dissenting.
This saga appears to be drawing to an unjust close. Eleven years ago we reviewed Abdur’Rahman’s ineffective assistance of counsel claim after the district court found it meritorious and granted Abdur’Rahman a new penalty-phase trial. See Abdur'Rahman v. Bell (Abdur'Rahman I), 999 F.Supp. 1073 (M.D.Tenn. 1998), aff'd in part and rev’d in part by Abdur'Rahman v. Bell (Abdur'Rahman II), 226 F.3d 696 (6th Cir.2000). I agreed with the district court’s conclusion then because I believed, as I still do now, that had Abdur’Rahman’s lawyer unearthed the breathtaking deprivations and serious mental impairments that shaped Abdur’Rahman and used those events and disabilities to paint a human portrait, at least one penalty-phase juror would have voted to spare his life.
My colleagues disagreed. While leaving undisturbed the district court’s finding that counsel performed deficiently, they undid its prejudice determination — even though the Warden had never quarreled with it. See Abdur'Rahman II, 226 F.3d at 708-09. Deploying the familiar logic of the double-edged sword, the majority obliterated in a few sentences the mitigating value of Abdur’Rahman’s horrific upbringing — the worst case of abuse the testifying psychologist had seen in twenty-five years of practice — and litany of psychotic disorders, with the spectre of his prior violent acts. Id. They reasoned vaguely that the evidence cut both ways because it “contained a description of Petitioner’s motive for killing a fellow prison inmate and a history of violent character traits.” Id. at 709. No matter that competent counsel would have persuasively framed these pri- or acts as symptomatic of Abdur’Rahman’s schizoid personality, and then integrated them into a nuanced depiction of Abdur’Rahman worthy of a juror’s mercy.
The majority’s prejudice analysis was wrong then and it has aged poorly. In Rompilla v. Beard, 545 U.S. 374, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005), a case where the petitioner had a similar history of prior violent acts and suffered from a comparable degree of abuse and psychological impairment, the Supreme Court held the pe*479titioner was prejudiced by the failure to present these forms of mitigating evidence, because a defense including such facts would have borne “no relation to the few naked pleas for mercy actually put before the jury.” Id. at 393, 125 S.Ct. 2456. And, in another case where the petitioner endured an atrocious upbringing and had a prior history of violence, a unanimous Court granted the writ after rejecting as objectively unreasonable the Florida Supreme Court’s prejudice determination. Porter v. McCollum, — U.S. -, 130 S.Ct. 447, 448, 175 L.Ed.2d 398 (2009) (per curiam). Indeed, the Court reached this conclusion even though Porter offered in mitigation that he was not “mentally healthy,” and Porter’s ex-wife testified that Porter had a good relationship with his son. Id. at 449. Still, the Court found prejudice under AEDPA because “[t]he judge and jury ... heard almost nothing that would humanize Porter or allow them to accurately gauge his moral culpability.” Id. at 454. Abdur’Rahman’s trial was constitutionally deficient for the same reasons. See Abdur’Rahman II, 226 F.3d at 719-24 (Cole, J., concurring in part and dissenting in part).
This would all be water under the bridge if the majority’s Brady analysis did not ironically presume that Abdur’Rahman’s counsel was competent. But the fiction that the defense had the time and aptitude to discover what the prosecution had a constitutional obligation to provide underpins the majority’s dismissal of the exculpatory evidence at issue in the instant petition. With respect to the Garafola report, the majority washes its hands of the prosecution’s deliberate withholding of this evidence by insisting that Abdur’Rahman’s counsel knew the fuzzy contours of the report and that through investigation he “should have discovered,” Maj. Op. at 476, the essential facts that it contained. For the same reason, the majority dismisses compelling evidence from Devalle Miller’s own mouth that he lied on the stand. See Maj. Op. at 474-76.
The majority’s conclusion that this evidence — proving that the government’s star witness lied under oath and depicting Abdur’Rahman’s mental deficiency — falls outside Brady’s scope confounds me. The rationale offered is tenuous. The proposition that Abdur’Rahman’s testimony could substitute for the impeachment evidence because it too could be used to contradict Miller, or that, with the magic of cross-examination, Abdur’Rahman’s counsel could have forced Miller to confess the truth, reflects a poor understanding of the mechanics of trial. The notion that Abdur’Rahman’s counsel should have discovered the facts contained in the Garafola report presumes a competence that, as mentioned above, our prior deficient performance holding belies. See Abdur’Rahman II, 226 F.3d at 707-09 (leaving undisturbed the district court’s deficient performance finding).
Still, these findings are of a piece with other significant decisions the majority has made in this case over the years: declining to find that Abdur’Rahman’s counsel’s performance prejudiced him, though the state never argued this point; refusing to cumulate all evidence of prosecutorial misconduct in the instant appeal, though the Certificate of Appealability (“COA”) permits it, the Supreme Court requires it, see Kyles v. Whitley, 514 U.S. 419, 436, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995), and all the constituent claims were exhausted; and forbidding Abdur’Rahman even to proceed on the instant claims pursuant to Federal Rule of Civil Procedure 60(b) after a remand from the Supreme Court. See Abdur’Rahman v. Bell, 493 F.3d 738 (6th Cir.2007) (vacated en banc Oct. 19, 2007). To be sure, the majority has put forth support for its positions, as I have for *480mine; but viewed at a distance a pattern emerges, and it reveals that the majority’s animating concern — even in this pre-AEDPA case — has not been to ensure that a conviction was had without constitutional error, but to efface in the name of federalism, finality, and comity any errors that were present.
Getting there is easier than you think. It merely requires a ceaseless commitment to privilege formalism over every other legal value; nowhere is that simpler to do than in the thicket of the Great Writ. If we chop claims into enough pieces and deal with each in a way that is perfectly abstracted from the reality of the death-penalty courtroom, all the errors vanish. The spell does break eventually, when someone looks hard enough to see past the sleight of hand. Whether the revelation will come to a person with the authority to spare Abdur’Rahman, and in time, I do not know.
I. Brady Claims
A fair look at the suppressed Brady evidence, in the context of the penalty-phase trial that actually took place, undermines confidence in the verdict and demands issuance of the writ. The central issue there was whether Abdur’Rahman’s life should be spared because he was mentally disturbed. His psychological instabilities explained why he was susceptible to the Southeastern Gospel Ministry’s quasi-religious and militaristic message and why he erupted into the uncontrolled violence that resulted in Daniels’s death. The prosecution rejected this view out of hand, calling it “bunk,” (Penalty-Phase Tr., App’x at 727), and insisting to the jury that Abdur’Rahman was not impaired in the slightest. Through the prosecutor’s lens, the jury saw Abdur’Rahman as a base and depraved killer, in control of his actions, and who had killed wantonly before, in 1972.
To be clear, the prosecutor is required to do battle forcefully. But there are limits. The Constitution forbade him from fixing the fight by withholding every scrap of evidence that undermined the state’s case or would have allowed the jury to see Abdur’Rahman’s actions in a more sympathetic light. The prosecutor knew that Abdur’Rahman had raised insanity as a defense to the 1972 killing, but rather than comply with his ethical and constitutional obligations and disclose the transcript of that proceeding to the defense, the prosecutor lied to defense counsel, telling him that no evidence mitigated Abdur’Rahman’s prior crime, and (the more pernicious invention) that it was committed in furtherance of a drug turf-war. The drug-turf-war fabrication devastated the defense, and the fallout entailed much more than the missed opportunity to present the suppressed evidence.
Stretched thin by a crushing caseload, defense counsel ran triage on Abdur’Rahman’s trial, see Mark Curriden, A Life in the Balance, A.B.A. J., Mar. 2011, at 47; his harried state and the rapport he felt with the prosecutor (they had opposed each other in several prior cases) explain why the prosecutor’s lies were so terribly successful. (See Zimmerman Post-Conviction Dep., App’x at 415-22 (admitting he aimed to prevent defense counsel from “getting into ... the 1972 murder”)); Curriden, supra, at 51. Making up a false motive for the prior crime that was consistent with the prosecution’s theory of the instant one (drug-related robbery) had two mutually reinforcing effects: it (1) bolstered the trustworthiness of the prosecutor’s core misrepresentation that no evidence mitigated the prior assault, and (2) left defense counsel with the intimidating impression that there might be something *481real in the prosecution’s view of Abdur’Rahman’s moral culpability. Each of these, working in tandem with defense counsel’s trust in the prosecution, allowed Abdur’Rahman’s unprepared and overworked defense counsel to feel the false security that proceeding with his ad hoc trial strategy and an underdeveloped record would not cause his client harm.
The lie inflicted damage at another level because the prosecutor parroted it to the professional mental health evaluators at the Middle Tennessee Mental Health Institute who analyzed Abdur’Rahman’s mental state. (See Zimmerman Letter to MTMHI, App’x at 268-69.) It is hard to see how the professionals there could correctly assess the health of Abdur’Rahman’s mind without knowledge of his prior psychosis. So this single deception created a cascade nullifying every legal and administrative safeguard meant to ensure that the existence of Abdur’Rahman’s mental deficits reached the jury. A verdict resulting from a falsehood this disruptive cannot command confidence.
The prosecution’s mayhem continued with the suppression of the Garafola Report; its depiction of Abdur’Rahman banging his head against every surface of the police interrogation room would leave any reader with the impression that he was seriously disturbed. Armed with this evidence, defense counsel not only would have put on a more persuasive mitigation case to the jury, he also would have received a signal — in neon lights — urging him to delve further into Abdur’Rahman’s background. And the report would have prodded defense counsel to seriously doubt the prosecutor’s representations regarding the 1972 assault, as it quotes Abdur’Rahman saying “I only killed one man in my life and that was because he was trying to fuck me.” (Garafola Report, App’x at 178-75.)
Had defense counsel then inquired into the veracity of that statement, he would have discovered that Abdur’Rahman was involved in a trio of coercive and violent sexual relationships with other inmates in 1972, (see Elmer Bishop Dep., App’x at 617-18), and that the man he stabbed in 1972, Michael Stein, was a sexual “predator” who “preyed on ... younger, weaker inmates[, like Abdur’Rahman,] for sex.” (Id. at 624.) Furthermore, when Stein’s mother sued the prison for wrongful death under the Federal Tort Claims Act, the government took the position, based on an FBI investigation, that Stein “was a member of a group of inmates who were attempting to apply extortionate pressures on [Abdur’Rahman] to submit to Stein’s demands for homosexual activities. The assault ... [on Stein] arose out of an attempted assault on [Abdur’Rahman] approximately two weeks earlier by members of this group.” (Id. at 627-28.) The absence of this evidence depicting an abused man lacking normal psychological brakes, and the compounding, down-the-line, effects of that omission on an overtaxed defense counsel, prejudiced Abdur’Rahman.
There is more. Testifying for the prosecution, Devalle Miller sold the lie that sent Abdur’Rahman down the river. Instead of telling the jury what Miller told the prosecutors, that he and Abdur’Rahman went to Daniels’s home to further the mission of the SEGM to stop drug dealing in the community, and that both he and Abdur’Rahman were given weapons for that purpose by the charismatic leaders of the group, William Beard and Allen Boyd, Miller recited a motive that dovetailed with the prosecution’s case for death. In this alternative reality, Abdur’Rahman was the intimidating figure who compelled Miller to go to Daniels’s house and rob him — end of story. Miller left out that he had lied to Beard about where he would go into hiding *482because Miller was afraid that Beard, Boyd and other SEGM leaders might kill him, a precaution difficult to square with the notion that Abdur’Rahman alone pulled the strings. From these facts the jury might readily have concluded that Abdur’Rahman was similarly intimidated by the SEGM leadership. Had defense counsel known that Miller had told the prosecution something entirely different, the defense would have nullified Miller’s testimony and provided more evidence that Abdur’Rahman was fertile ground for the SEGM’s misguided message.
Lastly, on top of secreting away exculpatory evidence, the prosecutor had the .gall to taint the jury by showing them indictments from Abdur’Rahman’s prior crimes, in direct contravention of the trial court’s order and the prosecutor’s agreement. Those indictments revealed more than the admissible fact of the prior conviction, they also showed a separate robbery charge which never yielded a conviction. Though the jury was instructed by the trial court to disregard this improper evidence, the prejudicial effect of the indictments could not so easily be undone. Thus, through means that the Tennessee Court of Appeals found “bordered on deception” and “improper,” the prosecution received yet another affirmation of its view of Abdur’Rahman’s character, one which tracked precisely the prosecution’s theory of the crime of conviction — a depraved murder/robbery with no mitigating qualities. State v. Jones, 789 S.W.2d 545, 552 (Tenn.1990).
How to make sense of these discrete but mutually-reinforcing acts of malfeasance? The Supreme Court has emphasized that Brady “omission[s] must be evaluated in the context of the entire record,” United States v. Agurs, 427 U.S. 97, 112, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976), and then, “collectively, not item by item.” Kyles, 514 U.S. at 436, 115 S.Ct. 1555. This is so because “[t]he proper standard of materiality must reflect our overriding concern with the justice of the finding of guilt.” Agurs, 427 U.S. at 112, 96 S.Ct. 2392. That focus requires us to consider whether withheld evidence would have “rebutted” the prosecution’s arguments, Cone v. Bell, 556 U.S. 449, 129 S.Ct. 1769, 1786, 173 L.Ed.2d 701 (2009), shown them to be “false and misleading,” id., or “len[t] support” to the petitioner’s case for life, id. at 1784. The evidence suppressed here would have had all those effects. Miller’s prior inconsistent testimony undercuts dramatically the persuasive value of the prosecution’s case for death by showing that the testimony of its key witness was false and misleading. See id. at 1783-86 (describing how the Brady evidence substantially enhanced the, case for life and diminished that for death). The transcript from Abdur’Rahman’s 1972 assault trial along with the Garafola Report lends support to the case for life by strengthening the inference that Abdur’Rahman was mentally disturbed. Showing the jury the prejudicial indictments improperly biased it against the defense. The sum of these parts invalidates the verdict. At least one juror could reasonably be predicted to see the case in a different light and vote for life after considering all the withheld evidence in mitigation and the detrimental effect that evidence would have had on the prosecution’s case for death. See id.; Kyles, 514 U.S. at 435, 115 S.Ct. 1555.
The majority’s refusal to conduct this cumulative Brady analysis with the claims on which the court denied a stand-alone COA (the 1972 transcript and the prejudicial indictments) has no support in the case law or the instant COA. Smith v. Secretary, Department of Corrections, 572 F.3d 1327, 1347 (11th Cir.2009), did hold, as the majority contends, that procedurally barred individual claims may not be cumu*483lated, but none of the claims Abdur’Rahman knits together was defaulted; all of them were raised and exhausted over the long course of this litigation. Moreover, the COA plainly contemplates a cumulative analysis of the prosecutorial-misconduct claims. (See Jan. 20, 2010 Order at 5 (“Abdur’Rahman is permitted to make cumulative-effect arguments with respect to the subclaims on which we grant him a COA, even if it involves referring to factual allegations that underpin prosecutorial misconduct subclaims on which we have denied his COA request.”).) Finally, the Warden waived any challenge to the inclusion of that material in the cumulative effect calculus. There is no support in the law or the record for the majority’s dodge.
II. Brady/Strickland Claim
Abdur’Rahman mounts a final attack by banding together the Brady violations with his long-settled Strickland claims. See Abdur’Rahman II, 226 F.3d at 707-09. I agree with Abdur’Rahman that the COA creates no jurisdictional bar to our review of this hybrid claim because the COA allows Abdur’Rahman to make cumulative effect arguments related to his Brady sub-claims. (See Jan. 20, 2010 Order at 5 (“Abdur’Rahman is permitted to make cumulative-effect arguments with respect to the subclaims on which we grant him a COA.”).) The clause the majority reads to limit the scope of those cumulative claims (“even if [bringing cumulative claims] involves referring to factual allegations that underpin prosecutorial misconduct sub-claims on which we have denied his COA request”) is properly read to provide an example of one kind of cumulative effect claim, not to limit the permutations of permissible Brady hybrids. (Id.) Even so, I am constrained to agree with the majority that Abdur’Rahman procedurally defaulted the Strickland/Brady claim by failing to raise it in state court. See Keith v. Mitchell, 455 F.3d 662, 679 (6th Cir.2006) (citing Lorraine v. Coyle, 291 F.3d 416, 447 (6th Cir.2002)).
Were I able to reach this last claim, I would grant it for the reasons detailed above. The Brady violations and Strickland ineffective assistance fed off each other at trial in a perverse symbiosis that infected the verdict with constitutional error. Perhaps if Abdur’Rahman could have pursued his petition in another circuit his life might be spared in this procedural posture. See, e.g., Derden v. McNeel, 978 F.2d 1453, 1456-57 (5th Cir.1992) (permitting a hybrid cumulative error argument where none was raised below). But I am powerless against our precedent and my colleagues’ contrary views.
III. Conclusion
A parting thought. Whatever your take on the merits of Abdur’Rahman’s claims, one thing about this case is undeniable: the prosecutor desecrated his noble role. He failed grossly in his duty to act as “the representative ... of a sovereignty ... whose interest ... in a criminal prosecution is not that it shall win a case, but that justice shall be done.” Berger v. United States, 295 U.S. 78, 88, 55 S.Ct. 629, 79 L.Ed. 1314 (1935). Abdur’Rahman may face the ultimate penalty as a result; Justice will bear a scar.
I dissent.