OPINION OF THE COURT
FUENTES, Circuit Judge:Appellee-Petitioner Ernest Simmons was convicted in Pennsylvania state court of robbery and murder in the first degree, and subsequently sentenced to death. Both his direct appeal and petition for post-conviction relief were denied in state court, but the District Court granted Simmons’s habeas petition oh the ground that the state prosecutors had withheld several pieces of material exculpatory evidence in violation of Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963). Appellants, various Pennsylvania state officials (“the Commonwealth”), now challenge the grant of the petition. Because we agree with the District Court that the cumulative effect of the multiple Brady violations is to undermine confidence in the verdict, we will affirm.
I.
A. Evidence at Trial
The victim in this case, an eighty-year-old woman named Anna Knaze, was killed on the afternoon of May 5, 1992, at her home in Johnstown, Pennsylvania. Around 5 p.m. on May 6, Knaze’s son, Stephen Knaze, visited his mother’s house after receiving a phone call that her mail had not been retrieved from her mailbox. Upon approaching the house, he found the door ajar. When he entered he saw Knaze lying dead on the dining room floor. Stephen Knaze searched the house and noticed his mother’s favorite pocketbook was missing; it was never found.
An autopsy showed that Knaze had been strangled, suffered blunt trauma that severed her spine, and all twelve pairs of her ribs had been broken. The death was deemed a homicide, with the cause of death listed as manual strangulation causing asphyxia. Based on the level of rigor in her body, her time of death was estimated to have been between 10 a.m. and 2 p.m. on May 5,1992.
Simmons soon became a suspect, and was arrested by the police on May 14, 1992, for an unrelated parole violation. While incarcerated, he was charged with the murder and robbery of Knaze. He was tried on these charges in June 1993.
At trial, the prosecution established a timeline of Simmons’s activities on the day of the murder. That morning, Simmons drove his girlfriend, LaCherie Pletcher, along with Pletcher’s friend Kitty McKinney and McKinney’s mother, to various appointments in the Johnstown area. He first took Pletcher to the Cambria County Domestic Relations office, where she was having blood work done around 10:05 a.m. Next, at around 10:35 a.m., Simmons dropped McKinney and her mother off at a gas station in Conemaugh, near Johns-town, so McKinney could pick up her car. According to McKinney, although Simmons at that point told her he would drive *227back to pick up Pletcher, she saw him drive off in a different direction — toward the Woodvale section of Johnstown, where Knaze lived.
Pletcher’s appointment concluded around 10:30 a.m. She waited for Simmons, and tried but was unable to reach him at work. Simmons finally picked Pletcher up at 11:45 a.m. He then dropped her off at school in the early afternoon, and she testified that she did not see him again until early in the morning on May 6, when he showed up at her door drunk and smelling of alcohol. At trial, the defense attempted to undermine this account by pointing to the fact that in Pletcher’s initial statements, at a coroner’s inquest in August 1992, she said that Simmons came home around 10 p.m. on May 5 and stayed with her the rest of the night, and was not drunk. Pletcher asserted that her recollection had been refreshed as to the exact series of events by an intervening letter that Simmons sent her.
Meanwhile, between 11 and 11:30 a.m. on May 5, three witnesses saw Knaze outside of her house speaking to a black man wearing distinctive sunglasses with wide stems and large gold trim on the hinges.1 These witnesses were Thelma Blough, her son Gary Blough, and Tammy Ickes, who all lived in a house next door to Knaze. At trial, all three of these witnesses identified Simmons, who is black, as the man they had seen, and Simmons’s sunglasses as those the man had been wearing that morning. Thelma Blough and Ickes testified that they had heard Simmons ask Knaze if he could use her telephone because his car had broken down, at which point the two entered Knaze’s house. Thelma and Gary Blough, along with Ickes, stated that they never saw Simmons leave the house and Thelma said that, unusually, Knaze never turned on her lights that evening.
Simmons challenged the credibility of the Blough and Ickes identifications. Tammy Ickes had initially stated that she only had a quick look at the man speaking to Knaze, and she and Thelma Blough had been unable to identify Simmons from a mug book and an array of six photographs in May 1992. Only in June 1992 did Thelma Blough inform Detective Richard Rok, head of the Knaze investigation, that her son Gary could identify the man in question. Notably, this breakthrough came after Gary Blough had been in the same jail as Simmons, whom he knew had been charged with Knaze’s murder, while incarcerated for a parole violation. This was also the first time the Bloughs told the police that Gary Blough was at his mother’s home that day. Gary Blough explained at trial that he had not informed the police of his presence earlier because he knew they were looking for him for his parole violation. After Gary Blough’s identification, Tammy Ickes identified Simmons only after he was pointed out to her by Gary Blough when she visited him in prison. Thelma Blough identified Simmons at a preliminary hearing in September 1992 after she had seen his picture in news reports naming Simmons as a suspect in Cobaugh’s death.
Another witness, David Mack, testified that he saw Simmons on May 6 near the beauty shop where Simmons was employed. He noticed that Simmons had a bandage on the back of his left hand, al*228though he could not remember Simmons’s explanation of the injury.
The principal witness against Simmons was Margaret Cobaugh, a sixty-two-year old woman who lived in Johnstown. She testified that, on April 1, 1992, her purse, with her driver’s license inside, was stolen from the senior activity center where she worked. Sometime not long before that incident, Simmons had been to the center asking if he could volunteer there. Pletcher offered testimony that between April 1 and May 5, 1992, she found a driver’s license belonging to an elderly woman in Simmons’s wallet. Pletcher stated that the photo on that license matched the one on Cobaugh’s current driver’s license.
Cobaugh also testified that at 1 a.m. on May 6, 1992, while returning to her home (just a few blocks from Knaze’s house) from a neighbor’s house, where she had been providing help with a medical emergency, a man smelling strongly of alcohol grabbed her from behind and placed his hand over her mouth. The man was wearing distinctive sunglasses, even though it was nighttime. He told her that “if you open your mother fucking mouth, you’ll get the same thing Anna Knaze got,” and then attempted to rape her. At that time, no one had yet discovered Knaze’s body. Cobaugh reported the assault to the police later on the morning of May 6, after arriving at work at the senior activity center, describing her attacker as a tall, black man. According to Cobaugh, the attack was so traumatic that she repressed her memory of the experience. She did not mention the statement about Knaze or otherwise link the attack to the Knaze case until October 9, 1992, when she was first called in by Detective Rok for an interview about the April 1992 theft of her purse. She only identified Simmons in a six-person photo array in October 1992, after she had seen his picture in news reports portraying him as the man charged with Knaze’s murder. Cobaugh also later identified Simmons as her attacker in a lineup requested by Simmons’s counsel and confirmed that identification in her trial testimony.
Presented with this evidence, the jury convicted Simmons of robbery and murder. Simmons was sentenced to death by lethal injection. Both the verdict and sentence were affirmed on direct appeal by the Pennsylvania Supreme Court.
B. Brady Material
In Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), the Supreme Court set out the rule that “the suppression by the prosecution of evidence favorable to an accused upon request violates due process where the evidence is material either to guilt or to punishment. ...” Id. at 87, 83 S.Ct. 1194. Appellants do not dispute that the Commonwealth prosecutor failed to disclose four items of information that were favorable to Simmons’s case.
1. Electronic Surveillance by LaCherie Pletcher
In the course of the investigation of Knaze’s murder, Pletcher cooperated with the prosecution by allowing the police to record telephone and in-person conversations between her and Simmons that occurred after his arrest. The existence and contents of these electronic surveillance tapes were not disclosed to Simmons before trial. Only subsequently did Simmons discover that, after he was arrested, Detective Rok had told Pletcher that she was a suspect in Knaze’s murder, indicating that Knaze’s killer had been seen with someone resembling Pletcher, and threatened to put her in jail if she did not cooperate with investigators. Pletcher therefore allowed the police to record her *229conversations with Simmons, and at Rok’s behest she tried to elicit incriminating statements from Simmons. Simmons contends that if he had known of the pressure Rok placed on Pletcher and the subsequent surveillance, he could have used that information to impeach Pletcher’s credibility, particularly with respect to her changed statements as to when he arrived home the night of May 5.
2. Margaret Cobaugh’s Firearms Forms
Soon after the assault, Cobaugh attempted to buy a handgun for protection. In filling out the state and federal forms necessary for the purchase, she did not disclose, as required, that she had been convicted of felony burglary in 1951. That felony conviction would have rendered her ineligible to buy a gun. The forms were filled out subject to the penalties of perjury for any misstatements. Simmons argues he could have used Cobaugh’s false statements on these forms to impeach her testimony at trial.
Because of Cobaugh’s omission, she was charged with violating the Pennsylvania Uniform Firearms Act, 18 Pa. Cons.Stat. § 6105(a), which bars a convicted felon from purchasing a firearm. The charge was dropped, however, after Detective Rok and Assistant District Attorney (“ADA”) Patrick Kiniry, the prosecutor on the Knaze case, arranged for Cobaugh to voluntarily surrender her gun to the police in exchange for the dismissal of the charges against her. Meanwhile, Kiniry kept possession of the forms that Cobaugh had filled out falsely rather than forwarding them to the appropriate authorities as he had done in other cases of suspected perjury. Simmons also asserts that this evidence would have been useful to support the argument that, because of the investigators’ assistance, Cobaugh had a motive to offer testimony helpful to the Commonwealth at his trial.
3. Lab Reports Regarding Cobaugh Assault
After Cobaugh reported the assault that took place on May 6, 1992, the police collected forensic evidence including three human hairs that were found on two of Cobaugh’s nightgowns. Three lab reports indicated that the hairs did not match Simmons’s head or pubic hair, but two were similar to Cobaugh’s head hair while the third was consistent with her pubic hair standard. The reports also stated that no blood or seminal fluid was found on the clothing Cobaugh wore the night of the assault, possibly because she had washed her underwear afterward. These reports were not disclosed to Simmons before trial, and did not come to his attention until 1998. Although they are inconclusive, Simmons contends that these reports might still have raised doubts about Cobaugh’s testimony.
4. Cobaugh Mug Book Identification Attempt
At some point after Cobaugh was assaulted and Simmons was arrested as a possible suspect, Detective Rok showed her a mug book containing Simmons’s picture, but she was unable to identify him from the book. Rok later testified at a preliminary hearing that Cobaugh had never looked at a mug book. According to Simmons, if this failed identification had been disclosed, his trial counsel would not have requested the in-person lineup at which Cobaugh positively identified him because that request was an attempt to elicit a failed identification that could be used to impeach her credibility at trial. Simmons contends that his trial counsel could instead have used the mug book incident itself for that purpose.
*230C. Procedural History
1. The Court of Common Pleas Decision
On April 12, 1996, Simmons filed a petition for post-conviction relief pursuant to Pennsylvania’s Post-Conviction Relief Act (“PCRA”), 42 Pa. Cons.Stat. Ann. §§ 9541 et seq., in the Court of Common Pleas of Cambria County, Pennsylvania, Criminal Division (“the PCRA court”). Among his other claims, Simmons raised the Commonwealth’s alleged Brady violations as grounds for granting his petition. After holding a series of evidentiary hearings, the PCRA court rejected this line of argument, concluding that Pletcher had voluntarily agreed to cooperate in the electronic surveillance; that the gun charge evidence, even if disclosed, would not have altered the outcome of the trial because of the other evidence implicating Simmons in Knaze’s murder; that the lab report regarding forensic evidence of Cobaugh’s assault was inconclusive rather than exculpatory; and that even if Cobaugh’s testimony were impeached by evidence of her initial failure to identify Simmons from a mug book, it would not have been enough to overcome the other successful identifications by Cobaugh and other witnesses. The PCRA court therefore denied Simmons’s petition. Simmons appealed that decision to the Pennsylvania Supreme Court.
2. The Pennsylvania Supreme Court Decision
The Supreme Court affirmed the denial of Simmons’s PCRA petition in a notably fractured decision. Commonwealth v. Simmons, 569 Pa. 405, 804 A.2d 625 (2001). Chief Justice Flaherty, joined by Justice Newman, announced the judgment of the Court, rejecting three of the Brady claims as proeedurally deficient under the PCRA’s rule requiring “contextualization” in pleading,2 and the fourth Brady claim (regarding the failed mug book identification) as non-meritorious. Justices Castille and Nigro concurred in the result without writing separately. Justice Cappy concurred, also citing procedural concerns. Finally, Justices Saylor and Zappala dissented, stating that Simmons was entitled to a new trial because the prosecution had suppressed material evidence and its actions “create[d] a cloud upon the reliability of the verdict and judgment of sentence.” Id. at 643.
3. The District Court Decision
Simmons next filed a habeas petition before the District Court for the Western District of Pennsylvania, pursuant to the Antiterrorism and Effective Death Penalty Act (“AEDPA”), 28 U.S.C. § 2254. The District Court, rejecting the substantive ruling of the PCRA court, granted the petition based on Simmons’s Brady claims.3 Simmons v. Beard, 356 F.Supp.2d 548, 555 (W.D.Pa.2005). The Court found *231that the evidence regarding threats against Pletcher and the prosecution’s intercession to prevent Cobaugh from facing charges for purchasing a firearm would have provided the jury with a compelling basis for mistrusting the testimony of these two key witnesses. Id. at 563-64. In reaching this conclusion, the District Court held that the PCRA court’s factual finding that Pletcher had cooperated voluntarily was unreasonable in light of her testimony that she agreed to cooperate only to avoid prosecution. Id. at 564 n. 15. The District Court also reasoned that the lab reports might have further undermined the validity of Cobaugh’s account in the jury’s eyes. Id. at 565. Finally, the Court agreed that Simmons’s lack of knowledge of the failed mug book identification directly impacted his defense since his counsel, if provided with that evidence to attack Cobaugh’s credibility, would not have attempted the “risky” tactic of seeking an in-person lineup even after Cobaugh had seen Simmons’s picture in the news. Id. Considering these Brady violations collectively, the District Court concluded that “[a]l-though [I] cannot say with certainty that the jury would have reached a different conclusion on its verdict [absent the violations], Simmons has demonstrated a ‘reasonable probability’ that it would have done so.” Id. at 566 (citation omitted). The District Court ruled that Simmons’s other claims were not meritorious, but based on the Brady violations it granted habeas relief and ordered that Simmons be given a new trial. Id. at 576.
II.
The question of what standard of review applies to this petition is a complex one.4 We exercise plenary review over a district court decision on a habeas petition where, as here, the District Court did not hold an evidentiary hearing. Jacobs v. Horn, 395 F.3d 92, 99 (3d Cir.2005). We similarly review de novo the District Court’s legal conclusion as to whether AEDPA deference applies to this petition. Hackett v. Price, 381 F.3d 281, 287 (3d Cir.2004).
AEDPA provides that, where a habeas petitioner’s claim was “adjudicated on the merits” in state court, the petition may not be granted unless the state court decision “was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States” or “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d). Under the § 2254 standard, a district court is bound to presume that the state court’s factual findings are correct, with the burden on the petitioner to rebut those findings by clear and convincing evidence. Id. § 2254(e)(1).
Where the state court decision is not on the merits, § 2254(d) does not apply and instead a federal habeas court applies the pre-AEDPA standard, reviewing pure legal questions and mixed questions of law and fact de novo. Appel v. Horn, 250 F.3d 203, 210 (3d Cir.2001). However, § 2254(e)(1) still mandates that the state court’s factual determinations are presumed correct unless rebutted by clear and convincing evidence. Appel, 250 F.3d at 210 (citing 28 U.S.C. § 2254(e)(1)).
III.
A. Reviewing the PCRA Court Decision or the Pennsylvania Supreme Court Decision
In considering a § 2254 petition, we review the “last reasoned decision” of *232the state courts on the petitioner’s claims. Bond v. Beard, 539 F.3d 256, 289-90 (3d Cir.2008). The application of that approach is somewhat difficult here, where the different blocs of the Pennsylvania Supreme Court cited both procedural or substantive bases, or none at all, for their votes. The four-way split among the justices, with no ground for the case’s disposition receiving majority support, means that we cannot find a single rationale explaining the Supreme Court’s decision.
However, the fact that the result was supported by multiple lines of reasoning does not allow us to deem it “unreasoned” and look past it to the decision of the PCRA court. The policy of the United States Supreme Court is that “[wjhen a fragmented Court decides a case and no single rationale explaining the result enjoys the assent of five Justices, the holding of the Court may be viewed as that position taken by those Members who concurred in the judgments on the narrowest grounds.” Marks v. United States, 430 U.S. 188, 193, 97 S.Ct. 990, 51 L.Ed.2d 260 (1977) (internal quotation marks and citation omitted). The Pennsylvania Supreme Court, though taking a narrower view, has differentiated between the reasoning of a plurality decision, which is not binding authority, and the “conclusion” of such a decision, which is still “binding on the parties in that particular case.” In re Interest of O.A., 552 Pa. 666, 717 A.2d 490, 496 n. 4 (1998). Here, the Pennsylvania Supreme Court issued a judgment, with explanation, binding on the parties before it, and that is the decision the District Court properly reviewed. See Ylst v. Nunnemaker, 501 U.S. 797, 805, 111 S.Ct. 2590, 115 L.Ed.2d 706 (1991) (looking to the “last explained state-court judgment” to determine state courts’ basis for rejecting a habeas petitioner’s claims, ignoring only a “silent” disposition by a higher court). Cf. Holley v. Yarborough, 568 F.3d 1091, 1099 (9th Cir.2009) (reviewing a state appellate court decision in an AEDPA case where the state supreme court had denied the petition for review of that decision without comment); Joseph v. Coyle, 469 F.3d 441 (6th Cir.2006) (treating state supreme court decision as “last reasoned decision” as to claims that it explicitly decided, and state appeals court ruling as “last reasoned decision” as to claims the supreme court declined to address). Therefore, we look to the Pennsylvania Supreme Court decision in our review under AEDPA.
B. Decision on the Merits
A state court decision is an “adjudication on the merits,” reviewed under thé deferential standard of § 2254(d), where it is “a decision finally resolving the parties’ claims, with res judicata effect, that is based on the substance of the claim advanced, rather than on a procedural, or other, ground.” Rompilla v. Horn, 355 F.3d 233, 247 (3d Cir.2004), rev’d on other grounds sub nomine Rompilla v. Beard, 545 U.S. 374, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005) (internal quotation marks and citation omitted). Our next task is therefore to determine whether the Pennsylvania Supreme Court’s decision qualifies for § 2254(d) deference under this definition.
As to some of Simmons’s claims, namely the three Brady claims dismissed for lack of compliance with the PCRA’s procedural “contextualization” rule, it is clear that there was no adjudication on the merits. At least three of the five justices supporting the judgment, a plurality, did not reach the substance of these claims. As we discuss further below, we find we may consider the merits of those claims, and in doing so we will apply de novo *233review rather than § 2254(d) deference.5 Cf. Thomas v. Horn, 570 F.3d 105, 115 (3d Cir.2009) (finding no adjudication on the merits triggering § 2254(d) deference where Pennsylvania Supreme Court decided a defendant’s claims “on purely procedural, not substantive, grounds,” and in so doing “stripped the PCRA court’s substantive determination of [those] claims of preclusive effect”).
However, the fractured nature of the court’s decision as to the mug book identification claim makes reaching a conclusion as to whether that claim was adjudicated on the merits difficult. The opinion of the two justices announcing the judgment addressed the failed mug book identification on the merits, but the concurrence of Justice Cappy discussed only procedural concerns, and it is unclear why the two silent concurring justices supported the judgment. Thus, there does not appear to be a majority rationale regarding the mug book identification issue. However, the amount of deference afforded to this aspect of the Pennsylvania Supreme Court’s decision is irrelevant. As we will discuss further below, even if we assume the deferential standard of § 2254(d) applies to this claim, the District Court was correct in rejecting the state court’s overall conclusion based on its assessment of the cumulative effect of the Commonwealth’s Brady violations.
Finally, we do not defer to the Pennsylvania Supreme Court on the question of the cumulative materiality of the prosecution’s Brady violations because, finding only one of Simmons’s Brady claims not to be procedurally barred, the court did not reach the issue of the collective effect of multiple violations. Cf. Smith v. Sec’y, Dep’t of Corr., 572 F.3d 1327, 1348-49 (11th Cir.2009) (concluding that the state court “did not reasonably decide” the issue of cumulative materiality “because it did not decide this issue at all,” where the state court had erroneously found some alleged Brady evidence not to be favorable and thus had not included it in discussing materiality); see also Appel v. Horn, 250 F.3d 203, 210 (3d Cir.2001) (“[W]hen ... the state court has not reached the merits of a claim thereafter presented to a federal habeas court, the deferential standards provided by AEDPA ... do not apply.”). Nor can we “look through” the Pennsylvania Supreme Court opinion to the PCRA court’s assessment, as the PCRA court simply failed to consider the effect of the four Brady violations in combination. Therefore, we may review the District Court’s cumulative materiality analysis without reference to the state courts’ decisions.
C. Materiality of the State’s Brady Violations
“To establish a Brady violation, it must be shown that (1) evidence was suppressed; (2) the evidence was favorable to the defense; and (3) the evidence was material to guilt or punishment.” United States v. Risha, 445 F.3d 298, 303 (3d Cir.2006). Appellant does not dispute that the materiality prong is the only compo*234nent of the Brady inquiry at issue in this case, and thus that element is where we will focus our analysis. (See Appellant’s Br. 8-9 (conceding that the prosecution suppressed the four pieces of evidence discussed above).)
Evidence is “material” where “there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different. A ‘reasonable probability’ is a probability sufficient to undermine confidence in the outcome.” United States v. Bagley, 473 U.S. 667, 682, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985). In judging materiality in a case involving multiple Brady violations, the Supreme Court has mandated that the effect of the violations should be “considered collectively, not item by item.” Kyles v. Whitley, 514 U.S. 419, 436, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995).
Before considering the materiality of the evidence suppressed by the prosecution, we must first address the ruling of Chief Justice Flaherty and Justice Newman of the Pennsylvania Supreme Court that Simmons had procedurally defaulted on three of his four Brady claims. The District Court disregarded that portion of the opinion on two grounds: first, that the procedural default reasoning had garnered only two votes and thus did not represent a majority or even plurality view of the seven justices; and second, that the “contextualization” rule cited was not a bar to federal habeas review because it was not a “firmly established and regularly followed” rule at the time the Pennsylvania Supreme Court reached its decision. Simmons, 356 F.Supp.2d at 558.
The first of these rationales is somewhat debatable, given that Justice Gappy’s concurrence also relied on procedural deficiencies as grounds for denial of Simmons’s petition. However, the Commonwealth expressly conceded before the District Court that the “contextualization” rule on which the justices relied was a novel rule that had not previously been applied in the same manner. See Ford v. Georgia, 498 U.S. 411, 423-24, 111 S.Ct. 850, 112 L.Ed.2d 935 (1991) (explaining that a state procedural rule cannot bar review of a federal constitutional claim where it is not strictly or regularly followed and thus the defendant was not on proper notice of the rule when seeking review in state court). Furthermore, the Commonwealth did not pursue the procedural default issue in its opening brief, and we therefore deem it waived. See United States v. Pelullo, 399 F.3d 197, 222 (3d Cir.2005) (citing “the well-established rule that the failure to identify or argue an issue in an opening brief constitutes waiver of that issue on appeal”). Therefore, we consider all four of Simmons’s Brady claims on the merits. As noted above, we consider the first three Brady claims de novo as claims not adjudicated on the merits by the Pennsylvania Supreme Court, while according § 2254(d) deference to that court’s ruling on the materiality of the mug book claim.
We first review each individual Brady claim, and then discuss their collective effect on Simmons’s trial. See Kyles, 514 U.S. at 437 n. 10, 115 S.Ct. 1555. We conclude that, as the District Court aptly explained, the Brady evidence at issue here is material because it calls into question the credibility of the two witnesses at the heart of the case — Cobaugh and Pletcher. Had Simmons had access to the information suppressed by the prosecution, there is a reasonable probability that his trial would have had a different outcome.6
*2351. Electronic Surveillance by Pletcher
Had the jury known that Detective Rok pressured Pletcher into assenting to electronic surveillance of her conversations with Simmons by suggesting there was evidence tying her to Knaze’s murder, it might well have distrusted Pletcher’s testimony as another attempt to be helpful to the prosecution and avoid legal trouble herself. This evidence was particularly salient given that Pletcher initially stated that Simmons had been with her at the time of the assault on Cobaugh, and only altered her story after meeting with Detective Rok. When pressed about this inconsistency at trial, Pletcher attributed the change to having her memory refreshed by an intervening letter from Simmons. Undoubtedly the defense would have taken advantage of the opportunity to offer a competing explanation: that Pletcher only implicated Simmons in the Cobaugh assault because of a fear that otherwise she herself might face criminal liability.
In an analogous case, United States v. Scheer, 168 F.3d 445 (11th Cir.1999), the Eleventh Circuit found that the prosecution’s attempts to intimidate a “crucial prosecution witness” into testifying “correctly” by threatening to send him back to prison qualified as material information under Brady because it could have been used to impeach that witness at trial. Similarly, had the defense in this case had access to the information about Rok’s efforts to pressure Pletcher into cooperating with the prosecution, it would have been much better positioned to east doubt on her credibility.7 The Commonwealth asserts that Detective Rok’s threats to Pletcher were irrelevant because the jury would have understood that Simmons’s girlfriend would not testify against him absent some sort of threat or pressure and would have taken that into account in judging her credibility. To the contrary, the fact that Pletcher seemed willing to condemn her boyfriend without any outside incentive to do so may very well have reinforced the damning nature of her testimony. The information about the threats against Pletcher was essential to explain why she might testify against Simmons even if he was not in fact guilty.
2. Cobaugh’s Firearms Forms
Cobaugh’s misrepresentation of her criminal history in attempting to purchase a handgun, and the prosecution’s *236assistance in helping her to avoid liability for that misrepresentation, were also highly relevant to Simmons’s case.8 First, Cobaugh made a misstatement under penalty of perjury, a fact that itself could have been used to impeach her testimony. The D.C. Circuit has described such prior perjurious statements as an “infirmity ... that is almost unique in its detrimental effect on a witness’ credibility.” United States v. Cuffie, 80 F.3d 514, 517-18 (D.C.Cir.1996) (holding that undisclosed evidence of witness’s prior perjury was material, despite availability of other types of impeachment material).
Additionally, exposing Detective Rok’s and ADA Kiniry’s assistance to Cobaugh could have cast a new light on her October 1992 revelations regarding the statement made about Knaze during her assault and her ability to identify Simmons as her attacker. It is not clear from the record whether Rok and Kiniry found out about the gun charges and intervened in Cobaugh’s case before or after her October meeting with Rok; however, the chronology of events is irrelevant, since the defense could just as well have argued either that Cobaugh came up with the October statements with the intent of leveraging them for help from the investigators, or that she was pressured into making those statements by the investigators armed with knowledge of her predicament.
The Supreme Court has expressly stated that evidence regarding a witness’s arrangements with the prosecution regarding pending crimina,! charges may affect the witness’s credibility and thus may be material for Brady purposes. In Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972), the Court held that where the government’s case “depended almost entirely on” one witness’s testimony, the credibility of that witness “was therefore an important issue in the case, and evidence of any understanding or agreement as to a future prosecution would be relevant to his credibility and the jury was entitled to know of it.” Id. at 154-55, 92 S.Ct. 763. Similarly, the Ninth Circuit has concluded that knowledge of a leniency deal between the prosecution and a key witness was important Brady material because it constituted “powerful and unique impeachment evidence demonstrating that [the witness] had an interest in fabricating his testimony.” Horton v. Mayle, 408 F.3d 570, 579 (9th Cir.2005). The Commonwealth suggests that knowledge of Rok’s and Kiniry’s actions would have made no difference because the disclosure of similar efforts that the prosecution made on behalf of Gary Blough did not affect the verdict; however, that reasoning ignores the far more central role that Cobaugh played at trial compared to Blough, a role similar to that of the witnesses in Giglio and Horion. The prosecution’s case rested in no small part on Cobaugh’s credibility, and thus evidence relevant to her motives for testifying cannot be so easily dismissed as immaterial.
3. Lab Reports Regarding Cobaugh Assault
This is the weakest claim offered by Simmons, given that the lab reports did not eliminate him as the potential perpetrator of Cobaugh’s sexual assault. Cf. Gary v. Hall, 558 F.3d 1229, 1256 (11th *237Cir.2009) (holding that inconclusive bite mark exemplar was not material in rape case where rapist had bitten victim, given that other evidence did link defendant to crime). Such “negative” or “inconclusive” results, however, may be exculpatory even where they do not provide definitive evidence on a particular issue. As the Fourth Circuit explained in Patter v. Slayton, 503 F.2d 472 (4th Cir.1974), neutral forensic evidence “may, because of its neutrality, tend to be favorable to the accused. While it does not by any means establish his absence from the scene of the crime, it does demonstrate that a number of factors which could link the defendant to the crime do not.” Id. at 479. In line with this approach, the Pennsylvania Supreme Court itself categorized rape kit results showing a lack of semen and foreign pubic hair as relevant evidence in a rape trial; though the results were inconclusive as to whether intercourse occurred, they were consistent with the defendant’s testimony that he did not have sexual intercourse with the victim and thus might be probative of the defendant’s credibility. Commonwealth v. Hawk, 551 Pa. 71, 709 A.2d 373, 377 (1998). Therefore, we must at least consider the potential effect of this evidence on the jury’s verdict in combination with the other Brady material.
4. Cobaugh, Mug Book Identification Attempt
The Pennsylvania Supreme Court held that “foreclosure of the mere opportunity of impeaching Cobaugh’s identification at trial by informing the jury that she had failed to identify appellant from a mug-book does not begin to establish a reasonable probability that the result of the trial would have been different” given that Cobaugh was looking for one photograph among hundreds, later identified Simmons in a lineup, and “was one of multiple witnesses who identified” Simmons. Simmons, 804 A.2d at 637. Assuming arguendo that we review this claim under § 2254(d), see Part III.B, supra, we must defer to that conclusion unless it was “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States,” or “was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d).
Appellee does not call to our attention any Supreme Court precedent that would directly undermine the above reasoning. It does appear, however, that the Court’s assumption that the lineup would still have occurred had the failed mug book identification been disclosed was an unreasonable construction of the factual evidence presented to the PCRA court. A state court’s fact-finding may qualify as unreasonable where “the state court ... had before it, and apparently ignored,” evidence supporting the habeas petitioner’s claim. Miller-El v. Cockrell, 537 U.S. 322, 346, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003). Here, the Pennsylvania Supreme Court ignored the apparently undisputed contention that defense counsel would not have requested the lineup if it already had a failed identification that it could use to impeach Cobaugh’s identification.
The absence of a lineup identification would have left the defense at trial with a witness who told the police immediately after her assault that she did not get a good look at her attacker, identified Simmons from a photo array only after seeing his picture in the media as a suspect in Knaze’s murder, and then failed to pick him out of a mug book. That is a far different scenario than that sketched out by the Pennsylvania Supreme Court, of a failed mug book identification during a brief attempt versus a successful lineup *238identification. Consequently, we will give this Brady violation more weight than the Pennsylvania Supreme Court did in considering whether it contributed to Simmons’s conviction in combination with the other three violations.
5. Cumulative Effect of the Brady Violations
Both Cobaugh and Pletcher were critical witnesses at Simmons’s trial, providing the main testimony that tied him to Knaze’s murder. The other witnesses identifying Simmons as the man talking to Knaze on the day of her death, Thelma Blough, her son Gary Blough, and Tammy Ickes, only belatedly did so after each had already seen him either in person or in a photograph identified as the person charged with Knaze’s murder. The prosecution appears to have recognized Cobaugh’s central role in its case, beginning its opening statement with the threat that Simmons allegedly made against Cobaugh during her assault and referencing that threat several more times as the key to identifying Simmons as Knaze’s killer. The Commonwealth itself called Cobaugh a “critical” witness. (Supp.App.269.) Pletcher’s testimony was also pivotal, as her deviation from her original story that Simmons had been with her the night of May 5 deprived him of a potential alibi defense as to the attack on Cobaugh.
Yet Cobaugh’s and Pletcher’s testimony had certain flaws. Cobaugh had not asserted that she could identify her attacker, or even mentioned the attacker’s statement regarding Knaze, until after Knaze’s murder was publicized and Simmons was charged as her killer. Meanwhile, Pletcher had changed her account of Simmons’s whereabouts on the evening of Cobaugh’s assault, at first stating that he had been with her the entire night of May 5 and only later recollecting that he had actually come home late and drunk on that date.
Cobaugh’s testimony might have been further debunked by some of the Brady material suppressed by the prosecution. The defense could have cited the inconclusive lab reports as highlighting the lack of physical evidence tying Simmons to Cobaugh’s assault. Additionally, the knowledge that Cobaugh had been unable to identify Simmons from a mug book even after seeing his picture in the media might have made the jury more dubious of her accusation that Simmons was the man who threatened and tried to rape her, especially since in her initial report Cobaugh did not suggest that she would be able to identify her attacker.
Most importantly, the inconsistencies in both Cobaugh’s and Pletcher’s stories, revealed on cross-examination, meant that the jurors were well-primed to hear that these two witnesses had good reason to come up with testimony helpful to the prosecution. Without the evidence suppressed by the prosecution, however, the defense could not credibly proffer such a theory. Had this information been available to the defense before trial, it could have much more effectively attacked the Commonwealth’s case on not just one, but two critical fronts. Cf. Kyles, 514 U.S. at 444-45, 115 S.Ct. 1555 (holding that nondisclosure of evidence undermining eyewitness identifications of the defendant by what the state identified as its “two best witnesses” was material under Brady).
Overall, the picture of what Simmons’s trial would have been like had these four Brady violations not occurred is vastly different from what actually happened. The two key witnesses presented by the state would have been substantially less credible, thus undermining the main evidence implicating Simmons in Knaze’s death and Cobaugh’s assault. Therefore, we agree *239with the District Court and hold that, cumulatively, the Commonwealth’s Brady violations leave us without confidence in Simmons’s conviction.
IV.
For the foregoing reasons, we will affirm the District Court’s grant of Simmons’s § 2254 petition and remand to the District Court for further proceedings in accordance with this opinion.
. Patricia Gonda, who worked at a day care center across the street from Knaze’s house, also testified that she had seen a black man walking on the street with a blond white woman sometime during the week Knaze was killed, and identified Simmons as that man. However, she could not remember exactly which day she had seen the man, only that it was before Knaze’s body was discovered.
. As explained by the Pennsylvania Supreme Court in Simmons, this pleading rule requires a PCRA petitioner to, for each claim he presents, include "a statement as to what is required in order for that particular claim to be cognizable under the PCRA and argument that fully supports each of the required elements.” 804 A.2d at 632 n. 2; see id. at 631-32.
. After Simmons filed his petition, Cobaugh made a statement indicating that she had not been truthful in her identification of Simmons. See Simmons v. Beard, 356 F.Supp.2d 548, 555 (W.D.Pa.2005). Simmons sought to incorporate this evidence into his habeas petition, but the District Court ruled that he must first exhaust any claims based on new evidence in state court. Id. In order to avoid delay, Simmons chose not to pursue exhaustion and instead to continue with his initial Brady claims. Id. at 556.
. We have jurisdiction over this appeal under 28 U.S.C. §§ 1291 and 2253. The District Court had jurisdiction under 28 U.S.C. §§ 1331, 2241, and 2254.
. Appellant suggests we must also defer to the PCRA court's conclusions as to the individual materiality of each Brady violation as factual findings accorded a presumption of correctness under AEDPA § 2254(e)(1). Even if the Pennsylvania Supreme Court decision had not stripped the PCRA court opinion of its res judicata effect by resolving Simmons’s claims on different grounds, that would not be the correct approach; the issue of materiality is a "mixed question of law and fact” not subject to § 2254(e)(1). See Carter v. Rafferty, 826 F.2d 1299, 1306 (3d Cir.1987) (stating, preAEDPA, that "the materiality of evidence under Brady is a mixed question of law and fact”); Duncan v. Morton, 256 F.3d 189, 197 (3d Cir.2001) (noting that "a mixed question of law and fact ... [is] not subject to the presumption of correctness [under section 2254(e)(1)]”).
. The Commonwealth mainly attacks the District Court for what it claims was an im*235properly exclusive focus on the issue of "confidence in the verdict” rather than the "reasonable probability of a different result.” However, we cannot credit this argument because it is clear that these two portions of the materiality standard are in fact synonymous. As the Supreme Court explained in Kyles, "[a] ‘reasonable probability’ of a different result is ... shown when the government’s evidentiary suppression 'undermines confidence in the outcome of the trial.' ” 514 U.S. at 434, 115 S.Ct. 1555 (quoting United States v. Bagley, 473 U.S. 667, 678, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985)). The District Court clearly recognized that equivalence, alternatively phrasing its ruling as a conclusion that "the long-suppressed information would have affected the trial in such a way as to undermine the Court's confidence in the jury’s verdict” and that "Simmons has demonstrated a ‘reasonable probability' that” the jury would have reached a different conclusion had the Brady violations not occurred. Simmons, 356 F.Supp.2d at 550, 566.
. As the District Court noted, the PCRA court found that Pletcher’s cooperation had in fact been voluntary. Simmons, 356 F.Supp.2d at 564 n. 15. The District Court rejected this view on the ground that it was unreasonable in light of Pletcher’s testimony before the PCRA court that “she was afraid of prosecution and afraid of what would happened [sic] to her daughter if she did not cooperate.” Id. We agree with the assessment that Pletcher’s "willing” cooperation was in fact the product of Detective Rok's threats of prosecution.
. The PCRA court hypothesized that the defense would not have utilized this evidence, even had it been available, for fear of putting too much emphasis on the terrifying nature of the assault on Cobaugh. However, the nature of the attack was not at issue at Simmons’s trial; the identity of the attacker was. Certainly the defense might have chosen to run the risk of making the jury more sympathetic to Cobaugh if it could have, at the same time, effectively suggested that Simmons was not the person who attacked her.