Buck v. Davis, 580 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

_________________

No. 15–8049

_________________

DUANE EDWARD BUCK, PETITIONER v. LORIE DAVIS, DIRECTOR, TEXAS DEPARTMENT OF CRIMINAL JUSTICE, CORRECTIONAL INSTITUTIONS DIVISION

on writ of certiorari to the united states court of appeals for the fifth circuit

[February 22, 2017]

Chief Justice Roberts delivered the opinion of the Court.

A Texas jury convicted petitioner Duane Buck of capital murder. Under state law, the jury could impose a death sentence only if it found that Buck was likely to commit acts of violence in the future. Buck’s attorney called a psychologist to offer his opinion on that issue. The psychologist testified that Buck probably would not engage in violent conduct. But he also stated that one of the factors pertinent in assessing a person’s propensity for violence was his race, and that Buck was statistically more likely to act violently because he is black. The jury sentenced Buck to death.

Buck contends that his attorney’s introduction of this evidence violated his Sixth Amendment right to the effective assistance of counsel. This claim has never been heard on the merits in any court, because the attorney who represented Buck in his first state postconviction proceeding failed to raise it. In 2006, a Federal District Court relied on that failure—properly, under then-governing law—to hold that Buck’s claim was procedurally defaulted and unreviewable.

In 2014, Buck sought to reopen that 2006 judgment by filing a motion under Federal Rule of Civil Procedure 60(b)(6). He argued that this Court’s decisions in Martinez v. Ryan,566 U. S. 1 (2012), and Trevino v. Thaler, 569 U. S. ___ (2013), had changed the law in a way that provided an excuse for his procedural default, permitting him to litigate his claim on the merits. In addition to this change in the law, Buck’s motion identified ten other factors that, he said, constituted the “extraordinary circumstances” required to justify reopening the 2006 judgment under the Rule. See Gonzalez v. Crosby,545 U. S. 524,535 (2005).

The District Court below denied the motion, and the Fifth Circuit declined to issue the certificate of appealability (COA) requested by Buck to appeal that decision. We granted certiorari, and now reverse.

I

A

On the morning of July 30, 1995, Duane Buck arrived at the home of his former girlfriend, Debra Gardner. He was carrying a rifle and a shotgun. Buck entered the home, shot Phyllis Taylor, his stepsister, and then shot Gardner’s friend Kenneth Butler. Gardner fled the house, and Buck followed. So did Gardner’s young children. While Gardner’s son and daughter begged for their mother’s life, Buck shot Gardner in the chest. Gardner and Butler died of their wounds. Taylor survived.

Police officers arrived soon after the shooting and placed Buck under arrest. An officer would later testify that Buck was laughing at the scene. He remained “happy” and “upbeat” as he was driven to the police station, “[s]miling and laughing” in the back of the patrol car. App. 134a–135a, 252a.

Buck was tried for capital murder, and the jury convicted. During the penalty phase of the trial, the jury was charged with deciding two issues. The first was what the parties term the “future dangerousness” question. At the time of Buck’s trial, a Texas jury could impose the death penalty only if it found—unanimously and beyond a reasonable doubt—“a probability that the defendant would commit criminal acts of violence that would constitute a continuing threat to society.” Tex. Code Crim. Proc. Ann., Art. 37.071, §2(b)(1) (Vernon 1998). The second issue, to be reached only if the jury found Buck likely to be a future danger, was whether mitigating circumstances nevertheless warranted a sentence of life imprisonment instead of death. See §2(e).

The parties focused principally on the first question. The State called witnesses who emphasized the brutality of Buck’s crime and his evident lack of remorse in its aftermath. The State also called another former girlfriend, Vivian Jackson. She testified that, during their relationship, Buck had routinely hit her and had twice pointed a gun at her. Finally, the State introduced evidence of Buck’s criminal history, including convictions for delivery of cocaine and unlawfully carrying a weapon. App. 125a–127a, 185a.

Defense counsel answered with a series of lay witnesses, including Buck’s father and stepmother, who testified that they had never known him to be violent. Counsel also called two psychologists to testify as experts. The first, Dr. Patrick Lawrence, observed that Buck had previously served time in prison and had been held in minimum custody. From this he concluded that Buck “did not present any problems in the prison setting.” Record in No. 4:04–cv–03965 (SD Tex.), Doc. 5–116, pp. 12–13. Dr. Lawrence further testified that murders within the Texas penal system tend to be gang related (there was no evidence Buck had ever been a member of a gang) and that Buck’s offense had been a “crime of passion” occurring within the context of a romantic relationship. Id., at 4, 19, 21. Based on these considerations, Dr. Lawrence determined that Buck was unlikely to be a danger if he were sentenced to life in prison. Id., at 20–21.

Buck’s second expert, Dr. Walter Quijano, had been appointed by the presiding judge to conduct a psychological evaluation. Dr. Quijano had met with Buck in prison prior to trial and shared a report of his findings with defense counsel.

Like Dr. Lawrence, Dr. Quijano thought it significant that Buck’s prior acts of violence had arisen from romantic relationships with women; Buck, of course, would not form any such relationships while incarcerated. And Dr. Quijano likewise considered Buck’s behavioral record in prison a good indicator that future violence was unlikely. App. 36a, 39a–40a.

But there was more to the report. In determining whether Buck was likely to pose a danger in the future, Dr. Quijano considered seven “statistical factors.” The fourth factor was “race.” His report read, in relevant part: “4. Race. Black: Increased probability. There is an over-representation of Blacks among the violent offenders.” Id., at 19a.

Despite knowing Dr. Quijano’s view that Buck’s race was competent evidence of an increased probability of future violence, defense counsel called Dr. Quijano to the stand and asked him to discuss the “statistical factors” he had “looked at in regard to this case.” Id., at 145a–146a. Dr. Quijano responded that certain factors were “know[n] to predict future dangerousness” and, consistent with his report, identified race as one of them. Id., at 146a. “It’s a sad commentary,” he testified, “that minorities, Hispanics and black people, are over represented in the Criminal Justice System.” Ibid. Through further questioning, counsel elicited testimony concerning factors Dr. Quijano thought favorable to Buck, as well as his ultimate opinion that Buck was unlikely to pose a danger in the future. At the close of Dr. Quijano’s testimony, his report was admitted into evidence. Id., at 150a–152a.

After opening cross-examination with a series of general questions, the prosecutor likewise turned to the report. She asked first about the statistical factors of past crimes and age, then questioned Dr. Quijano about the roles of sex and race: “You have determined that the sex factor, that a male is more violent than a female because that’s just the way it is, and that the race factor, black, increases the future dangerousness for various complicated reasons; is that correct?” Id., at 170a. Dr. Quijano replied, “Yes.” Ibid.

During closing arguments, defense counsel emphasized that Buck had proved to be “controllable in the prison population,” and that his crime was one of “jealousy, . . . passion and emotion” unlikely to be repeated in jail. Id., at 189a–191a. The State stressed the crime’s brutal nature and Buck’s lack of remorse, along with the inability of Buck’s own experts to guarantee that he would not act violently in the future—a point it supported by reference to Dr. Quijano’s testimony. See id., at 198a–199a (“You heard from Dr. Quijano, . . . who told you that . . . the probability did exist that [Buck] would be a continuing threat to society.”).

The jury deliberated over the course of two days. During that time it sent out four notes, one of which requested the “psychology reports” that had been admitted into evidence. Id., at 209a. These reports—including Dr. Quijano’s—were provided. The jury returned a sentence of death.

B

Buck’s conviction and sentence were affirmed on direct appeal. Buck v. State, No. 72,810 (Tex. Crim. App., Apr. 28, 1999). His case then entered a labyrinth of state and federal collateral review, where it has wandered for the better part of two decades.

Buck filed his first petition for a writ of habeas corpus in Texas state court in 1999. The four claims advanced in his petition, however, were all frivolous or noncognizable. See Ex parte Buck, No. 699684–A (Dist. Ct. Harris Cty., Tex., July 11, 2003), pp. 6–7. The petition failed to mention defense counsel’s introduction of expert testimony that Buck’s race increased his propensity for violence.

But Dr. Quijano had testified in other cases, too, and in 1999, while Buck’s first habeas petition was pending, one of those cases reached this Court. The petitioner, Victor Hugo Saldano, argued that his death sentence had been tainted by Dr. Quijano’s testimony that Saldano’s Hispanic heritage “was a factor weighing in the favor of future dangerousness.” App. 302a. Texas confessed error on that ground and asked this Court to grant Saldano’s petition for certiorari, vacate the state court judgment, and remand the case. In June 2000, the Court did so. Saldano v. Texas,530 U. S. 1212.

Within days, the Texas Attorney General, John Cornyn, issued a public statement concerning the cases in which Dr. Quijano had testified. The statement affirmed that “it is inappropriate to allow race to be considered as a factor in our criminal justice system.” App. 213a. In keeping with that principle, the Attorney General explained that his office had conducted a “thorough audit” and “identified eight more cases in which testimony was offered by Dr. Quijano that race should be a factor for the jury to consider in making its determination about the sentence in a capital murder trial.” Ibid. Six of those cases were “similar to that of Victor Hugo Saldano”; in those cases, letters had been sent to counsel apprising them of the Attorney General’s findings. Id., at 213a–214a. The statement closed by identifying the defendants in those six cases. Buck was one of them. Id., at 215a–217a. By the close of 2002, the Attorney General had confessed error, waived any avail-able procedural defenses, and consented to resentencing in the cases of five of those six defendants. See Alba v. Johnson, 232 F. 3d 208 (CA5 2000) (Table); Memorandum and Order in Blue v. Johnson, No. 4:99–cv–00350 (SD Tex.), pp. 15–17; Order in Garcia v. Johnson, No. 1:99–cv–00134 (ED Tex.), p. 1; Order in Broxton v. Johnson, No. 4:00–cv–01034 (SD Tex.), pp. 10–11; Final Judgment in Gonzales v. Cockrell, No. 7:99–cv–00072 (WD Tex.), p. 1.

Not, however, in Buck’s. In 2002, Buck’s attorney filed a new state habeas petition alleging that trial counsel had rendered ineffective assistance by introducing Dr. Quijano’s testimony. The State was not represented by the Attorney General in this proceeding—the Texas Attorney General represents state respondents in federal habeas cases, but not state habeas cases—and it did not confess error. Because Buck’s petition was successive, the Texas Court of Criminal Appeals dismissed it as an abuse of the writ. Ex parte Buck, Nos. 57,004–01, 57,004–02 (Tex. Crim. App., Oct. 15, 2003) ( per curiam).

Buck turned to the federal courts. He filed a petition for habeas corpus under28 U. S. C. §2254 in October 2004, by which time Attorney General Cornyn had left office. See Buck v. Dretke, 2006 WL 8411481, *2 (SD Tex., July 24, 2006). Buck sought relief on the ground that trial counsel’s introduction of Dr. Quijano’s testimony was constitutionally ineffective. The State responded that the state court had dismissed Buck’s ineffective assistance claim because Buck had failed to press it in his first petition, raising it for the first time in a procedurally improper second petition. The State argued that such reliance on an established state rule of procedure was an adequate and independent state ground precluding federal review. Texas acknowledged that it had waived similar procedural defenses in Saldano’s case. But it argued that Buck’s case was different because “[i]n Saldano’s case Dr. Quijano testified for the State”; in Buck’s, “it was Buck who called Dr. Quijano to testify.” Answer and Motion for Summary Judgment in No. 4:04–cv–03965 (SD Tex.), p. 20.

Buck countered that, notwithstanding his procedural default, the District Court should reach the merits of his claim because a failure to do so would result in a miscarriage of justice. Buck did not argue that his default should be excused on a showing of “cause” and “prejudice”—that is, cause for the default, and prejudice from the denial of a federal right. And for good reason: At the time Buck filed his §2254 petition, our decision in Coleman v. Thompson,501 U. S. 722–753 (1991), made clear that an attorney’s failure to raise an ineffective assistance claim during state postconviction review could not constitute cause. The District Court rejected Buck’s miscarriage of justice argument and held that, because of his procedural default, his ineffective assistance claim was unreviewable. Buck v. Dretke, 2006 WL 8411481, at *8. Buck unsuccessfully sought review of the District Court’s ruling. See Buck v. Thaler, 345 Fed. Appx. 923 (CA5 2009) ( per curiam) (denying application for a COA), cert. denied,559 U. S. 1072 (2010).

In 2011, Buck sought to reopen his case, arguing that the prosecution had violated the Equal Protection and Due Process Clauses by asking Dr. Quijano about the relationship between race and future violence on cross-examination and referring to his testimony during summation. Buck also argued that the State’s decision to treat him differently from the other defendants affected by Dr. Quijano’s testimony justified relieving him of the District Court’s adverse judgment. The Fifth Circuit disagreed, see Buck v. Thaler, 452 Fed. Appx. 423, 427–428 (CA5 2011) ( per curiam), and we denied certiorari, Buck v. Thaler,565 U. S. 1022 (2011). Buck, still barred by Coleman from avoiding the consequences of his procedural default, did not pursue his ineffective assistance claim.

C

In 2012, this Court “modif[ied] the unqualified statement in Coleman that an attorney’s ignorance or inadvertence in a postconviction proceeding does not qualify as cause to excuse a procedural default.” Martinez, 566 U. S., at 9. We held that when a state formally limits the adjudication of claims of ineffective assistance of trial counsel to collateral review, a prisoner may establish cause for procedural default if (1) “the state courts did not appoint counsel in the initial-review collateral proceeding,” or “appointed counsel in [that] proceeding . . . was ineffective under the standards of Strickland v. Washington,466 U. S. 668 (1984)”; and (2) “the underlying . . . claim is a substantial one, which is to say that . . . the claim has some merit.” Id., at 14.

By its terms, Martinez did not bear on Buck’s ineffective assistance claim. At the time of Buck’s conviction and appeal, Texas did not formally require criminal defendants to reserve such claims for collateral review. In Tre-vino, however, the Court concluded that the exception an-nounced in Martinez extended to state systems that, as a practical matter, deny criminal defendants “a meaningful opportunity” to press ineffective assistance claims on direct appeal. 569 U. S., at ___ (slip op., at 13). The Court further concluded that the system in Texas, where petitioner had been convicted, was such a system. Ibid. The upshot: Had Martinez and Trevino been decided before Buck filed his §2254 petition, a federal court could have reviewed Buck’s ineffective assistance claim if he demonstrated that (1) state postconviction counsel had been constitutionally ineffective in failing to raise it, and (2) the claim had “some merit.” Martinez, 566 U. S., at 14.

D

When Trevino was decided, Buck’s third state habeas petition was pending in Texas court. That petition was denied in November 2013. Ex parte Buck, 418 S. W. 3d 98 (Tex. Crim. App. 2013) ( per curiam). Two months later, Buck returned to federal court, where he filed a motion to reopen his §2254 case under Federal Rule of Civil Procedure 60(b)(6). Rule 60(b) enumerates specific circumstances in which a party may be relieved of the effect of a judgment, such as mistake, newly discovered evidence, fraud, and the like. The Rule concludes with a catchall category—subdivision (b)(6)—providing that a court may lift a judgment for “any other reason that justifies relief.” Relief is available under subdivision (b)(6), however, only in “extraordinary circumstances,” and the Court has explained that “[s]uch circumstances will rarely occur in the habeas context.” Gonzalez, 545 U. S., at 535.

In his motion, Buck identified 11 factors that, in his view, justified reopening the judgment. These included his attorney’s introduction of expert testimony linking Buck’s race to violence, the central issue at sentencing; the prosecution’s questions about race and violence on cross-examination and reliance on Dr. Quijano’s testimony in summation; the State’s confession of error in other cases in which Dr. Quijano testified, but its refusal to concede error in Buck’s case; and the change in law effected by Martinez and Trevino, which, if they had been decided earlier, would have permitted federal review of Buck’s defaulted claim. App. 283a–285a.

The District Court denied relief on two grounds. First, the court concluded that Buck had failed to demonstrate extraordinary circumstances. To that end, the court observed that a change in decisional law is rarely extraordinary by itself. Buck v. Stephens, 2014 WL 11310152, *4 (SD Tex., Aug. 29, 2014). It further determined that the State’s “promise” not to oppose resentencing did not count for much, reasoning that “Buck’s case is different in critical respects from the cases in which Texas confessed error” in that Buck’s lawyer, not the prosecutor, had first elicited the objectionable testimony. Id., at *4–*5. The court also dismissed the contention that the nature of Dr. Quijano’s testimony argued for reopening the case. Although “the introduction of any mention of race was,” in the court’s view, “ill[ ]advised at best and repugnant at worst,” it was also “de minimis”: Dr. Quijano had discussed the connection between race and violence only twice. Id., at *5. The court accordingly concluded that Buck had failed to make out the predicate for Rule 60(b)(6) relief.

Second, the court determined that—even if the circumstances were extraordinary—Buck’s claim would fail on the merits. The court noted that under Strickland, Buck was obliged to show that counsel’s performance was both deficient and prejudicial. The court held that Buck’s lawyer had indeed performed deficiently in calling Dr. Quijano to give testimony that “len[t] credence to any potential latent racial prejudice held by the jury.” 2014 WL 11310152, at *6. But, the court concluded, Buck had failed to demonstrate prejudice. It observed that Buck’s crime had been “horrific.” Ibid. And the court had already concluded that “the introduction of any mention of race was . . . de minimis.” Id., at *5. For those reasons, it held, Buck had failed to show a reasonable probability that he would not have been sentenced to death but for Dr. Quijano’s testimony about race and violence.

Buck sought to appeal the denial of his Rule 60(b)(6) motion. He accordingly filed an application for a COA with the Fifth Circuit. To obtain a COA, Buck was required to make “a substantial showing of the denial of a constitutional right.”[1]*28 U. S. C. §2253(c)(2).

The Fifth Circuit denied a COA, concluding that Buck’s case was “not extraordinary at all in the habeas context.” Buck v. Stephens, 623 Fed. Appx. 668, 673 (2015). The panel agreed with the District Court that Martinez and Trevino were not significant factors in the analysis. It characterized most of the other factors Buck had identified as “variations on the merits” of his claim, which was “at least unremarkable as far as [ineffective assistance] claims go.” 623 Fed. Appx., at 673. The panel likewise rejected Buck’s argument that he was entitled to relief because the State had issued a press release indicating that his case would be treated like Saldano’s, and then had confessed error in the other cases identified as similar in the statement, but not in Buck’s. Id., at 674. Because Buck had “not shown extraordinary circumstances that would permit relief under Federal Rule of Civil Procedure 60(b)(6),” the panel “den[ied] the application for a COA.” Id., at 669.

Buck’s motion for rehearing en banc was denied over two dissenting votes. Buck v. Stephens, 630 Fed. Appx. 251 (CA5 2015) ( per curiam). We granted certiorari. Buck v. Stephens, 578 U. S. ___ (2016).

II

A state prisoner whose petition for a writ of habeas corpus is denied by a federal district court does not enjoy an absolute right to appeal. Federal law requires that he first obtain a COA from a circuit justice or judge.28 U. S. C. §2253(c)(1). A COA may issue “only if the applicant has made a substantial showing of the denial of a constitutional right.” §2253(c)(2). Until the prisoner secures a COA, the Court of Appeals may not rule on the merits of his case. Miller-El v. Cockrell,537 U. S. 322 (2003).

The COA inquiry, we have emphasized, is not coextensive with a merits analysis. At the COA stage, the only question is whether the applicant has shown that “jurists of reason could disagree with the district court’s resolution of his constitutional claims or that jurists could conclude the issues presented are adequate to deserve encouragement to proceed further.” Id., at 327. This threshold question should be decided without “full consideration of the factual or legal bases adduced in support of the claims.” Id., at 336. “When a court of appeals sidesteps [the COA] process by first deciding the merits of an appeal, and then justifying its denial of a COA based on its adjudication of the actual merits, it is in essence deciding an appeal without jurisdiction.” Id., at 336–337.

The court below phrased its determination in proper terms—that jurists of reason would not debate that Buck should be denied relief, 623 Fed. Appx., at 674—but it reached that conclusion only after essentially deciding the case on the merits. As the court put it in the second sentence of its opinion: “Because [Buck] has not shown extraordinary circumstances that would permit relief under Federal Rule of Civil Procedure 60(b)(6), we deny the application for a COA.” Id., at 669. The balance of the Fifth Circuit’s opinion reflects the same approach. The change in law effected by Martinez and Trevino, the panel wrote, was “not an extraordinary circumstance.” 623 Fed. Appx., at 674. Even if Texas initially indicated to Buck that he would be resentenced, its “decision not to follow through” was “not extraordinary.” Ibid. Buck “ha[d] not shown why” the State’s alleged broken promise “would justify relief from the judgment.” Ibid.

But the question for the Fifth Circuit was not whether Buck had “shown extraordinary circumstances” or “shown why [Texas’s broken promise] would justify relief from the judgment.” Id., at 669, 674. Those are ultimate merits determinations the panel should not have reached. We reiterate what we have said before: A “court of appeals should limit its examination [at the COA stage] to a threshold inquiry into the underlying merit of [the] claims,” and ask “only if the District Court’s decision was debatable.” Miller-El, 537 U. S., at 327, 348.

The dissent does not accept this established rule, arguing that a reviewing court that deems a claim nondebatable “must necessarily conclude that the claim is meritless.” Post, at 2 (opinion of Thomas, J.). Of course when a court of appeals properly applies the COA standard and determines that a prisoner’s claim is not even debatable, that necessarily means the prisoner has failed to show that his claim is meritorious. But the converse is not true. That a prisoner has failed to make the ultimate showing that his claim is meritorious does not logically mean he failed to make a preliminary showing that his claim was debatable. Thus, when a reviewing court (like the Fifth Circuit here) inverts the statutory order of operations and “first decid[es] the merits of an appeal, . . . then justif[ies] its denial of a COA based on its adjudication of the actual merits,” it has placed too heavy a burden on the prisoner at the COA stage. Miller-El, 537 U. S., at 336–337. Miller-El flatly prohibits such a departure from the procedure prescribed by §2253. Ibid.

The State defends the Fifth Circuit’s approach by arguing that the court’s consideration of an application for a COA is often quite thorough. The court “occasionally hears oral argument when considering whether to grant a COA in a capital case.” Brief for Respondent 50. Indeed, in one recent case, it “received nearly 200 pages of initial briefing, permitted a reply brief, considered the parties’ supplemental authorities, invited supplemental letter briefs from both sides, and heard oral argument before denying the request for a COA.” Id., at 50–51.

But this hurts rather than helps the State’s case. “[A] claim can be debatable even though every jurist of reason might agree, after the COA has been granted and the case has received full consideration, that petitioner will not prevail.” Miller-El, 537 U. S., at 338. The statute sets forth a two-step process: an initial determination whether a claim is reasonably debatable, and then—if it is—an appeal in the normal course. We do not mean to specify what procedures may be appropriate in every case. But whatever procedures are employed at the COA stage should be consonant with the limited nature of the inquiry.

Given the approach of the court below, it is perhaps understandable that the parties have essentially briefed and argued the underlying merits at length. See, e.g., Brief for Petitioner 32 (“[T]rial counsel rendered deficient performance under Strickland.”); id., at 39 (“[T]here is a reasonable probability that Dr. Quijano’s race-as-dangerousness opinion swayed the judgment of jurors in favor of death.” (internal quotation marks and alteration omitted)); id., at 59 (Buck “has demonstrated his entitlement to relief under Rule 60(b)(6)”); Brief for Respondent 40 (“The particular facts of petitioner’s case do not establish extraordinary circumstances justifying relief from the judgment.” (boldface type deleted)). With respect to this Court’s review, §2253 does not limit the scope of our consideration of the underlying merits, and at this juncture we think it proper to meet the decision below and the arguments of the parties on their own terms.

III

Buck’s request for a COA raised two separate questions for the Fifth Circuit, one substantive and one procedural: first, whether reasonable jurists could debate the District Court’s conclusion that Buck was not denied his right to effective assistance of counsel under Strickland; and second, whether reasonable jurists could debate the District Court’s procedural holding that Buck had not made the necessary showing to reopen his case under Rule 60(b)(6).

A

We begin with the District Court’s determination (not specifically addressed by the Fifth Circuit) that Buck’s constitutional claim failed on the merits. The Sixth Amendment right to counsel “is the right to the effective assistance of counsel.” Strickland, 466 U. S., at 686 (quoting McMann v. Richardson,397 U. S. 759, n. 14 (1970)). A defendant who claims to have been denied ef-fective assistance must show both that counsel performed deficiently and that counsel’s deficient performance caused him prejudice. 466 U. S., at 687.

1

Strickland’s first prong sets a high bar. A defense lawyer navigating a criminal proceeding faces any number of choices about how best to make a client’s case. The lawyer has discharged his constitutional responsibility so long as his decisions fall within the “wide range of professionally competent assistance.” Id., at 690. It is only when the lawyer’s errors were “so serious that counsel was not functioning as the ‘counsel’ guaranteed . . . by the Sixth Amendment” that Strickland’s first prong is satisfied. Id., at 687.

The District Court determined that, in this case, counsel’s performance fell outside the bounds of competent representation. We agree. Counsel knew that Dr. Quijano’s report reflected the view that Buck’s race disproportionately predisposed him to violent conduct; he also knew that the principal point of dispute during the trial’s penalty phase was whether Buck was likely to act violently in the future. Counsel nevertheless (1) called Dr. Quijano to the stand; (2) specifically elicited testimony about the connection between Buck’s race and the likelihood of future violence; and (3) put into evidence Dr. Quijano’s expert report that stated, in reference to factors bearing on future dangerousness, “Race. Black: Increased probability.” App. 19a, 145a–146a.

Given that the jury had to make a finding of future dangerousness before it could impose a death sentence, Dr. Quijano’s report said, in effect, that the color of Buck’s skin made him more deserving of execution. It would be patently unconstitutional for a state to argue that a defendant is liable to be a future danger because of his race. See Zant v. Stephens,462 U. S. 862,885 (1983) (identifying race among factors that are “constitutionally impermissible or totally irrelevant to the sentencing process”). No competent defense attorney would introduce such evidence about his own client. See Buck v. Thaler, 565 U. S., at 1022 (statement of Alito, J., joined by Scalia and Breyer, JJ., respecting denial of certiorari) (Buck’s case “concerns bizarre and objectionable testimony”).

2

To satisfy Strickland, a litigant must also demonstrate prejudice—“a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” 466 U. S., at 694. Accordingly, the question before the District Court was whether Buck had demonstrated a reasonable probability that, without Dr. Quijano’s testimony on race, at least one juror would have harbored a reasonable doubt about whether Buck was likely to be violent in the future. The District Court concluded that Buck had not made such a showing. We disagree.

In arguing that the jury would have imposed a death sentence even if Dr. Quijano had not offered race-based testimony, the State primarily emphasizes the brutality of Buck’s crime and his lack of remorse. A jury may conclude that a crime’s vicious nature calls for a sentence of death. See Wong v. Belmontes,558 U. S. 15 (2009) ( per curiam). In this case, however, several considerations convince us that it is reasonably probable—notwithstanding the nature of Buck’s crime and his behavior in its aftermath—that the proceeding would have ended differently had counsel rendered competent representation.

Dr. Quijano testified on the key point at issue in Buck’s sentencing. True, the jury was asked to decide two issues—whether Buck was likely to be a future danger, and, if so, whether mitigating circumstances nevertheless justified a sentence of life imprisonment. But the focus of the proceeding was on the first question. Much of the penalty phase testimony was directed to future dangerousness, as were the summations for both sides. The jury, consistent with the focus of the parties, asked during deliberations to see the expert reports on dangerousness. See App. 187a–196a, 198a–203a, 209a.

Deciding the key issue of Buck’s dangerousness involved an unusual inquiry. The jurors were not asked to determine a historical fact concerning Buck’s conduct, but to render a predictive judgment inevitably entailing a degree of speculation. Buck, all agreed, had committed acts of terrible violence. Would he do so again?

Buck’s prior violent acts had occurred outside of prison, and within the context of romantic relationships with women. If the jury did not impose a death sentence, Buck would be sentenced to life in prison, and no such romantic relationship would be likely to arise. A jury could conclude that those changes would minimize the prospect of future dangerousness.

But one thing would never change: the color of Buck’s skin. Buck would always be black. And according to Dr. Quijano, that immutable characteristic carried with it an “[i]ncreased probability” of future violence. Id., at 19a. Here was hard statistical evidence—from an expert—to guide an otherwise speculative inquiry.

And it was potent evidence. Dr. Quijano’s testimony appealed to a powerful racial stereotype—that of black men as “violence prone.” Turner v. Murray,476 U. S. 28,35 (1986) (plurality opinion). In combination with the substance of the jury’s inquiry, this created something of a perfect storm. Dr. Quijano’s opinion coincided precisely with a particularly noxious strain of racial prejudice, which itself coincided precisely with the central question at sentencing. The effect of this unusual confluence of factors was to provide support for making a decision on life or death on the basis of race.

This effect was heightened due to the source of the testimony. Dr. Quijano took the stand as a medical expert bearing the court’s imprimatur. The jury learned at the outset of his testimony that he held a doctorate in clinical psychology, had conducted evaluations in some 70 capital murder cases, and had been appointed by the trial judge (at public expense) to evaluate Buck. App. 138a–141a. Reasonable jurors might well have valued his opinion concerning the central question before them. See Satterwhite v. Texas,486 U. S. 249,259 (1988) (testimony from “a medical doctor specializing in psychiatry” on the question of future dangerousness may have influenced the sentencing jury).

For these reasons, we cannot accept the District Court’s conclusion that “the introduction of any mention of race” during the penalty phase was “de minimis.” 2014 WL 11310152, at *5. There were only “two references to race in Dr. Quijano’s testimony”—one during direct examination, the other on cross. Ibid. But when a jury hears expert testimony that expressly makes a defendant’s race directly pertinent on the question of life or death, the impact of that evidence cannot be measured simply by how much air time it received at trial or how many pages it occupies in the record. Some toxins can be deadly in small doses.

The State acknowledges, as it must, that introducing “race or ethnicity as evidence of criminality” can in some cases prejudice a defendant. Brief for Respondent 31. But it insists that this is not such a case, because Buck’s own counsel, not the prosecution, elicited the offending testimony. We are not convinced. In fact, the distinction could well cut the other way. A prosecutor is seeking a conviction. Jurors understand this and may reasonably be expected to evaluate the government’s evidence and arguments in light of its motivations. When a defendant’s own lawyer puts in the offending evidence, it is in the nature of an admission against interest, more likely to be taken at face value.

The effect of Dr. Quijano’s testimony on Buck’s sentencing cannot be dismissed as “de minimis.” Buck has demonstrated prejudice.

B

1

We now turn to the lower courts’ procedural holding: that Buck failed to demonstrate that he was entitled to have the judgment against him reopened under Rule 60(b)(6). We have held that a litigant seeking a COA must demonstrate that a procedural ruling barring relief is itself debatable among jurists of reason; otherwise, the appeal would not “deserve encouragement to proceed further.” Slack v. McDaniel,529 U. S. 473,484 (2000) (quoting Barefoot v. Estelle, 463 U. S. 880, n. 4 (1983)).

The Rule 60(b)(6) holding Buck challenges would be reviewed for abuse of discretion during a merits appeal, see 11 C. Wright, A. Miller, & M. Kane, Federal Practice and Procedure §2857 (3d ed. 2012), and the parties agree that the COA question is therefore whether a reasonable jurist could conclude that the District Court abused its discretion in declining to reopen the judgment. See Brief for Petitioner 54–57; Brief for Respondent 34.

Buck brought his Rule 60(b) motion under the Rule’s catchall category, subdivision (b)(6), which permits a court to reopen a judgment for “any other reason that justifies relief.” Rule 60(b) vests wide discretion in courts, but we have held that relief under Rule 60(b)(6) is available only in “extraordinary circumstances.” Gonzalez, 545 U. S., at 535. In determining whether extraordinary circumstances are present, a court may consider a wide range of factors. These may include, in an appropriate case, “the risk of injustice to the parties” and “the risk of undermining the public’s confidence in the judicial process.” Liljeberg v. Health Services Acquisition Corp.,486 U. S. 847–864 (1988).

In the circumstances of this case, the District Court abused its discretion in denying Buck’s Rule 60(b)(6) motion. The District Court’s conclusion that Buck “ha[d] failed to demonstrate that this case presents extraordinary circumstances” rested in large measure on itsdetermination that “the introduction of any mention of race”—though “ill[ ]advised at best and repugnant at worst”—played only a “de minimis” role in the proceeding. 2014 WL 11310152, at *5. The Fifth Circuit, for its part, failed even to mention the racial evidence in concludingthat Buck’s claim was “at least unremarkable as far as [ineffective assistance] claims go.” 623 Fed. Appx., at 673. But our holding on prejudice makes clear that Buck may have been sentenced to death in part because of his race. As an initial matter, this is a disturbing departure from a basic premise of our criminal justice system: Our law punishes people for what they do, not who they are. Dispensing punishment on the basis of an immutable characteristic flatly contravenes this guiding principle. As petitioner correctly puts it, “[i]t stretches credulity to characterize Mr. Buck’s [ineffective assistance of counsel] claim as run-of-the-mill.” Brief for Petitioner 57.

This departure from basic principle was exacerbated because it concerned race. “Discrimination on the basis of race, odious in all aspects, is especially pernicious in the administration of justice.” Rose v. Mitchell,443 U. S. 545,555 (1979). Relying on race to impose a criminal sanction “poisons public confidence” in the judicial process. Davis v. Ayala, 576 U. S. ___, ___ (2015) (slip op., at 28). It thus injures not just the defendant, but “the law as an institution, . . . the community at large, and . . . the democratic ideal reflected in the processes of our courts.” Rose, 443 U. S., at 556 (internal quotation marks omitted). Such concerns are precisely among those we have identified as supporting relief under Rule 60(b)(6). See Liljeberg, 486 U. S., at 864.

The extraordinary nature of this case is confirmed by what the State itself did in response to Dr. Quijano’s testimony. When the case of Victor Hugo Saldano came before this Court, Texas confessed error and consented to resentencing. The State’s response to Saldano’s petition for certiorari succinctly expressed the injustice Saldano had suffered: “the infusion of race as a factor for the jury to weigh in making its determination violated his constitutional right to be sentenced without regard to the color of his skin.” App. 306a.

The Attorney General’s public statement, issued shortly after we vacated the judgment in Saldano’s case, reflected this sentiment. It explained that the State had responded to Saldano’s troubling petition by conducting a “thorough audit” of criminal cases, finding six similar to Saldano’s “in which testimony was offered by Dr. Quijano that race should be a factor for the jury to consider.” Id., at 213a. The statement affirmed that “it is inappropriate to allow race to be considered as a factor in our criminal justice system.” Ibid. Consistent with this position—and to its credit—the State confessed error in the cases of five of the six defendants identified in the Attorney General’s statement, waiving all available procedural defenses and consenting to resentencing.

These were remarkable steps. It is not every day that a State seeks to vacate the sentences of five defendants found guilty of capital murder. But then again, these were—as the State itself put it at oral argument here—“extraordinary” cases. Tr. of Oral Arg. 41; see Buck v. Thaler, 565 U. S., at 1030 (Sotomayor, J., joined by Kagan, J., dissenting from denial of certiorari) (“Especially in light of the capital nature of this case and the express recognition by a Texas attorney general that the relevant testimony was inappropriately race charged, Buck has presented issues that ‘deserve encouragement to proceed further.’ ” (quoting Miller-El, 537 U. S., at 327)).

To be sure, the State has repeatedly attempted to justify its decision to treat Buck differently from the other five defendants identified in the Attorney General’s statement, including on asserted factual grounds that the State has been required to abjure. See Brief for Respondent 46, n. 10 (the State’s initial opposition to Buck’s habeas petition “erroneously” argued that Buck was treated differently because defense counsel, not the State, called Dr. Quijano as a witness; that was also true of two of the other defendants). The State continues its efforts before this Court, arguing that Buck’s was the only one of the six cases in which defense counsel, not the prosecution, first elicited Dr. Quijano’s opinion on race. See also post, at 8 (opinion of Thomas, J.).

But this is beside the point. The State’s various explanations for distinguishing Buck’s case have nothing to do with the Attorney General’s stated reasons for confessing error in Saldano and the cases acknowledged as similar. Regardless of which party first broached the subject, race was in all these cases put to the jury “as a factor . . . to weigh in making its determination.” App. 306a. The statement that “it is inappropriate to allow race to be considered as a factor in our criminal justice system” is equally applicable whether the prosecution or ineffective defense counsel initially injected race into the proceeding. Id., at 213a. The terms of the State’s announcement provide every reason for originally including Buck on the list of defendants situated similarly to Saldano, and no reason for later taking him off.

In opposition, the State reminds us of the importance of preserving the finality of judgments. Brief for Respondent 34. But the “whole purpose” of Rule 60(b) “is to make an exception to finality.” Gonzalez, 545 U. S., at 529. And in this case, the State’s interest in finality deserves little weight. When Texas recognized that the infusion of race into proceedings similar to Saldano’s warranted confession of error, it effectively acknowledged that the people of Texas lack an interest in enforcing a capital sentence obtained on so flawed a basis. In concluding that the value of finality does not demand that we leave the District Court’s judgment in place, we do no more than acknowledge what Texas itself recognized 17 years ago.

2

Our Rule 60(b)(6) analysis has thus far omitted one significant element. When Buck first sought federal ha-beas relief in 2004, Coleman barred the District Court from hearing his claim. Today, however, a claim of ineffective assistance of trial counsel defaulted in a Texas postconviction proceeding may be reviewed in federal court if state habeas counsel was constitutionally ineffective in failing to raise it, and the claim has “some merit.” Martinez, 566 U. S., at 14; see Trevino, 569 U. S., at ___ (slip op., at 13). Buck cannot obtain relief unless he is entitled to the benefit of this rule—that is, unless Martinez and Trevino, not Coleman, would govern his case were it reopened. If they would not, his claim would remain unreviewable, and Rule 60(b)(6) relief would be inappropriate. See 11 Wright & Miller, Federal Practice and Procedure §2857 (showing “a good claim or defense” is a precondition of Rule 60(b)(6) relief ).

Until merits briefing in this Court, both parties litigated this matter on the assumption that Martinez and Trevino would apply if Buck reopened his case. See Pet. for Cert. 27–28; Brief in Opposition 11–13; Amended Application for Certificate of Appealability and Brief in Support 26, Respondent-Appellee’s Opposition to Pet. for En Banc Rehearing 9–11, and Respondent’s Opposition to Application for Certificate of Appealability 15–17 in No. 14–70030 (CA5); Amended Response to Motion for Relief from Judgment in No. 4:04–cv–03965 (SD Tex.), pp. 11–13. But the State’s brief adopts a new position on this issue. The State now argues that those cases announced a “new rule” that, under Teague v. Lane,489 U. S. 288 (1989) (plurality opinion), does not apply retroactively to cases (like Buck’s) on collateral review. Brief for Respondent 38–40. Buck responds that Teague analysis applies only to new rules of criminal procedure that govern trial proceedings—not new rules of habeas procedure that govern collateral proceedings—and that the State has in any event waived its Teague argument. Reply Brief 20.

We agree that the argument has been waived. See Danforth v. Minnesota,552 U. S. 264,289 (2008) (“States can waive a Teague defense . . . by failing to raise it in a timely manner . . . .”). It was not advanced in District Court, before the Fifth Circuit, or in the State’s brief in opposition to Buck’s petition for certiorari. Although we may reach the issue in our discretion, we have observed before that a State’s failure to raise a Teague argument at the petition stage is particularly “significant” in deciding whether such an exercise of discretion is appropriate. Schiro v. Farley,510 U. S. 222–229 (1994). When “a legal issue appears to warrant review, we grant certiorari in the expectation of being able to decide that issue.” Id., at 229. If we were to entertain the State’s eleventh-hour Teague argument and find it persuasive, Buck’s Strickland and Rule 60(b)(6) contentions—the issues we thought worthy of review—would be insulated from our consideration. We therefore decline to reach the Teague question and conclude that Martinez and Trevino apply to Buck’s claim. We reach no broader determination concerning the application of these cases.

C

For the foregoing reasons, we conclude that Buck has demonstrated both ineffective assistance of counsel under Strickland and an entitlement to relief under Rule 60(b)(6). It follows that the Fifth Circuit erred in denying Buck the COA required to pursue these claims on appeal.

The judgment of the United States Court of Appeals for the Fifth Circuit is reversed, and the case is remanded for further proceedings consistent with this opinion.

It is so ordered.

SUPREME COURT OF THE UNITED STATES

_________________

No. 15–8049

_________________

DUANE EDWARD BUCK, PETITIONER v. LORIE DAVIS, DIRECTOR, TEXAS DEPARTMENT OF CRIMINAL JUSTICE, CORRECTIONAL INSTITUTIONS DIVISION

on writ of certiorari to the united states court of appeals for the fifth circuit

[February 22, 2017]

Justice Thomas, with whom Justice Alito joins, dissenting.

Having settled on a desired outcome, the Court bull-dozes procedural obstacles and misapplies settled law to justify it. But the majority’s focus on providing relief to petitioner in this particular case has at least one upside: Today’s decision has few ramifications, if any, beyond the highly unusual facts presented here. The majority leaves entirely undisturbed the black-letter principles of collateral review, ineffective assistance of counsel, and Rule 60(b)(6) law that govern day-to-day operations in federal courts.

I

In reversing the judgment below, the majority relies on three grounds: that the Fifth Circuit misapplied the standard for granting a certificate of appealability (COA); that the District Court erroneously rejected petitioner’s Sixth Amendment ineffective-assistance-of-counsel claim; and that the District Court abused its discretion in rejecting petitioner’s Federal Rule of Civil Procedure 60(b)(6) motion to reopen the court’s earlier judgment denying habeas relief. On each ground, the majority simply dis-agrees with the courts below over the application of the governing standard. The majority does not announce any new standards or suggest that the District Court or the Fifth Circuit applied the wrong standards altogether. See, e.g., ante, at 13 (“The court below phrased its determination in proper terms . . . ”). I agree with the majority that the courts below identified the correct standards for all three of these inquiries. Contrary to the majority’s conclusion, however, I would hold that they correctly applied those standards, too.

A

At the outset, the Court wrongly criticizes the Fifth Circuit for its application of the COA standard. A COA is warranted only if the district court’s ruling is “debatable amongst jurists of reason.” Miller-El v. Cockrell, 537 U. S. 322, 336 (2003) . To answer this question, a court must conduct a “general assessment” of the merits of a defendant’s claim. Ibid. The majority faults the Fifth Circuit for concluding outright that petitioner “ ‘has not shown extraordinary circumstances that would permit relief under [Rule] 60(b)(6).’ ” Ante, at 13 (quoting Buck v. Stephens, 623 Fed. Appx. 668, 669 (CA5 2015)). In the majority’s view, the existence of extraordinary circumstances represents an “ultimate merits determinatio[n] the panel should not have reached.” Ante, at 13–14. Instead, according to the majority, the panel should have limited itself to the threshold question whether the merits were debatable.

The majority’s criticism of the Fifth Circuit is misplaced. A court may grant a COA even if it might ultimately conclude that the underlying claim is meritless, so long as the claim is debatable. Miller-El, supra, at 336. But to deny a COA, a court must necessarily conclude that the claim is meritless. A reviewing court cannot determine that a claim is indisputably meritless (that is, nondebat-able) without first deciding that it is meritless. See Slack v. McDaniel, 529 U. S. 473, 484 (2000) (“Where a plain procedural bar is present and the district court is correct to invoke it to dispose of the case,” the claim is not debatable (emphasis added)); Weeks v. Angelone, 528 U. S. 225, 231, 234 (2000) (affirming denial of COA after determining that petitioner’s claim was meritless).

In concluding that petitioner’s claims were indisputably meritless as a prelude to denying the COA, the Fifth Circuit thus adopted the approach that courts must take in denying a COA. The majority might disagree with the conclusion that petitioner’s claims are indisputably meritless, but its criticism of the Fifth Circuit’s approach is most certainly misguided. The majority’s contrary approach would prevent a court of appeals from denying a COA in any case, an outcome that Congress and our precedents have plainly foreclosed. See 28 U. S. C. §2253(c)(2) (COA warranted only if petitioner makes “a substantial showing of the denial of a constitutional right”); Miller-El, supra, at 337 (The “issuance of a COA must not be pro forma or a matter of course”). The majority’s comment that “[u]ntil the prisoner secures a COA, the Court of Appeals may not rule on the merits of his case,” ante, at 12, should not be taken at face value.

In any event, after chastising the Court of Appeals for making an end run around the COA standard in order to reach the merits of petitioner’s Rule 60(b) claim, the Court does precisely that. See ante, at 26 (“[W]e conclude that Buck has demonstrated . . . an entitlement to relief under Rule 60(b)(6)”). Astonishingly, the Court also decides the merits of petitioner’s Sixth Amendment claim—an issue that was not even “addressed by the Fifth Circuit.” Ante, at 16; see ante, at 26 (“Buck has demonstrated . . . ineffective assistance of counsel under Strickland [v. Washington, 466 U. S. 668 (1984) ]”).

After reaching out to adjudicate the merits, the Court relies on its merits disposition to justify reversing the Fifth Circuit’s denial of a COA. Ante, at 26 (“It follows that the Fifth Circuit erred in denying Buck the COA required to pursue these claims on appeal” (emphasis added)). This unapologetic course reversal—made without so much as a hint of the irony—is striking. The majority also has things just backwards. It criticizes the Fifth Circuit for undertaking a merits inquiry to deny a COA (when such an inquiry is required) and then it conducts a merits inquiry to decide that petitioner’s claim is debat-able (when such an inquiry is inappropriate).

B

The Court’s application of the standard in Strickland v. Washington, 466 U. S. 668 (1984) , is similarly misguided. In particular, the Court erroneously finds that petitioner’s claim satisfies Strickland’s second prong, which requires a defendant to show that his counsel’s mistake materially prejudiced his defense. Prejudice exists only when correcting the alleged error would have produced a “substantial” likelihood of a different result. Harrington v. Richter, 562 U. S. 86 –112 (2011). Here, the sentence of death hinged on the jury’s finding that petitioner posed a threat of future dangerousness. Texas’ standard for making such a finding is not difficult to satisfy: “The facts of the offense alone may be sufficient to sustain the jury’s finding of future dangerousness,” and “[a] jury may also infer a defendant’s future dangerousness from evidence showing a lack of remorse.” Buntion v. State, 482 S. W. 3d 58, 66–67 (Tex. Crim. App. 2016).

The majority neglects even to mention the relevant legal standard in Texas, relying instead on rhetoric and speculation to craft a finding of prejudice. But the prosecution’s evidence of both the heinousness of petitioner’s crime and his complete lack of remorse was overwhelming. Accordingly, Dr. Quijano’s “de minimis” racial testimony, Buck v. Stephens, 2014 WL 11310152, *5 (SD Tex., Aug. 29, 2014), did not prejudice petitioner.

First, the facts leave no doubt that this crime was premeditated and cruel. The Court recites defense testimony describing the killing spree here as a “crime of passion,” ante, at 4 (internal quotation marks omitted), but the record belies that characterization. The rampage occurred at the home of Debra Gardner, petitioner’s ex-girlfriend. Prior to the shooting, petitioner called her house. His stepsister, Phyllis Taylor, answered, and petitioner asked to speak with Gardner. Gardner declined, and petitioner hung up. Petitioner then retrieved a shotgun and rifle, loaded both guns, and drove 28 miles to Gardner’s house. Upon arrival, he broke down the door and opened fire without provocation. The shooting did not occur in the heat of the moment.

In addition to describing this as a crime of passion, the majority also parrots defense testimony that petitioner’s violence was limited to “the context of romantic relationships.” Ante, at 18. But this assertion is also quite wrong. Upon entering Gardner’s house, petitioner first shot at an acquaintance, Harold Ebnezer. He next approached his stepsister, Taylor, who was seated on the couch. He said, “ ‘I’m going to shoot your ass too.’ ” App. 82a. She begged him, “ ‘Duane, please don’t shoot me. I’m your sister. I don’t deserve to be shot. Remember I do have children.’ ” Id., at 83a. Petitioner ignored her pleas, placed the gun on her chest, and shot her. Petitioner does not claim that he was in a romantic relationship with either Ebnezer or Taylor.

After shooting Taylor, petitioner cornered one of Gardner’s friends, Kenneth Butler, and shot him, as well. He then exited the house and chased Gardner into the middle of the street. She turned to him and pleaded, “ ‘Please don’t shoot me. Please don’t shoot me. Why are you doing this in front of my kids?’ ” Id., at 104a. Her son, Devon, watched from the sidewalk. Her daughter, Shennel, begged petitioner to spare her mother and even attempted to restrain him. Petitioner pointed the gun at Gardner and said, “ ‘I’m going to shoot you. I’m going to shoot your A[ss].’ ” Id., at 117a. He then did so. The flight path of the bullet suggests that Gardner was on her knees when petitioner shot her.

Second, the evidence of petitioner’s lack of remorse, largely ignored by the majority, is startling. After shooting Gardner, petitioner walked back to his car and placed the firearms in the trunk. He then returned to taunt Gardner where she lay mortally wounded and bleeding in the street. He said, “ ‘It ain’t funny now. You ain’t laughing now.’ ” Id., at 106a. Police arrived shortly thereafter and arrested him. In the patrol car, petitioner was “laughing and joking and taunting.” Id., at 71a. He continued to smile and laugh during the drive to the police station. When one of the officers informed petitioner that he did not find the situation humorous, petitioner replied that “ ‘[t]he bitch got what she deserved.’ ” Id., at 135a. He remained happy and upbeat for the remainder of the drive, even commenting that he was going to heaven because God had already forgiven him.

C

Finally, the majority incorrectly concludes that the District Court erred in denying petitioner’s motion under Rule 60(b)(6), which permits district courts to reopen otherwise final judgments only in “extraordinary circumstances.” Ackermann v. United States, 340 U. S. 193, 199 (1950) . Although the majority pays lip service to the fact that the District Court’s decision on this point is subject to “limited and deferential appellate review,” Gonzalez v. Crosby, 545 U. S. 524, 535 (2005) ; see ante, at 20–21, it proceeds to conduct a de novo review. Indeed, the majority references the District Court’s analysis only once in the entire section of its opinion addressing Rule 60(b)(6). But the question is not whether this Court thinks the circumstances are extraordinary; the question is whether reasonable jurists would debate that the District Court abused its discretion in reaching the equitable, highly factbound conclusion that they are not. Particularly in light of our admonition that such circumstances “will rarely occur in the habeas context,” 545 U. S., at 535, I think it is not debatable that the District Court acted within its discretion in denying Rule 60(b)(6) relief here.

In reversing the Fifth Circuit, the centerpiece of the Court’s analysis is its observation that “ ‘[d]iscrimination on the basis of race, odious in all aspects, is especially pernicious in the administration of justice.’ ” Ante, at 22 (quoting Rose v. Mitchell, 443 U. S. 545, 555 (1979) ). I agree that racial classifications are categorically impermissible under the Equal Protection Clause—but peti-tioner is not raising an equal protection claim. The many precedents the Court cites interpreting that Clause are therefore beside the point. Instead, petitioner is claiming that his case presents extraordinary circumstances under Rule 60(b)(6). The majority identifies no precedents regarding race in the Rule 60(b)(6) context. And certainly nothing in the text or history of Rule 60(b)(6)—unlike the text and history of the Equal Protection Clause—suggests that race-based claims demand unique solicitude in this context. At the very least, the District Court did not abuse its discretion in determining that they do not.[1]

In a similar vein, the majority suggests that the use of race in petitioner’s capital proceeding injured the public’s confidence in the integrity of our judicial system. Ante, at 22. This argument cannot be squared with the District Court’s finding that the challenged racial testimony was “de minimis.” 2014 WL 11310152, at *5. It also ignores the fact that petitioner’s own counsel elicited the testi-mony. The majority obscures this point by citing cases concerning alleged racial discrimination by an agent of the state. See ante, at 22 (citing, e.g., Davis v. Ayala, 576 U. S. ___ (2015) (addressing the prosecutor’s use of peremptory strikes and holding that any constitutional error was harmless)). There, the injury to public confidence derives from the fact that the government itself is discriminating against the defendant. The same cannot be said, however, when defense counsel introduces harmful testimony or makes a bad strategic choice.[2]

In conjunction with its observations about race, the Court notes that the Texas attorney general, in response to similar testimony from Dr. Quijano in another case, issued a press release decrying the use of race in the justice system and subsequently waived all procedural obstacles to resentencing in several cases in which Dr. Quijano testified. But Texas had good reason for treating this case differently from the others. Of those cases, this is the only one where “it can be said that the responsibility for eliciting the offensive testimony lay squarely with the defense.” Buck v. Thaler, 565 U. S. 1022, 1025 (2011) (Alito, J., statement respecting denial of certiorari).

Lastly, the Court belittles Texas’ claimed interest in finality. In the majority’s view, Texas effectively forfeited its finality interest when it waived its procedural defenses in purportedly similar cases. See ante, at 24. But Texas did not waive its procedural defenses in this case, and, in any event, Texas is not alone in possessing an interest in the finality of petitioner’s sentence. Society at large has the same interest. Finality advances values “essential to the operation of our criminal justice system.” Teague v. Lane, 489 U. S. 288, 309 (1989) (plurality opinion). It promotes the law’s deterrent effect; it provides peace of mind to a wrongdoer’s victims; it promotes public confidence in the justice system; it conserves limited public resources; and it ensures the clarity of legal rights and statuses.

The Court’s finality analysis also ignores the lengthy passage of time (nearly eight years) between the District Court’s original rejection of habeas relief in this case and petitioner’s filing of the instant Rule 60(b)(6) motion. See Gonzalez, 545 U. S., at 542, n. 4 (Stevens, J., dissenting) (“In cases where significant time has elapsed between a habeas judgment and the relevant change in procedural law, it would be within a district court’s discretion to leave such a judgment in repose”). Permitting a defendant to file a Rule 60(b) motion years after the fact functionally eviscerates the statute of limitations contained in the Antiterrorism and Effective Death Penalty Act of 1996, thereby undermining its purpose of “lend[ing] finality to state court judgments within a reasonable time.” Day v. McDonough, 547 U. S. 198 –206 (2006) (internal quotation marks omitted).

II

Despite its errors, today’s opinion should have little effect on the broader law, for two reasons. For one thing, the Court’s reasoning is highly factbound, and the facts presented here are unlikely to arise again. For another, although the majority misapplies settled principles, it does not purport to actually alter any of those principles.

A

This is an unusual case, and the majority’s single-minded focus on according relief to this petitioner on these facts naturally limits the reach of its decision. In cases presenting different facts, today’s decision will provide little guidance. The Court’s ultimate conclusion relies on the convergence of three critical factors that will rarely, if ever, recur. See ante, at 19 (describing this case as involving an “unusual confluence of factors” that together create a “perfect storm”).

First, the Court places special weight on the fact that this is a capital case. See, e.g., ante, at 23 (noting “the capital nature of this case” as a factor favoring Rule 60(b)(6) relief (internal quotation marks omitted)). Second, the Court notes that the testimony at issue was expressly racial and suggested that petitioner was more deserving of the death penalty because he is black. See, e.g., ante, at 21 (“Buck may have been sentenced to death in part because of his race. . . . [T]his is a disturbing departure from a basic premise of our criminal justice system”).[3] Third, the Court explains that the state attorney general took the “remarkable steps,” ante, at 23, of publicly declaring that Dr. Quijano’s testimony in this case was inappropriate and waiving all procedural defenses to resentencing in similar cases. See, e.g., ante, at 22 (“The extraordinary nature of this case is confirmed by what the State itself did in response to Dr. Quijano’s testimony”).

B

The effect of today’s decision is also limited for a second reason. Although the majority misapplies many exist-ing doctrines, it refrains from announcing any new principles of law. In particular, it leaves untouched—andcourts should accordingly continue to apply asusual—established principles governing collateral review, ineffective-assistance-of-counsel claims, and Rule 60(b)(6) motions.

At the outset, the opinion leaves intact Miller-El’s well-worn COA standard for habeas petitions: Courts of appeals should deny applications when the district court’s ruling is not “debatabl[y]” wrong. 537 U. S., at 336; see ante, at 13. “A prisoner seeking a COA must prove something more than the absence of frivolity or the existence of mere good faith on his or her part.” 537 U. S., at 338 (internal quotation marks omitted). Courts have substantial discretion in deciding how to structure this inquiry. See ante, at 15 (“We do not mean to specify what procedures may be appropriate in every case”).

The Court also reaffirms the “ ‘highly deferential’ ” character of the Strickland standard. Harrington, 562 U. S., at 105 (quoting Strickland, 466 U. S., at 689); see, e.g., ante, at 16 (first prong of Strickland “sets a high bar”). Courts applying Strickland must respect “the constitutionally protected independence of counsel and the wide latitude counsel must have in making tactical decisions.” Cullen v. Pinholster, 563 U. S. 170, 195 (2011) (internal quotation marks and alteration omitted). Counsel should be “strongly presumed to have rendered adequate assistance andmade all significant decisions in the exercise of reasonable professional judgment.” Id., at 189 (internal quotation marks omitted). And a defendant can show prejudice only if the likelihood of a different result would have been “substantial, not just conceivable.” Harrington, supra, at 112.

Perhaps most significantly, the test for reopening judgments under Rule 60(b)(6) remains the same. A “ ‘very strict interpretation of Rule 60(b) is essential if the finality of judgments is to be preserved.’ ” Gonzalez, 545 U. S., at 535 (quoting Liljeberg v. Health Services Acquisition Corp., 486 U. S. 847, 873 (1988) (Rehnquist, C. J., dissenting)). A district court may grant relief only if the movant can show “extraordinary circumstances,” which “will rarely occur in the habeas context.” 545 U. S., at 535. (internal quotation marks omitted); see ante, at 10. A change in law alone is not enough. See 545 U. S., at 536–537. And a district court’s decision to deny relief is subject to only “limited and deferential” appellate review. Id.,at 535.

Although petitioner argues that the change in law effected by Martinez v. Ryan, 566 U. S. 1 (2012) , and Tre-vino v. Thaler, 569 U. S. ___ (2013), is central to the Rule 60(b)(6) inquiry, see Brief for Petitioner 56; Reply Brief 19, the Court does not even count those decisions in its tally of extraordinary circumstances. Instead, it treats a potentially viable Martinez claim as a “precondition” to relief. Ante, at 25. This makes sense: Unless petitioner has the ability to invoke Martinez and Trevino, reopening the judgment would be futile. For those in petitioner’s position, the absence of a potentially valid Martinez claim is disqualifying, but the presence of one does nothing to demonstrate the existence of extraordinary circumstances.

III

Finally, the Court’s opinion does not require the lower courts to reflexively accord relief to petitioner on remand. In order to succeed under Martinez and Trevino, petitioner must establish that his state habeas counsel was constitutionally ineffective for failing to raise a Strickland claim as to his trial counsel. Today’s decision does not address that showing, and the court on remand should not treat it as a foregone conclusion.

I respectfully dissent.