Walker Evidence undermines confidence in the outcome of Conley’s trial. The impeachment evidence contained in the four documents was largely minor and cumulative; much of the information was already known to the defense. Finally, none of the new evidence undercuts Walker’s core testimony at trial — corroborated by Cox and Brown — that Walker observed Cox running to the fence right behind Brown. The majority’s decision today needlessly postpones the inevitable day when this Court or another finally concludes that there is no reasonable probability that, had the Walker Evidence been properly disclosed, the results of Conley’s trial would *32have been different. I respectfully dissent.
. The court stated that it made these findings on an "enlarged record,” suggesting that it was considering more or different evidence than it did in its earlier opinion. Conley IV, 164 F.Supp.2d at 222. Nowhere does the court explain what that evidence is.
. Under Wright, the moving party must demonstrate the following four elements: (1) the evidence claimed to be newly discovered was unknown or unavailable to the defendant at the time of trial; (2) failure to learn of the evidence was not attributable to lack of diligence by the defendant; (3) the evidence is material, and not merely cumulative or impeaching; and (4) it will probably result in an acquittal upon retrial of defendant. 625 F.2d at 1019; see also United States v. Gonzalez-Gonzalez, 258 F.3d 16, 20 (1st Cir.2001).
. In its opinion in Conley II, the district court discussed in some depth the testimony of Bullard and Brown, and made clear that its holding as to prejudice encompassed the entirety of that newly discovered and/or disclosed evidence. 103 F.Supp.2d at 51-54. The district court’s opinion in Conley IV, however, referenced only the Walker evidence as a basis for its allowance of Conley’s § 2255 motion. 164 F.Supp.2d at 221-22. Although it mentioned the Bullard and Brown testimony, it did so only in the context of "Background Criminal Proceedings.” Id. at 218. We read Conley IV, therefore, as premising its determination of a Wright violation solely on the Walker evidence.
. For a comprehensive summary of the evidence presented at trial, see United States v. Conley, 186 F.3d 7, 7-15, 19-21 (1st Cir.1999) ("Conley I").
.A word should be added regarding the specific testimony of Richard Walker (now the subject of Conley's Brady challenge). Walker testified that he observed Cox chasing closely behind Brown, saw Brown scale the fence, and then saw Cox trying to reach for him. According to Walker, once Brown made it over the fence successfully, Walker ran toward a hole in the fence. Walker then proceeded through the hole and down an incline, falling a couple of times. Once on the other side of the fence, he observed two white plainclothes police officers standing in the street. One was tall, about 6'2", and the other was shorter. Walker followed the tall officer in pursuit of the suspect. At some point during the pursuit, the taller officer dropped his radio. After Brown had been apprehended, Walker returned the radio to the tall officer. Walker did not attempt to identify the tall white officer at trial, but did testify that the officer was the same height and size as Conley.
. For this reason, I find the majority's comment that this evidence was “adequate for conviction, but was always circumstantial” an extraordinary one considering that the sufficiency of the evidence was long ago decided. Although I am sure unintended, the comment constitutes an inappropriate diminution of Conley I, which cannot but have an influence on future proceedings.
. Although Conley previously argued that the Bullard and Brown testimony were part of the new evidence justifying a new trial, because neither the Bullard nor the Brown testimony can reasonably be considered exculpatory, the testimonies’ materiality need not be assessed under Brady. Strickler, 527 U.S. at 281-82, 119 S.Ct. 1936. Additionally, at oral argument Conley’s attorney conceded that because the district court determined that the Bullard and Brown evidence was not exculpatory, it was not part of Conley’s current Brady challenge.
. Indeed, Conley not only possessed evidence of Walker’s prior inconsistent statements and chose not to impeach Walker with them at trial, the record indicates that the defense actually objected to the government’s attempt to introduce Walker’s prior inconsistent statements. See Tr. II at 51-52.
. Conley has not shown how Walker's refusal to submit to a polygraph examination would have been admissible at trial. See deVries v. St. Paul Fire and Marine Ins. Co., 716 F.2d 939, 945 (1st Cir.1983) (holding that evidence of a witnesses’s refusal to take a polygraph exam is inadmissible); Aetna Ins. Co. v. Barnett Bros. Inc., 289 F.2d 30, 34 (8th Cir.1961).