Turner v. United States, 582 U.S. ___ (2017)

NOTICE: This opinion is subject to formal revision before publication in the preliminary print of the United States Reports. Readers are requested to notify the Reporter of Decisions, Supreme Court of the United States, Washington, D. C. 20543, of any typographical or other formal errors, in order that corrections may be made before the preliminary print goes to press.

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 15–1503 and 15–1504

_________________

CHARLES S. TURNER, et al., PETITIONERS

15–1503v.

UNITED STATES

RUSSELL L. OVERTON, PETITIONER

15–1504v.

UNITED STATES

on writs of certiorari to the district of columbia court of appeals

[June 22, 2017]

Justice Breyer delivered the opinion of the Court.

In Brady v. Maryland, 373 U. S. 83 (1963) , this Court held that the government violates the Constitution’s Due Process Clause “if it withholds evidence that is favorable to the defense and material to the defendant’s guilt or punishment.” Smith v. Cain, 565 U. S. 73, 75 (2012) (emphasis added) (summarizing Brady holding). In 1985 the seven petitioners in these cases were tried together in the Superior Court for the District of Columbia for the kidnaping, armed robbery, and murder of Catherine Fuller. Long after petitioners’ convictions became final, it emerged that the Government possessed certain evidence that it failed to disclose to the defense. The only question before us here is whether that withheld evidence was “material” under Brady. The D. C. Superior Court, after a 16-day evidentiary hearing, determined that the withheld evidence was not material. Catlett v. United States, Crim. No. 8617–FEL–84 etc. (Aug. 6, 2012), App. to Pet. for Cert. in No. 15–1503, pp. 84a, n. 4, 81a–131a. The D. C. Court of Appeals reviewed the record, reached the same conclusion, and affirmed the Superior Court. 116 A. 3d 894 (2015). After reviewing the record, we reach the same conclusion as did the lower courts.

I

In these fact-intensive cases, we set out here only a basic description of the record facts along with our reasons for reaching our conclusion. We refer those who wish more detail to the opinions of the lower courts. App. to Pet. for Cert. in No. 15–1503, at 81a–131a; 116 A. 3d 894.

A

The Trial

On March 22, 1985, a grand jury indicted the seven petitioners—Timothy Catlett, Russell Overton, Levy Rouse, Kelvin Smith, Charles Turner, Christopher Turner, and Clifton Yarborough—and several others for the kidnaping, robbery, and murder of Catherine Fuller. The evidence produced at their joint trial showed that on October 1, 1984, at around 4:30 p.m., Catherine Fuller left her home to go shopping. At around 6 p.m., William Freeman, a street vendor, found Fuller’s body inside an alley garage between Eighth and Ninth Street N. E., just a few blocks from Fuller’s home. See Appendix, infra (showing a map of the area in which the murder was committed). Fuller had been robbed, severely beaten, and sodomized with an object that caused extensive internal injuries.

The Government advanced the theory at trial that Fuller had been attacked in the alley by a large group of individuals, including petitioners; codefendants Steve Webb, Alfonso Harris, and Felicia Ruffin; as well as by Calvin Alston and Harry Bennett. The Government’s evidentiary centerpiece consisted of testimony by Alston and Bennett, who confessed to participating in the offense and who cooperated with the Government in return for leniency. Although the testimony of Alston and Bennett diverged on minor details, it was consistent in stating that, and describing how, Fuller was attacked by a siz-able group of individuals, including petitioners and they themselves.

Alston testified that at about 4:10 p.m. on the day of the murder, he arrived in a park located on H Street between Eighth and Ninth Streets. He said he found a group of people gathered there. It included petitioners Levy Rouse, Russell Overton, Christopher Turner, Charles Turner, Kelvin Smith, Clifton Yarborough, and Timothy Catlett, as well as several codefendants and others. Those in the group were talking and singing while Catlett was banging out a beat. Alston suggested “getting paid” by robbing someone. App. A467. Catlett, Overton, Rouse, Smith, Charles Turner, Christopher Turner, Yarborough, and several others agreed. Alston pointed at Catherine Fuller, who was walking on the other side of H Street near the corner of H and Eighth Streets. Those in the group said they were “game for getting paid.” Id., at A471–A472. Alston, Rouse, Yarborough, and Charles Turner crossed H Street moving toward Eighth Street and followed Fuller down Eighth Street. The rest of the group crossed H Street and moved toward Ninth Street. When Alston’s group approached Fuller, Charles Turner shoved her into an alley that runs between Eighth and Ninth Streets. Charles Turner, Rouse, and Alston began punching Fuller. They were soon joined by Christopher Turner, Smith, and others. All of them continued to hit and kick Fuller until she fell to the ground. Rouse and Charles Turner then carried Fuller to the center of the alley and dropped her in front of a garage located at the point where the alley joins another, perpendicular alley that runs toward I Street. Someone dragged Fuller into the garage. Alston, Rouse, Charles Turner, Overton, Yarborough, and Catlett followed. Others stood outside. Members of the group tore Fuller’s clothes off and struggled over her change purse. Overton and Charles Turner then held Fuller’s legs, and Alston, Catlett, Harris, and Yarborough stood around her while Rouse sodomized her with a foot-long pipe. Shortly after, the group dispersed and left the alley.

Harry Bennett’s testimony was similar. Bennett also described a group attack. He said that he had gone to the H Street park, where he saw Rouse, Overton, Christopher Turner, Smith, Catlett, and others gathered. Alston was talking to the group about “[g]etting paid” and said “let’s go get that lady.” Id., at A368–A370. At that point Alston, Rouse, Overton, and Webb crossed H Street and approached Fuller, while Catlett, Christopher Turner, Charles Turner, and Harris followed in a separate group. Bennett added that he himself went to the corner of Eighth and H Streets to watch for police. He then went into the alley and joined the group in kicking and beating Fuller. He testified that at least 12 people were there, with some beating Fuller and others watching or picking up her jewelry. Overton then dragged Fuller into the garage, and Bennett, Rouse, Christopher Turner, Charles Turner, Catlett, Smith, Harris, and Webb followed, as did some “girls.” Id., at A402–A405. Alston and Steve Webb held Fuller’s legs, and Rouse sodomized her with a pole. The group then dispersed from the garage and alley.

The Government presented several other witnesses who corroborated aspects of Alston’s and Bennett’s testimony, including the fact that Fuller was attacked by a group. Melvin Montgomery testified that he was in the H Street park on the afternoon of the murder. He saw Overton, Catlett, Rouse, Charles Turner, and others gathered there. The group was being noisy and singing a song about needing money. Somebody then said they were “going to get that one,” and Montgomery saw that Overton was pointing to a woman standing on the corner of Eighth Street. Id., at 77–79. Overton, Catlett, Rouse, Charles Turner, and others crossed H Street. Some headed toward Eighth Street while others went toward Ninth Street. Montgomery did not follow them.

Maurice Thomas, then 14 years old, testified that he witnessed the attack itself. Thomas lived in the neighborhood and knew many of the defendants. As he was walking home, he glanced down the Eighth Street alley and saw a group surrounding Fuller. Thomas saw Catlett pat Fuller down and then hit her. He then saw everyone in the group join in hitting her. Thomas said he knew Catlett, Yarborough, Rouse, Charles Turner, Christopher Turner, and Smith and recognized them in the group. Thomas heard Fuller calling for help. He ran home where he found his aunt, who told him not to tell anyone what he saw. Later that day, Thomas saw Catlett at a corner store, and heard Catlett say to someone that they “had to kill her” because “she spotted someone he was with.” Id., at 127–128.

On the afternoon of the murder, Carrie Eleby and Linda Jacobs were looking for petitioner Smith, who was Eleby’s boyfriend, near the corner of H and Eighth Streets. They heard screams coming from where a “gang of boys” was beating somebody near the garage in the alley. Id., at A539–A541. Eleby and Jacobs approached the group. Eleby recognized Christopher Turner, Smith, Catlett, Rouse, Overton, Alston, and Webb kicking Fuller while Yarborough stood nearby. Both Eleby and Jacobs testified that they saw Rouse sodomize Fuller with a pole. Eleby added that Overton held Fuller’s legs.

Finally, the Government played a videotape of a recorded statement that Yarborough, one of the petitioners, had given to detectives on December 9, 1984, approximately two months after the murder. Names were redacted. The video shows Yarborough describing in detail how he was part of a large group that forced Fuller into the alley, jointly robbed and assaulted her, and dragged her into the garage.

None of the defendants testified, nor did any of them try, through witnesses or other evidence, to rebut the prosecution witnesses’ claim that Fuller was killed in a group attack. Rather, each petitioner pursued what was essentially a “not me, maybe them” defense, namely, that he was not part of the group that attacked Fuller. Each tried to establish this defense by impeaching witnesses who had placed that particular petitioner at the scene. Some, for example, provided evidence that Eleby and Jacobs had used PCP the day of Fuller’s murder. Some also tried to establish alibis for the time of Fuller’s death.

The jury convicted all seven petitioners, along with codefendant Steve Webb (who subsequently died). The jury acquitted codefendants Alfonso Harris and Felicia Ruffin. On direct appeal, the D. C. Court of Appeals affirmed petitioners’ convictions, though it remanded for resentencing. 545 A. 2d 1202, 1219 (1988). The trial court resentenced petitioners to the same amount of prison time. App. to Pet. for Cert. in No. 15–1503, at 82a, n. 2.

B

The Brady Claims

Beginning in 2010, petitioners pursued postconviction proceedings in which they sought to vacate their convictions or to be granted a new trial. App. to Pet. for Cert. in No. 15–1503, at 84a, n. 4. After petitioners’ convictions became final, it emerged that the Government possessed certain evidence that it had withheld from the defense at the time of trial. Petitioners discovered other withheld evidence in their review of the trial prosecutor’s case file, which the Government turned over to petitioners in the course of the postconviction proceedings. Among other postconviction claims, petitioners contended that the withheld evidence was both favorable and material, entitling them to relief under Brady.

The D. C. Superior Court considered petitioners’ Brady claims as part of a 16-day evidentiary hearing. It rejected those claims, finding that “none of the undisclosed information was material.” App. to Pet. for Cert. in No. 15–1503, at 130a. The D. C. Court of Appeals affirmed. 116 A. 3d, at 901. It similarly concluded that the withheld evidence was not material under Brady. 116 A. 3d, at 913–926. At issue in those proceedings were the following seven specific pieces of evidence:

1. The identity of James McMillan. Freeman, the vendor who discovered Fuller’s body in the alley garage, testified at trial that, while he was waiting for police to arrive, he saw two men run into the alley and stop near the garage for about five minutes before running away when an officer approached. One of the men had a bulge under his coat. Early in the trial, codefendant Harris’ counsel had requested the identity of the two men to confirm that her client was not one of them. But the Government refused to disclose the men’s identity.

In their postconviction review of the prosecutor’s files, petitioners learned that Freeman had identified the two men he saw in the alley as James McMillan and Gerald Merkerson. McMillan lived in a house which opens in the back onto a connecting alley. In the weeks following Fuller’s murder, but before petitioners’ trial, McMillan was arrested for beating and robbing two women in the neighborhood. Neither attack included a sexual assault. Separately, petitioners learned that seven years after petitioners’ trial, McMillan had robbed, sodomized, and murdered a young woman in an alley.

2. The interview with Willie Luchie. The prosecutor’s notes also recorded an undisclosed interview with Willie Luchie, who told the prosecutor that he and three others walked through the alley on their way to an H Street liquor store between 5:30 and 5:45 p.m. on the evening of the murder. As the group walked by the garage, Luchie “heard several groans” and “remembers the doors to the garage being closed.” App. 25. Another person in the group recalled “hear[ing] some moans,” while the other two persons did not recall hearing anything unusual. Id., at 27, 53; id., at A992. The group continued walking without looking into the garage or otherwise investigating the source of the sounds. They did not see McMillan or any other person in the alley when they passed through.

3. The interviews with Ammie Davis. Undisclosed notes written by a police officer and the prosecutor refer to two interviews with Ammie Davis, who had been arrested for disorderly conduct a few weeks after Fuller’s murder. Davis initially told a police investigator that she had seen another individual, James Blue, beat Fuller to death in the alley. Shortly thereafter, she said she only saw Blue grab Fuller and push her into the alley. Davis also said that a girlfriend, whom she did not name, accompanied her. She promised to call the investigator with more details, but she did not do so.

About 9 months later (after petitioners were indicted but approximately 11 weeks before their trial), a prosecutor learned of the investigator’s notes and interviewed Davis. The prosecutor’s notes state that Davis did not provide any more details, except to say that the girlfriend who accompanied her was nicknamed “ ‘Shorty.’ ” Id., at 267–268. About two months later, which was shortly before petitioners’ trial, Blue murdered Davis in an unrelated drug dispute.

During the postconviction evidentiary hearing, the prosecutor who interviewed Davis testified that he did not disclose Davis’ statement because she acted “playful” and “not serious” during the interview and he found her to be “totally incredible.” Id., at 269–272. Additionally, the prosecutor stated that he knew Davis had previously falsely accused Blue of a different murder, and on another occasion had falsely accused a different individual of a different murder.

4. Impeachment of Kaye Porter and Carrie Eleby. Kaye Porter accompanied Eleby during an initial interview with homicide detectives. Porter agreed with Eleby that she had also heard Alston state that he was involved in robbing Fuller. An undisclosed prosecutorial note states that in a later interview with detectives, Porter stated that she did not actually recall hearing Alston’s statement and just went along with what Eleby said. The note also states that Eleby likewise admitted that she had lied about Porter being present during Alston’s statement and had asked Porter to support her.

5. Impeachment of Carrie Eleby. A prosecutor’s un-disclosed note revealed that Eleby said she had beenhigh on PCP during a January 9, 1985, meeting with investigators.

6. Impeachment of Linda Jacobs. An undisclosed note of an interview with Linda Jacobs said that the detective had “question[ed] her hard,” and that she had “vacillated” about what she saw. Id., at A1009. The prosecutor recalled that the detective “kept raising his voice” and was “smacking his hand on the desk” during the interview. Id., at A2298–A2299.

7. Impeachment of Maurice Thomas. An undisclosed note of an interview with Maurice Thomas’ aunt stated that she “does not recall Maurice ever telling her anything such as this.” Id., at A1010; see id., at 295–296.

II

A

The Government does not contest petitioners’ claim that the withheld evidence was “favorable to the accused, either because it is exculpatory, or because it is impeaching.” Strickler v. Greene, 527 U. S. 263 –282 (1999). Neither does the Government contest petitioners’ claim that it “suppressed” the evidence, “either willfully or inadvertently.” Id., at 282. It does, as it must, concede that the Brady rule’s “ ‘overriding concern [is] with the justice of the finding of guilt,’ ” United States v. Bagley, 473 U. S. 667, 678 (1985) (quoting United States v. Agurs, 427 U. S. 97, 112 (1976) ), and that the Government’s “ ‘interest . . . in a criminal prosecution is not that it shall win a case, but that justice shall be done,’ ” Kyles v. Whitley, 514 U. S. 419, 439 (1995) (quoting Berger v. United States, 295 U. S. 78, 88 (1935) ). Consistent with these principles, the Government assured the Court at oral argument that subsequent to petitioners’ trial, it has adopted a “generous policy of discovery” in criminal cases under which it discloses any “information that a defendant might wish to use.” Tr. of Oral Arg. 47–48. As we have recognized, and as the Government agrees, ibid., “[t]his is as it should be.” Kyles, supra, at 439 (explaining that a “ ‘prudent prosecutor[’s]’ ” better course is to take care to disclose any evidence favorable to the defendant (quoting Agurs, supra, at 108)).

Petitioners and the Government, however, do contest the materiality of the undisclosed Brady information. “[E]vidence is ‘material’ within the meaning of Brady when there is a reasonable probability that, had the evidence been disclosed, the result of the proceeding would have been different.” Cone v. Bell, 556 U. S. 449 –470 (2009) (citing Bagley, supra, at 682). “A ‘reasonable probability’ of a different result” is one in which the suppressed evidence “ ‘undermines confidence in the outcome of the trial.’ ” Kyles, supra, at 434 (quoting Bagley, supra, at 678). In other words, petitioners here are entitled to a new trial only if they “establis[h] the prejudice necessary to satisfy the ‘materiality’ inquiry.” Strickler, supra, at 282.

Consequently, the issue before us here is legally simple but factually complex. We must examine the trial record, “evaluat[e]” the withheld evidence “in the context of the entire record,” Agurs, supra, at 112, and determine in light of that examination whether “there is a reasonable prob-ability that, had the evidence been disclosed, the result of the proceeding would have been different.” Cone, supra, at 470 (citing Bagley, supra, at 682). Having done so, we agree with the lower courts that there was no such reasonable probability.

B

Petitioners’ main argument is that, had they known about McMillan’s identity and Luchie’s statement, they could have challenged the Government’s basic theory that Fuller was killed in a group attack. Petitioners contend that they could have raised an alternative theory, namely, that a single perpetrator (or two at most) had attacked Fuller. According to petitioners, the groans that Luchie and his companion heard when they walked through the alley between 5:30 and 5:45 p.m. suggest that the attack was taking place inside the garage at that moment. The added facts that the garage was small and that Luchie’s group saw no one in the alley could bolster a “single attacker” theory. Freeman’s recollection that one garage door was open when he found Fuller’s body at around 6 p.m., combined with Luchie’s recollection that both doors were shut around 5:30 or 5:45 p.m., could suggest that one or two perpetrators were in the garage when Luchie walked by but left before Freeman arrived. McMillan’s identity as one of the men Freeman saw enter the alley after Freeman discovered Fuller’s body would have revealed McMillan’s criminal convictions in the months before petitioners’ trial. Petitioners argue that together, this evidence would have permitted the defense to knit together a theory that the group attack did not occur at all—and that it was actually McMillan, alone or with an accomplice, who murdered Fuller. They add that they could have used the investigators’ failure to follow up on Ammie Davis’ claim about James Blue, and the various pieces of withheld impeachment evidence, to suggest that an incomplete investigation had ended up accusing the wrong persons.

Considering the withheld evidence “in the context of the entire record,” however, Agurs, supra, at 112, we conclude that it is too little, too weak, or too distant from the main evidentiary points to meet Brady’s standards. As petitioners recognize, McMillan’s guilt (or that of any other single, or near single, perpetrator) is inconsistent with petitioners’ guilt only if there was no group attack. But a group attack was the very cornerstone of the Government’s case. The witnesses may have differed on minor details, but virtually every witness to the crime itself agreed as to a main theme: that Fuller was killed by a large group of perpetrators. The evidence at trial was such that, even though petitioners knew that Freeman saw two men enter the alley after he discovered Fuller’s body, that one appeared to have a bulky object hidden under his coat, and that both ran when the police arrived, none of the petitioners attempted to mount a defense that implicated those men as alternative perpetrators acting alone.

Is it reasonably probable that adding McMillan’s identity, and Luchie’s ambiguous statement that he heard groans but saw no one, could have led to a different result at trial? We conclude that it is not. The problem for petitioners is that their current alternative theory would have had to persuade the jury that both Alston and Bennett falsely confessed to being active participants in a group attack that never occurred; that Yarborough falsely implicated himself in that group attack and, through coordinated effort or coincidence, gave a highly similar account of how it occurred; that Thomas, a disinterested witness who recognized petitioners when he happened upon the attack and heard Catlett refer to it later that night, wholly fabricated his story; that both Eleby and Jacobs likewise testified to witnessing a group attack that did not occur; and that Montgomery in fact did not see petitioners and others, as a group, identify Fuller as a target and leave the park to rob her.

With respect to the undisclosed impeachment evidence, the record shows that it was largely cumulative of impeachment evidence petitioners already had and used at trial. For example, the jury heard multiple times about Eleby’s frequent PCP use, including Eleby’s own testimony that she and Jacobs had smoked PCP shortly before they witnessed Fuller’s attack. In this context, it would not have surprised the jury to learn that Eleby used PCP on yet another occasion. Porter was a minor witness who was also impeached at trial with evidence about changes in her testimony over time, leaving little added significance to the note that she changed her mind about having agreed with Eleby’s claims. The jury was also well aware of Jacobs’ vacillation, as she was impeached on the stand with her shifting stories about what she witnessed. Knowledge that a detective raised his voice during an interview with her would have added little more. Nor do we see how the note about the statement by Thomas’ aunt could have mattered much, given the facts that neither side chose to call the aunt as a witness and that the jury already knew, from Thomas’ testimony, that his aunt had told him not to tell anyone what he saw. As for James Blue, petitioners argue that the investigators’ delay in following up on Ammie Davis’ statement could have led the jury to doubt the thoroughness of the investigation. But this likelihood is seriously undercut by notes about Davis’ demeanor and lack of detail, and by her prior false accusations that Blue committed a different murder and that yet another person committed yet a different murder.

We of course do not suggest that impeachment evidence is immaterial with respect to a witness who has already been impeached with other evidence. See Wearry v. Cain, 577 U. S. ___, ___–___ (2016) (per curiam) (slip op., at 7–9). We conclude only that in the context of this trial, with respect to these witnesses, the cumulative effect of the withheld evidence is insufficient to “ ‘undermine confidence’ ” in the jury’s verdict, Smith, 565 U. S., at 75–76 (quoting Kyles, 514 U. S., at 434; brackets omitted).

III

On the basis of our review of the record, we agree with the lower courts that there is not a “reasonable probability” that the withheld evidence would have changed the outcome of petitioners’ trial, id., at 434 (internal quotation marks omitted). The judgment of the D. C. Court of Appeals, accordingly, is affirmed.

It is so ordered.

Justice Gorsuch took no part in the consideration or decision of these cases.

APPENDIX

SUPREME COURT OF THE UNITED STATES

_________________

Nos. 15–1503 and 15–1504

_________________

CHARLES S. TURNER, et al., PETITIONERS

15–1503v.

UNITED STATES

RUSSELL L. OVERTON, PETITIONER

15–1504v.

UNITED STATES

on writs of certiorari to the district of columbia court of appeals

[June 22, 2017]

Justice Kagan, with whom Justice Ginsburg joins, dissenting.

Consider two criminal cases. In the first, the government accuses ten defendants of acting together to commit a vicious murder and robbery. At trial, each defendant accepts that the attack occurred almost exactly as the government describes—contending only that he wasn’t part of the rampaging group. The defendants thus undermine each other’s arguments at every turn. In the second case, the government makes the same arguments as before. But this time, all of the accused adopt a common defense, built around an alternative account of the crime. Armed with new evidence that someone else perpetrated the murder, the defendants vigorously dispute the government’s gang-attack narrative and challenge the credibility of its investigation. The question this case presents is whether such a unified defense, relying on evidence unavailable in the first scenario, had a “reason-able probability” (less than a preponderance) of shiftingeven one juror’s vote. Cone v. Bell, 556 U. S. 449, 452, 470 (2009) ; see Kyles v. Whitley, 514 U. S. 419, 434 (1995) .

That is the relevant question because the Government here knew about but withheld the evidence of an alternative perpetrator—and so prevented the defendants from coming together to press that theory of the case. If the Government’s non-disclosure was material, in the sense just described, this Court’s decision in Brady v. Maryland, 373 U. S. 83 (1963) , demands a new trial. The Court today holds it was not material: In light of the evidence the Government offered, the majority argues, the transformed defense stood little chance of persuading a juror to vote to acquit. That conclusion is not indefensible: The Government put on quite a few witnesses who said that the defendants committed the crime. But in the end, I think the majority gets the answer in this case wrong. With the undisclosed evidence, the whole tenor of the trial would have changed. Rather than relying on a “not me, maybe them” defense, ante, at 6, all the defendants would have relentlessly impeached the Government’s (thoroughly impeachable) witnesses and offered the jurors a way to view the crime in a different light. In my view, that could well have flipped one or more jurors—which is all Brady requires.

Before explaining that view, I note that the majority and I share some common ground. We agree on the universe of exculpatory or impeaching evidence suppressed in this case: The majority’s description of that evidence, and of the trial held without it, is scrupulously fair. See ante, at 2–6, 7–9. We also agree—as does the Government—that such evidence ought to be disclosed to defendants as a matter of course. See ante, at 10. Constitutional requirements aside, turning over exculpatory materials is a core responsibility of all prosecutors—whose professional interest and obligation is not to win cases but to ensure justice is done. See Kyles, 514 U. S., at 439. And finally, we agree on the legal standard by which to assess the materiality of undisclosed evidence for purposes of applying the constitutional rule: Courts are to ask whether there is a “reasonable probability” that disclosure of the evidence would have led to a different outcome—i.e., an acquittal or hung jury rather than a conviction. See ante, at 10.

But I part ways with the majority in applying that standard to the evidence withheld in this case. That evidence falls into three basic categories, discussed below. Taken together, the materials would have recast the trial significantly—so much so as to “undermine[ ] confidence” in the guilty verdicts reached in their absence. Kyles, 514 U. S., at 434.

First, the Government suppressed information identifying a possible alternative perpetrator. The defendants knew that, shortly before the police arrived, witnesses had observed two men acting suspiciously near the alleyway garage where Catherine Fuller’s body was found. But they did not know—because the Government never told them—that a witness had identified one of those men as James McMillan. Equipped with that information, the defendants would have discovered that in the weeks following Fuller’s murder, McMillan assaulted and robbed two other women of comparable age in the same neighborhood. And using that information, the defendants would have united around a common defense. They would all have pointed their fingers at McMillan (rather than at each other), arguing that he committed Fuller’s murder as part of a string of similar crimes.

Second, the Government suppressed witness statements suggesting that one or two perpetrators—not a large group—carried out the attack. Those statements were given by two individuals who walked past the garage around the time of Fuller’s death. They told the police that they heard groans coming from inside the garage; and one remarked that the garage’s doors were closed at the time. Introducing that evidence at trial would have sown doubt about the Government’s group-attack narrative, because that many people (as everyone agrees) couldn’t have fit inside the small garage. And the questions thus raised would have further supported the defendants’ theory that McMillan (and perhaps an accomplice) had committed the murder.

Third and finally, the Government suppressed a raft of evidence discrediting its investigation and impeaching its witnesses. Undisclosed files, for example, showed that the police took more than nine months to look into a witness’s claim that a man named James Blue had murdered Fuller. Evidence of that kind of negligence could easily have led jurors to wonder about the competence of all the police work done in the case. Other withheld documents revealed that one of the Government’s main witnesses was high on PCP when she met with investigators to identify participants in the crime—and that she also encouraged a friend to lie to the police to support her story. Using that sort of information, see also ante, at 9, the defendants could have undercut the Government’s witnesses—even while presenting their own account of the murder.

In reply to all this, the majority argues that “none of the [accused] attempted to mount [an alternative-perpetrator] defense” and that such a defense would have challenged “the very cornerstone of the Government’s case.” Ante, at 12. But that just proves my point. The defendants didn’t offer an alternative-perpetrator defense because the Government prevented them from learning what made it credible: that one of the men seen near the garage had a record of assaulting and robbing middle-aged women, and that witnesses would back up the theory that only one or two individuals had committed the murder. Moreover, that defense had game-changing potential exactly because it challenged the cornerstone of the Government’s case. Without the withheld evidence, each of the defendants had little choice but to accept the Government’s framing of the crime as a group attack—and argue only that he wasn’t there. That meant the defendants often worked at cross-purposes. In particular, each defendant not identified by a Government witness sought to bolster that witness’s credibility, no matter the harm to his co-defendants. As one defense lawyer remarked after another’s supposed cross-examination of a Government witness: “They’ve got [an extra] prosecutor[ ] in the courtroom now.” Saperstein & Walsh, 10 Defendants Complicate Trial, Washington Post, Nov. 17, 1985, p. A14, col. 1. Credible alternative-perpetrator evidence would have allowed the defendants to escape this cycle of mutually assured destruction. By enabling the defendants to jointly attack the Government’s “cornerstone” theory, the withheld evidence would have reframed the case presented to the jury.

Still, the majority claims, an alternative-perpetrator defense would have had no realistic chance of changing the outcome because the Government had ample evidence of a group attack, including five witnesses who testified that they had participated in it or seen it happen. See ante, at 12–13. But the Government’s case wasn’t nearly the slam-dunk the majority suggests. No physical evidence tied any of the defendants to the crime—a highly surprising fact if, as the Government claimed, more than ten people carried out a spur-of-the-moment, rampage-like attack in a confined space. And as even the majority recognizes, the Government’s five eyewitnesses had some serious credibility deficits. See ibid. Two had been charged as defendants, and agreed to testify only in exchange for favorable plea deals. See 116 A. 3d 894, 902 (D. C. 2015). Two admitted they were high on PCP at the time. See id., at 903, 911; App. A535–A536, A649. (As noted above, one was also high when she later met with police to identify the culprits.) One was an eighth-grader whose own aunt contradicted parts of his trial testimony. See 116 A. 3d, at 903, 911. Even in the absence of an alternative account of the crime, the jury took more than a week—and many dozens of votes—to reach its final verdict. Had the defendants offered a unified counter-narrative, based on the withheld evidence, one or more jurors could well have concluded that the Government had not proved its case beyond a reasonable doubt.

Again, the issue here concerns the difference between two criminal cases. The Government got the case it most wanted—the one in which the defendants, each in an effort to save himself, formed something of a circular firing squad. And the Government avoided the case it most feared—the one in which the defendants acted jointly to show that a man known to assault women like Fuller committed her murder. The difference between the two cases lay in the Government’s files—evidence of obvious relevance that prosecutors nonetheless chose to suppress. I think it could have mattered to the trial’s outcome. For that reason, I respectfully dissent.