dissenting:
I respectfully dissent.
In Banks v. Dretke, 540 U.S. 668, 124 S.Ct. 1256, 157 L.Ed.2d 1166 (2004), the Supreme Court held that, as to the suppression of prosecution witness Robert Farr’s secret police-informant status and its bearing on the reliability of the death penalty verdict, Banks had satisfied all three elements of a Brady claim. The Court also held that Banks was entitled to a certifícate of appealability on the question of whether he adequately raised a second Brady claim based on the suppression of prosecution witness Charles Cook’s September 1980 interrogation transcript. As the majority opinion correctly sets forth more fully, on remand the District Court held that Banks satisfied all three elements of a Brady claim in respect to the Cook transcript suppression as it bears on the reliability of Banks’s capital murder verdict and conviction. See Banks v. Quarterman, 2008 WL 906716 (E.D.Tex. 2008). The Director appealed.
The majority opinion reverses the District Court’s well-considered judgment on the ground that the Cook Brady pre-trial interrogation transcript is not material. I disagree and would affirm the District Court’s judgment.
A proper application of the principles developed by the Supreme Court for deciding Brady claims leads to the conclusion that the Cook interrogation transcript is material because there is a reasonable probability that, had the suppressed information been disclosed to the defense, the result in the guilt phase trial would have been different. See Kyles v. Whitley, 514 U.S. 419, 434, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995) (“The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence. A ‘reasonable probability’ of a different result is accordingly shown when the government’s evidentiary suppression ‘undermines confidence in the outcome of the trial.’ ”) The majority opinion correctly recites these principles, but it reaches the wrong result because it does not properly and completely apply them to this case. Instead, the majority relies mainly on a sufficiency-of-evidence test and reasons that the evidence was sufficient to convict even without Cook’s testimony. In doing so, the majority ignores its own *333recitation that the Supreme Court has repeatedly held that the materiality standard is not a sufficiency-of-evidence test, which is incompatible with the Brady materiality inquiry. Further, the majority fails to evaluate materiality properly because it grossly underestimates the reasonably probable effects that disclosure of the Cook interrogation transcript would have had as an impeachment tool and as a safeguard against the prosecutorial misconduct that occurred in this case.
Finally, the majority’s collective evaluation of the state’s suppression of both prosecutorial misconduct and impeachment evidence is perfunctory and incomplete. The majority believes that the state’s cover-up of prosecution witness Farr’s secret agent status during the guilt trial is really not part of this case; thus, the majority considers it only in diluted form in an alternative fall-back argument. Consequently, the majority fails to see that the collective effects of the Cook and Farr perjuries and prosecutorial cover-ups significantly further undermine any confidence that Banks received a fair trial.1 Although I agree with the District Court that the prosecution’s concealment and cover up of the Cook transcript is material in itself, when the Cook and Farr guilt-phase prosecutorial misconduct is considered collectively, it is all the more evident that Banks has satisfied materiality and the other elements of his guilt-phase Brady claim. The majority opinion acknowledges that “materiality must be assessed collectively, not point by point[,]” maj. op. at 328 (citing Kyles, 514 U.S. at 436, 115 S.Ct. 1555), but it does not seriously take the suppressed Farr secret informant status evidence into account or consider its full, legitimate weight cumulatively with the Cook transcript suppression evidence.
1.
A brief outline of the critical guilt-phase trial evidence helps to explain why the suppression of the Cook interrogation transcript, either individually or together with the suppressed Farr secret paid police-informant status evidence, is material for Brady purposes: Banks was convicted of committing the capital murder of Richard Wayne Whitehead on April 12,1980, in Bowie County, Texas by shooting him with a gun in the course of committing robbery against him. To convict Banks of the charged crime, the State had to prove that Banks committed both the murder and the robbery in the same criminal transaction. The other responsive verdicts the jury could have rendered were: guilty of non-capital murder; guilty of aggravated assault; or not guilty. The State’s proof of the essential elements of capital murder was crucially dependent upon the credibility of two witnesses, Charles Cook and Robert Farr.
Cook testified that on April 12, 1980, Banks drove up in a Mustang car in front of Cook’s house in Dallas; Banks confessed privately to Cook that he had killed a “white boy” near Texarkana and had taken his car; Banks left the car and a pistol with Cook; and Banks caught the bus back to Texarkana. Cook testified that he disposed of the car and its contents, and sold the pistol to his neighbor, Bennie Lee Jones. Robert Farr testified that a week or so later he and Banks drove from Texarkana to Cook’s house in Dallas, where Cook gave Banks a .22 pistol because “his *334gun was in West Dallas.” Cook’s testimony as to Banks’s confession was the only direct evidence that Banks committed the murder of Richard Whitehead on April 12, 1980, and that he had killed the victim while robbing him of his car. Cook’s testimony was the only direct evidence that Banks had relinquished to Cook possession of the car and a pistol that later ballistics tests determined to be the .25 pistol murder weapon. Farr’s testimony provided the only cogent corroboration of Cook’s testimony that Banks had brought the murder weapon to Dallas and had left it with Cook. There was no direct physical evidence, such as fingerprints, hair, blood or other residue linking Banks to the murder weapon, the corpse or the crime scene. The State failed to introduce the Mustang car or any of its contents into evidence. Other witnesses’ testimony provided suspicion but not proof of capital murder against Banks by reporting that he was with Whitehead in the victim’s Mustang near Texarkana as late as midnight April 11, 1980, and that he was in Dallas with a similar Mustang on April 12, 13, and 14, 1980. The State ballistics expert opined that his tests indicated the .25 pistol had fired the shots that killed Whitehead, but he was not asked by either the prosecution or the defense to explain the reliability or margin of error of his methodology or results. The coroner testified that three bullets in or near the body caused the death, but was not asked and did not estimate the time or date of the death. Whitehead’s watch, necklace, and ring had not been removed, and there was no evidence that money or other valuables had been taken from his person.
The State prosecution and investigative officers suppressed evidence in respect to Farr (the “Farr Brady evidence”) and Cook (the “Cook Brady evidence”) before and during the guilt phase trial that would have been favorable in several respects to Banks. That evidence, which was ultimately uncovered in the federal habeas proceedings, was as follows:
The Farr Brady evidence: Deputy Willie Huff, who suspected Banks of the murder, hired Farr, whose wife was the sister of Banks’s girlfriend, to act as a paid secret informant to help locate the murder weapon. Farr, an illicit drug user, agreed to do so because he feared that otherwise Huff would charge him with drug crimes, and he needed money for drugs. Farr instigated the Dallas trip with Banks by making up a story that he needed Banks to help him find a gun to use in robbing a pharmacy for drugs. In the guilt phase trial, Farr repeatedly testified falsely that he had not taken money from police officers, had not been promised anything by them, and had not discussed his testimony or the case with them. Huff testified in the guilt phase that he followed Farr and Banks to Cook’s house in Dallas, but he complicitly did not reveal Farr’s secret paid-informant relationship with him or that Farr tipped him off about the trip. Further, because of Farr’s reputation among the police as a drug user, Huff may have testified falsely when he denied knowing that Farr was a “doper” or an illicit drug user. The prosecution knowingly let Farr’s perjury stand uncorrected and capitalized on its suppression by pointing to Farr’s candor about his drug usage to implicitly assure the jury that his entire testimony was truthful.2
*335The Cook Brady evidence: Cook was intensively coached and his testimony was thoroughly rehearsed by prosecutors and law enforcement officers, including the District Attorney and Deputy Huff, in at least one session a few days prior to Banks’s trial, which began on September 29, 1980. The undisclosed 38 page transcript of that session in the District Attorney’s possession would have allowed Banks to impeach and discredit Cook’s testimony. In it Cook admitted at one point that he had been told before his April 24, 1980 statement to police that Banks was wanted for the murder of a white male near Texarkana on April 12, 1980, and that they had suggested that if Cook failed to cooperate he could be charged as an accessory to that capital murder. Further, the transcript would have revealed that without the intensive coaching and rehearsal Cook’s testimony at trial would have been inconsistent in significant respects with his April 24, 1980 statement to police about Banks’s confession and possession of the .25 pistol.3 Most important, on cross-examination at trial, Cook perjured himself three times in testifying falsely that he had not discussed his testimony with anyone prior to trial and that the District Attorney had put him on the stand without knowing what he would say.4 The District Attorney, who conducted the direct and redirect examination of Cook, had a marked-up copy of the 38 page transcript in his hand during the trial, but he did not disclose the transcript of the September 1980 rehearsal session, or correct Cook’s perjury. Instead, the District Attorney handed the defense attorney only a copy of Cook’s April 24,1980 statement, which was much more consistent with Cook’s trial testimony than Cook’s September 1980 interrogation transcript. In closing argument, an assistant prosecutor told the jury that Cook’s testimony admitting to two prior felony convictions and prison sentences and a pending arson charge showed that his entire testimony was truthful. The prosecutor added that all of the State’s witnesses had been thoroughly tested and that there was no reason to doubt that Cook had told them the complete and absolute truth.5 The prosecutor did not reveal Cook’s perjury, the pre-trial *336transcript, the prior coaching and rehearsal session, or that in August 1980, Cook had been charged as a habitual offender exposing him to a life sentence without parole.
*335Now Charles Cook didn't hide anything from you, either. He came right up with it. Have you been convicted? Twice. Did you go on trial. No, I pled guilty. Why? Because I was guilty. He didn't hide a thing from you. Now if we had been picking witnesses for this case, we would have done a better job, but life is as we find it. We take our witnesses and we take the proof as it is given to us, as we find it and as we search for it. That’s the truth about this matter. Charles Cook brought you the absolute truth.
*3362.
In reviewing the same record now before us in this case, the Supreme Court in Banks v. Dretke, 540 U.S. 668, 124 S.Ct. 1256, 157 L.Ed.2d 1166 (2004), clearly implied, without definitively deciding, that there is a reasonable probability that the suppression of Farr’s informant status and Cook’s interrogation transcript adversely and materially impacted the reliability of Banks’s guilty-as-charged verdict and capital murder conviction. In the opening paragraphs of its opinion, for example, the Supreme Court stated:
Prior to trial, the State advised Banks’s attorney there would be no need to litigate discovery issues, representing: “[W]e will, without the necessity of motions!,] provide you with all discovery to which you are entitled.” Despite that undertaking, the State withheld evidence that would have allowed Banks to discredit two essential prosecution witnesses. The State did not disclose that one of those witnesses was a paid police informant, nor did it disclose a pretrial transcript revealing that the other witness’ trial testimony had been intensively coached by prosecutors and law enforcement officers.
Furthermore, the prosecution raised no red flag when the informant testified, untruthfully, that he never gave the police any statement and, indeed, had not talked to any police officer about the case until a few days before the trial. Instead of correcting the informant’s false statements, the prosecutor told the jury that the witness “ha[d] been open and honest with you in every way,” and that his testimony was of the “utmost significance!.]” Similarly, the prosecution allowed the other key witness to convey, untruthfully, that his testimony was entirely unrehearsed. Through direct appeal and state collateral review proceedings, the State continued to hold secret the key witnesses’ links to the police and allowed their false statements to stand uncorrected.
Ultimately, through discovery and an evidentiary hearing authorized in a federal habeas corpus proceeding, the long-suppressed evidence came to light. The District Court granted Banks relief from the death penalty, but the Court of Appeals reversed. In the latter court’s judgment, Banks had documented his claims of prosecutorial misconduct too late and in the wrong forum; therefore he did not qualify for federal-court relief. We reverse that judgment. When police or prosecutors conceal significant exculpatory or impeaching material in the State’s possession, it is ordinarily incumbent on the State to set the record straight.
Banks, 540 U.S. at 675-76, 124 S.Ct. 1256 (alterations in original) (internal citations omitted). These statements and others in the Supreme Court’s opinion clearly indicate its view that the suppressed evidence in this ease constitutes significant exculpatory or impeachment material.
3.
On remand, the District Court in its thorough opinion first held that the Cook transcript was favorable evidence “because it establishes that Cook was extensively coached just days before giving his testimony at trial. The Cook Transcript contains repeated instances where the investigators told Cook that his story was inconsistent and needed to be altered,” Banks, 2008 WL 906716 at *3, and was even more valuable as impeachment evi*337dence because “at ... trial, Cook expressly denied having been coached,” id. The district court also found that the transcript was “favorable because it provides information regarding Cook’s state of mind at the time he signed the April 1980 affidavits,” and “highlights a number of inconsistencies between Cook’s proposed trial testimony and his actual trial testimony.” Id.
The district court further found the transcript to be “material” under Brady for three reasons: “(1) Cook’s testimony was both uncorroborated and was central to the prosecution’s case at the guilt phase of the trial; (2) Cook misrepresented the fact that he had been coached, and the prosecution improperly relied upon that misrepresentation; and (3) Cook substantially altered his testimony in response to the coaching that he received just prior to trial.” Id. The district court explained:
Although there were several State witnesses who testified about different aspects of the crime, Cook was the only witness to testify that Petitioner confessed to murdering the victim, as well as the only witness to give a motive for Petitioner committing the crime, and, importantly, Cook’s testimony on these issues was uncorroborated. No other witness could verify the fact that Petitioner had confessed and no other witness could supply a motive for Petitioner’s commission of this crime.
Even more central to Petitioner’s Brady argument is the fact that Cook testified three times at Petitioner’s trial that he did not speak to anyone about this case prior to trial. The Cook Transcript, however, reveals that Cook did, in fact, have an extensive conversation with investigators just prior to his trial testimony .... The State not only allowed this erroneous testimony to stand uncorrected, but it also represented to the jurors during closing argument that Cook “didn’t budge from the truth” and that Cook “did not hide anything from you ... [he] brought you the absolute truth.”
Cook’s Transcript also demonstrates that Cook altered his testimony at trial in response to the extensive coaching he received during his September 1980 interview .... Had the defense been able to cross-examine Cook on the suppressed statement, the defense may have persuaded one or more jurors to reject Cook’s trial version of events. Several inconsistencies in Cook’s testimony came to light in the transcript, and while each inconsistency by itself might not have had much impact on the case, taken together, they had the potential to significantly affect the jurors’ impression of Cook. The transcript also demonstrates the interview tactics and demeanor of the investigators which may have swayed the jury into believing Cook’s testimony was coached, and ultimately, might have caused the jury to distrust the prosecution.
Id. at *4-6. The district court specifically based this last finding on two responses that the transcript reveals were altered as a result of the coaching session. First, when asked at trial, Cook said he had forgotten who the victim of his prior assault conviction was, but the transcript revealed that he in fact knew the victim was a school teacher. Thus, his response concealed the fact that he had assaulted a school teacher, undoubtedly a fact relevant to Cook’s credibility. Second, in his 1980 statement to law enforcement and at trial Cook stated that Banks admitted to killing the victim just “for the hell of it,” but said during the pre-trial coaching session that Banks admitted to killing the victim be*338cause he wanted the victim’s car. As the district court noted, this alteration allowed prosecutors to avoid questions about whether the victim’s car, which by all accounts was in poor mechanical condition, “would motivate someone to commit murder.” Id. at *5-6.
In conclusion, the district court wrote: The combination of the importance of Cook’s testimony to the case against Petitioner, the coaching by authorities revealed in the transcript, the otherwise unknown inconsistent statements, and the prosecution’s failure to correct false testimony lead the Court to conclude “that the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.”
Id. at *6 (quoting Kyles, 514 U.S. at 435, 115 S.Ct. 1555).
4.
The majority opinion acknowledges that Banks has established that the prosecution suppressed evidence favorable to Banks and that the ultimate question is whether that evidence is material. The majority concludes, however, that there is not a reasonable probability that its disclosure would have changed the result and that its suppression does not undermine confidence in the outcome. But the reasons that the majority assigns for its conclusions reveal that it has inadvertently fallen into error in at least three respects.
First, the majority’s main reason for its conclusion that the Cook transcript is not material is that, in its estimation, even if Cook’s testimony had been discredited, the evidence supporting Banks’s conviction is still “ample,” “abundant,” and “a great deal of evidence.” This, however, is plainly the application of a sufficiency-of-evidence test, and the Supreme Court has made clear that the Brady materiality inquiry “is not a sufficiency of evidence test.” Kyles, 514 U.S. at 434, 115 S.Ct. 1555. Thus, “[a] defendant need not demonstrate that after discounting the inculpatory evidence in light of the undisclosed evidence, there would not have been enough left to convict.” Id. at 434-35, 115 S.Ct. 1555. The majority’s addition of the modifiers “ample,” “abundant,” and “a great deal of evidence” do not change the nature of the majority’s test; it is still a sufficiency-of-evidence test that is incompatible with a Brady materiality analysis. Materiality and sufficiency are conceptually different and not congruent. “The possibility of an acquittal on a criminal charge does not imply an insufficient evidentiary basis to convict. One does not show a Brady violation by demonstrating that some of the inculpatory evidence should have been excluded, but by showing that the favorable evidence could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict.” Id. at 435, 115 S.Ct. 1555. If material evidence in Banks’s favor was withheld by the State, he was deprived of a fair trial regardless of whether the evidence introduced by the State was sufficient to convict.
Second, the Supreme Court found that the Cook transcript was “significant exculpatory or impeaching material,” Banks, 540 U.S. at 675, 124 S.Ct. 1256, that it “would have allowed Banks to discredit [Cook as an] essential prosecution witness,” id., that it showed that Cook’s “trial testimony had been intensively coached by prosecutors and law enforcement officers,” id., that it “revealed that the State’s representatives had closely rehearsed Cook’s testimony [and] [i]n particular, [that] the officials told Cook how to reconcile his testimony with affidavits to which he had earlier subscribed recounting Banks’s visits to Dallas,” id. at 685, 124 S.Ct. 1256, *339and that “it provided compelling evidence that Cook’s testimony had been tutored by Banks’s prosecutors,” id.
Without explaining specifically how the Supreme Court’s characterization of the Cook transcript was wrong and can be ignored, the majority reaches the contrary conclusion that the “extensive ‘coaching’ claimed by Banks is simply not present[;]” maj. op. at 325, and evidently concludes that the Cook transcript would not have been an effective defensive tool because “there is little, if any, impeachment value to a mere showing that Cook had met with prosecutors prior to trial[;]” maj. op. at 323, and, because Banks’s defense counsel was given Cook’s April 24, 1980 affidavits, the Cook transcript “would have simply allowed Banks’s counsel to add one more example to the Cook-lied-about-not-talking-to-others point that counsel was already making,” maj. op. at 323.
I respectfully disagree with the majority’s de minimis or extremely low appraisal of the potential impeachment value and fair-trial safeguard effect of the Cook transcript. See United States. v. Bagley, 473 U.S. 667, 676, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) (“Impeachment evidence, however, as well as exculpatory evidence, falls within the Brady rule. See Giglio v. United States, 405 U.S. 150, 154, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972). Such evidence is ‘evidence favorable to an accused,’ Brady, 373 U.S. at 87, 83 S.Ct. 1194, so that, if disclosed and used effectively, it may make the difference between conviction and acquittal. Cf. Napue v. Illinois, 360 U.S. 264, 269, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959) (‘The jury’s estimate of the truthfulness and reliability of a given witness may well be determinative of guilt or innocence, and it is upon such subtle factors as the possible interest of the witness in testifying falsely that a defendant’s life or liberty may depend’)”). The majority’s rejection of the Supreme Court’s evaluation of the same evidence in the context of this same case and record does not appear to be justifiable. Further, in terms of impeachment value and protection against perjury and prosecutorial misconduct, Cook’s two- and-a-half page April 24, 1980 Dallas County affidavit pales in comparison to the significance of the 38 page transcript of his September 1980 eve-of-trial interrogation and rehearsal by prosecutors and Deputy Huff. Also, the majority’s dismissal of the Cook transcript as not undermining confidence in the verdict is contrary to our own Circuit precedents. This court has frequently found suppressed evidence that undermined a key prosecution witness’s uncorroborated testimony on essential elements of the government’s ease to be material. Tassin v. Cain, 517 F.3d 770, 780-81 (5th Cir.2008) (suppressed evidence showing key prosecution witness had motive to lie was material); United States v. Sipe, 388 F.3d 471, 480, 488-90, 491-92 (5th Cir.2004) (suppressed evidence permitting impeachment of prosecution’s key witnesses was material); United States v. Fisher, 106 F.3d 622, 634-35 (5th Cir.1997) (holding that suppressed impeachment evidence “tending to discredit” government’s key witness was material), abrogated on other grounds by Ohler v. United States, 529 U.S. 753, 755, 120 S.Ct. 1851, 146 L.Ed.2d 826 (2000).
The majority’s reliance on Strickler v. Greene, 527 U.S. 263, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999), for its conclusion that the Cook transcript is not material or prejudicial is misplaced. The Supreme Court in Banks v. Dretke concluded that, in respect to prejudice or materiality, Striekler is clearly distinguishable from Banks’s case:
Regarding “prejudice,” the contrast between Striekler and Banks’s case is marked. The witness whose impeach*340ment was at issue in Strickler gave testimony that was in the main cumulative and hardly significant to one of the “two predicates for capital murder: [armed] robbery[.]” Other evidence in the record, the Court found, provided strong support for the conviction even if the witness’ testimony had been excluded entirely: Unlike the Banks prosecution, in Strickler, “considerable forensic and other physical evidence link[ed] [the defendant] to the crime” and supported the capital murder conviction. Most tellingly, the witness’ testimony in Strickler “did not relate to [[the issue under review,]] [the petitioner’s] eligibility for the death sentence”; it “was not relied upon by the prosecution at all during its closing argument at the penalty phase.”
Banks, 540 U.S. at 700-01, 124 S.Ct. 1256 (double bracketed material added; all other alterations in original) (internal citations omitted) (citing Strickler, 527 U.S. at 292-95, 119 S.Ct. 1936).
In contrast with Strickler, Cook’s testimony was the indispensable centerpiece of the prosecution’s guilt-phase case. Cook’s testimony that Banks had confessed to him that he alone committed both of the two elements of the capital murder charge, viz., murder during robbery, was the only direct evidence warranting a guilty-as-charged verdict rather than a lesser or not guilty verdict. Cook’s testimony that Banks had allowed Cook to take the .25 pistol from Banks, was the only direct evidence linking Banks to the purported murder weapon. The prosecution relied primarily on Cook’s testimony as to Banks’s confession and possession of the .25 pistol in its guilt-phase closing argument. Had Cook’s testimony been impeached or discredited, the prosecution would have had only inconclusive circumstantial evidence that Banks, acting alone, contemporaneously committed murder and robbery against Whitehead with the murder weapon. Moreover, the prosecution would have been unable to argue to the jury that Cook told the absolute truth that Banks had confessed to the particulars of the crime and had allowed him to peaceably take the murder weapon from Banks.
The majority, in part, approaches the Brady materiality issue as if it were merely a question of whether an attorney engaged in improper witness coaching in the context of typical lawyer trial preparation practice. The standard applicable here is not provided by ordinary law practice mores, however, but is one of constitutional dimension imposing on prosecutors a higher duty owed by the sovereign to the courts, the public and the accuseds. The Supreme Court in Banks v. Dretke soundly rejected the application of a lesser standard here, stating:
We have several times underscored the “special role played by the American prosecutor in the search for truth in criminal trials.” Strickler, 527 U.S., at 281[, 119 S.Ct. 1936]; accord Kyles, 514 U.S., at 439-440[, 115 S.Ct. 1555]; United States v. Bagley, 473 U.S. 667, 675, n. 6[, 105 S.Ct. 3375, 87 L.Ed.2d 481] (1985); Berger, 295 U.S., at 88[, 55 S.Ct. 629]. See also Olmstead v. United States, 277 U.S. 438, 484[, 48 S.Ct. 564, 72 L.Ed. 944] (1928) (Brandeis, J., dissenting). Courts, litigants, and juries properly anticipate that “obligations [to refrain from improper methods to secure a conviction] ... plainly restpng] upon the prosecuting attorney, will be faithfully observed.” Berger, 295 U.S., at 88[, 55 S.Ct. 629]. Prosecutors’ dishonest conduct or unwarranted concealment should attract no judicial approbation. See Kyles, 514 U.S., at 440[, 115 S.Ct. 1555] (“The prudence of the careful prosecutor should not ... be discouraged.”).
*341Banks, 540 U.S. at 696, 124 S.Ct. 1256 (alterations in original).
Thus, we are not called upon to decide merely whether the prosecutors overstepped the bounds of ordinary trial prep ethics but whether they concealed from Banks evidence favorable to him which his defense counsel could have used to ensure that he received a fair trial; whether they allowed prosecution witnesses to testify falsely without correction; whether they falsely vouched for the credibility of their witnesses in closing arguments; and, if so, whether that suppression of evidence and prosecutorial misconduct prevented Banks from receiving a fair trial by a jury of his peers.6
Our materiality inquiry must focus most intensely on whether there is a reasonable probability that the total conduct of the State’s prosecution team in respect to the Cook transcript prevented Banks’s lay person jurors from fairly judging the credibility of Cook’s testimony as a prosecution witness, and, if so, whether that undermines confidence in the outcome of the proceeding. Thus, it is not our function to arrive at a god’s eye view of what really transpired between Cook and Banks. That is within the prerogative of jurors. Nor is it our place to play oddsmakers by calculating whether Banks would more likely than not have received a different verdict with use of the Cook transcript; it is well settled that such a rule would place too heavy a burden on a defendant to show that the government’s withholding of evidence deprived him of a fair trial. Rather, we are called upon to decide whether the prosecution’s suppression of the Cook transcript and its consequences could reasonably be taken to put the whole case in such a different light as to undermine confidence in the guilty-as-charged capital murder verdict. See Kyles, 514 U.S. at 435, 115 S.Ct. 1555.
In doing so, we must take into consideration all of the effects that flowed from the prosecution’s suppression. The jury was ignorant of Cook’s three instances of perjury and may have been led to believe that his testimony had not been rehearsed and discussed with anyone, when in truth his testimony had been intensively coached, tutored and molded by prosecutors and police to fit the prosecution’s case. Banks’s defense counsel was deprived of the transcript of Cook’s interrogation as a means to vigorously cross-examine, impeach, and discredit Cook’s testimony. The jury and the defense counsel were falsely led to believe that Cook’s April 24, 1980 affidavit was his only recorded statement and the only version of his story. The prosecutor in closing argument was able to argue falsely and without fear of contradiction that Cook’s entire testimony *342was the absolute truth. All of these consequences that resulted from the prosecution’s suppression of Cook’s interrogation transcript combine to undermine confidence in the guilty-as-charged verdict. See Kyles, 514 U.S. at 453-54, 115 S.Ct. 1555 (considering all of the consequences of the State’s suppression and concluding that these effects deprived the defendant of a fair trial).
In the process of evaluating the Cook transcript for materiality, we must not usurp the function of the jury by attempting to quantify the precise magnitude and importance of each inconsistency or opening for impeachment; or by failing to view them collectively rather than individually in isolation; or by resolving the conflicts in the trial and suppressed impeachment material in favor of one party or the other. I respectfully disagree with the majority opinion’s methodology and conclusions that tend to preempt credibility questions that should be resolved by the jury. On one hand, the majority classifies many of the inconsistencies as “minor” and dismisses them as having no cumulative weight that could affect the jury’s judgment as to credibility; and, on the other, it broadly concludes, incorrectly I think, that the changing versions of Cook’s story are consistent enough as to offer no reasonably probable avenue for effective cross-examination and impeachment. On this score, I agree with the District Court’s conclusion: “Several inconsistencies in Cook’s testimony came to light in the transcript, and while each inconsistency by itself might not have had much impact on the case, taken together, they had the potential to significantly affect the jurors’ impression of Cook.” Banks v. Quarberman, 2008 WL 906716,*6 (E.D.Tex.2008) (citing Kyles, 514 U.S. at 449 n. 19, 115 S.Ct. 1555 (noting combined effect of suppressed evidence)). As this court has observed in similar situations before, “[i]t is possible to explain away the differences in these statements — and indeed the government expends considerable effort doing so in its briefs — but such argument is properly reserved for the jury.” Sipe, 388 F.3d at 482.
In sum, one can hardly be confident that Banks received a fair guilt phase trial, given the jury’s ignorance of Cook’s triple perjury giving the impression that his testimony was totally unrehearsed; the District Attorney’s knowingly allowing Cook’s triple perjury to stand uncorrected; the District Attorney’s artifice in offering Cook’s April 24, 1980 affidavit as Cook’s only recorded statement in the State’s possession; and the assistant prosecutor’s compounding of the State’s deception by arguing that Cook had brought the jury the absolute and complete truth in his testimony. See Banks, 540 U.S. at 703, 124 S.Ct. 1256 (quoting Kyles, 514 U.S. at 434, 115 S.Ct. 1555 (“The question is not whether the defendant would more likely than not have received a different verdict with the evidence, but whether in its absence he received a fair trial, understood as a trial resulting in a verdict worthy of confidence.”)). On the record before us, one could not plausibly deny the existence of the requisite “reasonable probability of a different result” had the suppressed Cook interrogation transcript been disclosed to the defense before or during the guilt phase trial, see id. at 703, 124 S.Ct. 1256; or that the suppression of the transcript and related prosecutorial misconduct undermines confidence in the guilt trial and its outcome, see Kyles, 514 U.S. at 434, 115 S.Ct. 1555. Accordingly, as to the suppression of the Cook interrogation transcript, its attendant prosecutorial misconduct and their bearing on the reliability of the jury’s verdict of guilty-as-charged to capital murder, all three elements of a Brady claim, including materiality, are satisfied.
*3435.
After mistakenly concluding that the Cook transcript suppression by itself is not material, the majority opinion further contends that we cannot consider the Farrrelated prosecutorial misconduct collectively with that related to Cook, because it is “already-disposed-of” or because Banks has somehow waived this argument. I must disagree. It is fundamental to the rule of Brady that the materiality of a failure to disclose favorable evidence “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). Relatedly, “the state’s obligation under Brady [ ] to disclose evidence favorable to the defense, turns on the cumulative effect of all such evidence suppressed by the government,” Kyles, 514 U.S at 421, 115 S.Ct. 1555; that is, materiality is defined in terms of suppressed evidence considered collectively, not just individually item by item, see id. at 436, 115 S.Ct. 1555;7 and “the prosecutor remains responsible for gauging that effect regardless of any failure by the police to bring favorable evidence to the prosecutor’s attention,” id. at 421, 437, 115 S.Ct. 1555. The majority presents no persuasive authority or argument for the proposition that Banks’s case is excepted from the Agurs and Kyles requirements.
Aternatively, however, the majority contends that, even if the Farr and Cook prosecutorial misconduct are considered collectively, they still do not undermine our confidence that Banks received a fair trial. I must continue to disagree.
Robert Farr was a key witness for the prosecution at the guilt as well as the penalty phase of Banks’s trial. Corroborating part of Cook’s testimony in the guilt phase, Farr testified to traveling from Texarkana to Dallas with Banks to retrieve a gun. On cross-examination, defense counsel questioned Farr and received the following answers:
Q. And have you ever taken any money from some police officers?
A. No.
Q. I see, and police officers promised you anything?
A. No, they have not.
Q. Well, Robert, did you ever give any police officers a statement?
A. No.
Q. You never did?
A. No.
Q. Who did you talk to? What police officers did you talk to about this?
A. I have talked to no one about this, outside of when they called us down referring to the case.
The evidence introduced during the federal habeas process has shown that this portion of Farr’s testimony at trial was false. In actuality, Farr, acting as Deputy Huffs secret paid informant, persuaded Banks to travel to Dallas to get a gun and tipped off Huff about their trip. Huff acknowledged in his evidentiary hearing testimony that Farr worked for him as a paid informant, that Farr — at Huffs request — “contacted Delma [Banks] to see what information he could find from him about this particular case,” and that Huff paid Farr $200 for his services. Farr further revealed:
*344I assumed that if I did not help [Huff] with his investigation of Delma that he would have me arrested for drug charges. That’s why I agreed to help [Huff]. I was afraid that if I didn’t help him, I would be arrested ....
Willie Huff asked me to help him find Delma’s gun. I told [Huff] that he would have to pay me money right away for my help on the case. I think altogether he gave me about $200.00 for helping him. He paid me some of the money before I set Delma up. He paid me the rest after Delma was arrested and charged with murder ....
In order to help Willie Huff, I had to set Delma up. I told Delma that I wanted to rob a pharmacy to get drugs and that I needed his gun to do it. I did not really plan to commit a robbery but I told Delma this so that he would give me his gun .... I convinced Delma to drive to Dallas with me to get the gun.
The prosecution knew or was charged with Deputy Huffs knowledge that Farr’s answers were false, yet allowed his testimony to stand uncorrected.
If Banks’s defense counsel had been apprised before or during trial of Farr’s secret paid informant arrangement with Deputy Huff, he would have been able to cross-examine and impeach the testimony of both Farr and Huff about Farr’s monetary and penal interest in helping Huff and the prosecution to convict Banks. The revelation that Farr worked as Huffs paid informant may have given the jury pause in crediting Farr’s testimony and caused it to question the good faith of the State’s entire investigation and prosecution.
The prosecution’s continued concealment of the secret Farr-Huff paid informant arrangement during the guilt-phase trial also enabled the prosecutor in closing argument to falsely portray Farr and Huff as having testified to the whole truth. In summation, the prosecution told the jury “[Deputy Willie Huff] brings us the truth and like I said, this is a search for truth.” Further, the prosecutor said, “The people that you saw come before you today and the day before were thoroughly tested and I ask you to search your memories and ... see if you can discern any reason why these people would falsify for vengeance or for advantage or to save their hides. I didn’t see any.” Emphasizing the importance and meaning of the Dallas trip instigated by Farr and Farr’s testimony, the prosecutor stated, “Now when Delma got in the car to go to Dallas, only Delma knew that was over there. Delma led the police to that residence and when Delma went up to Charles Cook’s door and knocked in the night and then walked away, what did he tell you? He spoke to you through Robert Farr and Marcus Jefferson ... He said, ‘They don’t have my gun.’ ” Finally, in vouching for the truthfulness and important corroborative effect of Farr’s testimony, the prosecutor argued, “I rather enjoyed Robert Farr’s [testimony], because Robert was pounded on ... And Robert came right back with ... the truth.” “Ask yourself if we rely on our conviction solely and strictly on Charles Cook, and the answer is no. Charles Cook didn’t lead us to Dallas. No, that’s not a coincidence. That’s a lot of good, hard, solid police work. That’s a lot of hours sitting and waiting for somebody that you suspect to make a wrong move.”
Because the Farr-related and Cook-related evidence suppressions and the total prosecutorial misconduct related to those suppressions, considered collectively, not just individually item by item, could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict, there is a reasonable probability that, had the suppressed evidence been disclosed to the defense, the *345result of the proceeding would have been different. Thus, Banks has satisfied the materiality and other elements of his guilt-phase Brady claim and should be entitled to a new trial on his capital murder indictment.
6.
Finally, the majority presents an assortment of unpersuasive arguments against this opinion. I respectfully address them as follows.
The majority argues that by highlighting the weaknesses in the State’s case against Banks, most notably the unexplained absence of the victim’s Mustang and the conclusory testimony of the State’s ballistics expert, I have somehow raised new claims that “in effect, seek to re-try[ ] this case.” See maj. op. at 312. Agurs, as discussed above, requires that the “materiality” inquiry consider the suppressed evidence in light of the entire record. See Agurs, 427 U.S. at 112, 96 S.Ct. 2392. The “touchstone of materiality is a ‘reasonable probability’ of a different result,” which is “shown when the government’s evidentiary suppression ‘undermines confidence in the outcome of the trial.’ ” See Kyles, 514 U.S. at 434, 115 S.Ct. 1555. To determine whether confidence in the verdict is undermined by the suppression, we must necessarily evaluate the strength or weakness of the State’s other evidence of guilt. See id. at 451-53, 115 S.Ct. 1555; see also Strickler, 527 U.S. at 293, 119 S.Ct. 1936.8
The majority next asserts that we should not consider the Supreme Court’s statements that the Cook transcript shows that Cook had been “intensively coached,” subjected to an “interrogation,” and that his “testimony had been tutored by Banks’s prosecutors”; and that the transcript “would have allowed Banks to discredit” Cook, an “essential prosecution witness.” Banks, 540 U.S. at 675, 685, 124 S.Ct. 1256. The majority argues that these statements are, in effect, only dicta and “not the law of the case with respect to the Cook-transcript Brady claim.” See maj. op. at 324.
First, even assuming arguendo that the these statements are merely dicta, we should only reluctantly disregard the Supreme Court’s assessment of the transcript and its impeachment value. See United States v. Becton, 632 F.2d 1294, 1296 n. 3 (5th Cir.1980) (“We are not bound by dicta, even of our own court .... Dicta of the Supreme Court are, of course, another matter.”); see also Schwab v. Crosby, 451 F.3d 1308, 1325 (11th Cir.2006) (explaining that “there is dicta and then there is dicta, and then there is Supreme Court dicta,” which is “not something to be lightly cast aside,” and collecting courts of appeal decisions expressing similar sentiments). And we should be especially hesitant to disregard the Supreme Court’s “dicta” discussing this very case.
Second, we must decide whether there is a reasonable probability that Banks could have used the Cook transcript to impeach *346Cook’s testimony and, as a result, cause the jury to discredit Cook’s testimony. Certainly, when a decisive majority of the Supreme Court describes the transcript as showing “intensive coaching” and “interrogation,” and as providing “compelling evidence that Cook’s testimony had been tutored by Banks’s prosecutors,” it is at least reasonably probable that the jury could have reached the same conclusion as to the transcript — thus causing it to discredit Cook’s testimony and, having discredited the State’s primary evidence against Banks, to reach a different outcome, including a lesser alternative verdict.
The majority disagrees that Cook and Farr provided crucial, uncorroborated testimony against Banks during the guilt phase of the trial. Certainly, marginal portions of Cook’s testimony were corroborated; for example, that Cook sold a pistol to his neighbor, or that Banks was seen driving a Mustang in Dallas on the weekend of April 12. The majority uses these partial corroborations, however, to obscure the fact that the most salient portions of Cook’s testimony were completely uncorroborated, including his testimony to Banks’s confession and linking of the murder weapon to Banks. In particular, the majority overlooks the importance of Cook’s testimony relating Banks’s confession to the jury. As the Supreme Court observed in Arizona v. Fulminante, 499 U.S. 279, 111 S.Ct. 1246, 113 L.Ed.2d 302 (1991), “A confession is like no other evidence. Indeed, ‘the defendant’s own confession is probably the most probative and damaging evidence that can be admitted against him’---- While some statements by a defendant may concern isolated aspects of the crime or may be incriminating only when linked to other evidence, a full confession in which the defendant discloses the motive for and means of the crime may tempt the jury to rely upon that evidence alone in reaching its decision.” Id. at 296, 111 S.Ct. 1246 (quoting Bruton v. United States, 391 U.S. 123, 139-40, 88 S.Ct. 1620, 20 L.Ed.2d 476 (1968) (White, J., dissenting)).
The majority further argues that Farr’s testimony during the guilt phase was of little importance because it was corroborated by the testimony of Marcus Jefferson, who accompanied Farr and Banks on the trip to Dallas. However, this argument overlooks Farr’s secret paid police informant role and the prosecutorial misconduct in covering it up. If the truth about Farr had been revealed at trial, it would have discredited not only Farr’s testimony but that of Jefferson and Deputy Huff, and perhaps the entire prosecution, as well. In any event, Jefferson was a 19 year old ineffectual witness who added nothing to Farr’s perjury-laden testimony.9 *347In closing argument, the prosecution expressly relied on and vouched for the truthfulness of Farr’s testimony while barely mentioning Marcus Jefferson’s.
In assessing the significance of the prosecution’s suppression of evidence favorable to Banks, it must of course be borne in mind that not every item of evidence introduced by the State would have been directly undercut if the truth about the Cook and Farr evidence and official misconduct had been disclosed. It is important, however, that the circumstantial evidence remaining unscathed would hardly have amounted to overwhelming proof that Banks was guilty of capital murder. The inconclusiveness of the physical and circumstantial evidence and the lack of confidence in the credibility of the testimony by Cook and Farr does not, of course, prove Banks’s innocence. Whether the unaffected evidence remaining after the impeachment of Cook’s and Farr’s testimony is sufficient to support a conviction of capital murder is debatable. But the issue is not whether the State would have had a case to go to the jury if it had forthrightly disclosed all evidence favorable to Banks. Rather, the question is whether one can have confidence in the jury’s verdict in light of the prosecution’s suppression of evidence that could have been used to impeach the credibility of two witnesses crucial to its case and of the prosecution’s further misconduct in covering up and capitalizing on that suppression of evidence. Because the prosecution knowingly relied on two perjurious witnesses to prove the core of its capital murder case, knowingly misrepresented to the jury that Cook and Farr had been vetted and found truthful by the State, and continued to hide and capitalize on that perjury and suppression of evidence favorable to Banks up until the federal habeas proceedings, “fairness” cannot be stretched to the point of calling this a fair trial.
The majority sincerely states that it does not condone the prosecution’s concealment of its witnesses’ perjury and other exculpatory evidence. But the state condoned all of this wrongful conduct in *348obtaining Banks’s conviction of capital murder through the violation of his constitutional rights. Thus, the state should be required to rectify this constitutional wrong by providing Banks a new, fair guilt phase trial.
For these reasons, I respectfully dissent.
. Prosecution witness Robert Farr testified in both the guilt and penalty phases of Banks's trial. Thus, the suppression of Farr’s secret paid police-informant status during the guilt phase must be evaluated collectively with the suppression of the Cook transcript, in the context of the entire record, for purposes of materiality. See Kyles, 514 U.S. at 421, 436-38, 115 S.Ct. 1555.
. The assistant prosecutor stated during closing argument: "I’ll be very honest. I don’t know if any testimony at all can be enjoyable in a proceeding like this, but I rather enjoyed Robert Farr's, because Robert was pounded on. You got tracks, don’t you, Robert? You use drugs, don’t you, Robert? And Robert came right back with the last thing in the world that the attorney asking him the ques*335tions expected, and that was the truth. He says, 'Yeah, I use drugs.' "
. The inconsistencies between Cook’s statements during the September 1980 interview, and his April 1980 statement and trial testimony related to: (1) when Cook first noticed blood on Banks’s pants, and when he gave Banks a change of clothes; (2) when Cook first noticed that Banks had a pistol; (3) Banks's motive for killing the victim; (4) when Cook disposed of the Mustang Banks left with him; (5) when Cook sold the pistol to his neighbor; and (6) the identity of the victim of Cook’s prior assault conviction.
. Cook testified:
Q. Who all have you talked to about this, Mr. Cook?
A. I haven’t talked to anyone about it.
Q. Haven’t talked to anybody?
A. No, sir.
Q. Mr. Raffaelli [the District Attorney] just put you on the stand, not knowing what you were going to testify to. Is that what you’re telling me?
A. That's what I’m telling you.
Q. Mr. Cook?
A. Yes, sir.
. The assistant prosecutor argued during the guilt-phase summation:
. As the Supreme Court noted in United States v. Ash, 413 U.S. 300, 93 S.Ct. 2568, 37 L.Ed.2d 619 (1973), counsel have for centuries "regularly interviewed witnesses before trial.” Id. at 318, 93 S.Ct. 2568. The historic nature of this practice, however, does not lessen the potential for prosecutorial misconduct. Nor does the routine use of pre-trial interviews excuse prosecutors from their ethical and constitutional obligations. "In many ways, the prosecutor, by accident or by design, may improperly subvert the trial. The primary safeguard against abuses of this kind is the ethical responsibility of the prosecutor, who as so often has been said, may ‘strike hard blows’ but not 'foul ones.' If that safeguard fails, review remains available under due process standards.” Id. at 320, 93 S.Ct. 2568 (internal citations omitted) (quoting Berger v. United States, 295 U.S. 78, 88, 55 S.Ct. 629, 79 L.Ed. 1314 (1935); Brady v. Maryland, 373 U.S. 83, 87-88, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963); citing Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972); Mooney v. Holohan, 294 U.S. 103, 55 S.Ct. 340, 79 L.Ed. 791 (1935); Miller v. Pate, 386 U.S. 1, 87 S.Ct. 785, 17 L.Ed.2d 690 (1967); Chambers v. Mississippi, 410 U.S. 284 (1973)).
. "We evaluate the tendency and force of the undisclosed evidence item by item; there is no other way. We evaluate its cumulative effect for purposes of materiality separately and at the end of the discussion[.]” Kyles, 514 U.S. at 437 n. 10, 115 S.Ct. 1555.
. The majority erroneously attempts to distinguish Banks’s case from Kyles and United States v. Sipe, 388 F.3d 471 (5th Cir.2004). However, Kyles and Sipe are not materially distinguishable: Only their procedural histories are different (resulting from the courts’ piecemeal resolution of Banks's various Brady claims), and those procedural distinctions are irrelevant here. This case, like Kyles and Sipe, involves multiple pieces of suppressed, favorable evidence and prosecutorial misconduct that, when considered cumulatively, are material for purposes of Brady. In other words, the unique procedural history of this case does not alter Kyles' requirement that we collectively consider all of the official misconduct in suppressing favorable evidence to determine whether the prosecution’s misconduct undermines confidence in the verdict.
. The majority’s attempt to paint Jefferson as a stalwart witness, see maj. op. at 329-30, is not supported by the record. Jefferson was a young 19 year old tag-along whose credibility would not have survived had Farr been unmasked at trial as Deputy Huff’s stooge. He could not remember the date or even the month that he went to Dallas with Banks and Farr. See Trial Tr. at 2086-87. He testified on direct examination:
Q. Okay, when he came back from the house, did he have anything with him?
A. No, I didn't see nothing with him.
Q. Okay, did Delma say anything to you or to Robert Farr at that time?
A. No.
Q. He did not make a statement to you at that time concerning the object that he had picked up?
A. No.
Q. Do you remember him making any statement to Robert while y'all were in the automobile?
A. Not while we were sitting still.
Q. Okay, after you were rolling, after you were moving, did he make a statement?
Trial Tr. at 2089-90. His testimony was interrupted at this point by defense counsel's request to approach the bench, and the district attorney subsequently requested to take *347up the matter outside the jury’s presence. Outside the presence of the jury, the district attorney attempted to confront Jefferson with a written statement, but the trial court sustained an objection to his reading it at that time. The district attorney then asked Jefferson a series of questions about what Banks said when they were driving off from the house. See Trial Tr. at 2092-93.
The jury was brought back into the courtroom and Jefferson then testified, in front of the jury, as follows:
Q. Did Delma Banks make any statement at that time to either you or to Robert Farr or just to the occupants of the car in general?
A. No.
* * *
Q. Did he state anything to you or to Robert Farr as you-all left the house?
A. Just that, what I told you [apparently referring to his examination outside of the jury’s presence].
Q. Could you tell us again, please, sir?
A. [T]hat Two-two — this Two-two dude gave him another gun. He didn’t have his.
Trial Tr. at 2095. Contrary to the majority’s description of the testimony, it was not at all clear from Jefferson's testimony to whom the indefinite pronoun “his” referred. Jefferson's testimony continued:
Q. Your honor, I'm not sure what his answer was, and I would ask him to say it a little louder and distincter [sic], so that I can understand you and the members of this jury can understand you, please, sir. What did he say when he came back from the door, as y'all were driving away?
A. He said that Two-two said that some girl had his gun, and so Two-two gave him another gun.
Trial Tr. at 2096.