delivered the opinion of the Court.
The writ of habeas corpus stands as a safeguard against imprisonment of those held in violation of the law. Judges must be vigilant and independent in reviewing petitions for the writ, a commitment that entails substantial judicial resources. Those resources are diminished and misspent, however, and confidence in the writ and the law it vindicates *92undermined, if there is judicial disregard for the sound and established principles that inform its proper issuance. That judicial disregard is inherent in the opinion of the Court of Appeals for the Ninth Circuit here under review. The Court of Appeals, in disagreement with the contrary conclusions of the Supreme Court of the State of California and of a United States District Court, ordered habeas corpus relief granted to set aside the conviction of Joshua Richter, respondent here. This was clear error.
Under 28 U. S. C. § 2254(d), the availability of federal ha-beas relief is limited with respect to claims previously “adjudicated on the merits” in state-court proceedings. The first inquiry this case presents is whether that provision applies when state-court relief is denied without an accompanying statement of reasons. If it does, the question is whether the Court of Appeals adhered to the statute’s terms, in this ease as it relates to ineffective-assistance claims judged by the standard set forth in Strickland v. Washington, 466 U. S. 668 (1984). A second case decided today, Premo v. Moore, post, p. 115, presents similar issues. Here, as in that case, it is necessary to reverse the Court of Appeals for failing to accord required deference to the decision of a state court.
I
It is necessary to begin by discussing the details of a crime committed more than a decade and a half ago.
A
Sometime after midnight on December 20, 1994, sheriff’s deputies in Sacramento County, California, arrived at the home of a drug dealer named Joshua Johnson. Hours before, Johnson had been smoking marijuana in the company of Richter and two other men, Christian Branscombe and Patrick Klein. When the deputies arrived, however, they found only Johnson and Klein. Johnson was hysterical and covered in blood. Klein was lying on a couch in Johnson’s living *93room, unconscious and bleeding. Klein and Johnson each had been shot twice. Johnson recovered; Klein died of his wounds.
Johnson gave investigators this account: After falling asleep, he awoke to find Richter and Branseombe in his bedroom, at which point Branseombe shot him. Johnson heard more gunfire in the living room and the sound of his assailants leaving. He got up, found Klein bleeding on the living •room couch, and called 911. A gun safe, a pistol, and $6,000 cash, all of which had been in the bedroom, were missing.
Evidence at the scene corroborated Johnson’s account. Investigators found spent shell casings in the bedroom (where Johnson said he had been shot) and in the living room (where Johnson indicated Klein had been shot). In the living room there were two casings, a .32 caliber and a .22 caliber. One of the bullets recovered from Klein’s body was a .32 and the other was a .22. In the bedroom there were two more casings, both .32 caliber. In addition detectives found blood spatter near the living room couch and bloodstains in the bedroom. Pools of blood had collected in the kitchen and the doorway to Johnson’s bedroom. Investigators took only a few blood samples from the crime scene. One was from a blood splash on the wall near the bedroom doorway, but no sample was taken from the doorway blood pool itself.
Investigators searched Richter’s residence and found Johnson’s gun safe, two boxes of .22-caliber ammunition, and a gun magazine loaded with cartridges of the same brand and type as the boxes. A ballistics expert later concluded the .22-caliber bullet that struck Klein and the .22-caliber shell found in the living room matched the ammunition found in Richter’s home and bore markings consistent with the model of gun for which the magazine was designed.
Richter and Branseombe were arrested. At first Richter denied involvement. He would later admit taking Johnson’s pistol and disposing of it and of the .32-caliber weapon Brans-combe used to shoot Johnson and Hein. Richter’s counsel *94produced Johnson’s missing pistol, but neither of the guns used to shoot Johnson and Klein was found.
B
Branscombe and Richter were tried together on charges of murder, attempted murder, burglary, and robbery. Only Richter’s case is presented here.
The prosecution built its case on Johnson’s testimony and on circumstantial evidence. Its opening statement took note of the shell casings found at the crime scene and the ammunition and gun safe found at Richter's residence. Defense counsel offered explanations for the circumstantial evidence and derided Johnson as a drug dealer, a paranoid, and a trigger-happy gun fanatic who had drawn a pistol on Brans-combe and Richter the last time he had seen them. And there were inconsistencies in Johnson’s story. In his 911 call, for instance, Johnson first said there were four or five men who had broken into his house, not two; and in the call he did not identify Richter and Branscombe among the intruders.
Blood evidence does not appear to have been part of the prosecution's planned case prior to trial, and investigators had not analyzed the few blood samples taken from the crime scene. But the opening statement from the defense led the prosecution to alter its approach. Richter's attorney outlined the theory that Branscombe had fired on Johnson in self-defense and that Klein had been killed not on the living room couch but in the crossfire in the bedroom doorway. Defense counsel stressed deficiencies in the investigation, including the absence of forensic support for the prosecution's version of events.
The prosecution took steps to adjust to the counterattack now disclosed. Without advance notice and over the objection of Richter’s attorney, one of the detectives who investigated the shootings testified for the prosecution as an expert in blood pattern evidence. He concluded it was unlikely *95Klein had been shot outside the living room and then moved to the couch, given the patterns of blood on Klein’s face, as well as other evidence including “high velocity” blood spatter near the couch consistent with the location of a shooting. The prosecution also offered testimony from a serologist. She testified the blood sample taken near the pool by the bedroom door could be Johnson’s but not Klein’s.
Defense counsel’s cross-examination probed weaknesses in the testimony of these two witnesses. The detective who testified on blood patterns acknowledged that his inferences were imprecise, that it was unlikely Klein had been lying down on the couch when shot, and that he could not say the blood in the living room was from either of Klein’s wounds. Defense counsel elicited from the serologist a concession that she had not tested the bedroom blood sample for cross-contamination. She said that if the year-old sample had degraded, it would be difficult to tell whether blood of Klein’s type was also present in the sample.
For the defense, Richter’s attorney called seven witnesses. Prominent among these was Richter himself. Richter testified he and Branscombe returned to Johnson’s house just before the shootings in order to deliver something to one of Johnson’s roommates. By Richter’s account, Branscombe entered the house alone while Richter waited in the driveway; but after hearing screams and gunshots, Richter followed inside. There he saw Klein lying not on the couch but in the bedroom doorway, with Johnson on the bed and Branscombe standing in the middle of the room. According to Richter, Branscombe said he shot at Johnson and Klein after they attacked him. Other defense witnesses provided some corroboration for Richter’s story. His former girlfriend, for instance, said she saw the gun safe at Richter’s house shortly before the shootings.
The jury returned a verdict of guilty on all charges. Richter was sentenced to life without parole. On appeal, his conviction was affirmed. People v. Branscombe, 72 Cal. *96Rptr. 2d 773 (App. 1998) (officially depublished). The California Supreme Court denied a petition for review, People v. Branscombe, No. S069751, 1998 Cal. LEXIS 4252 (June 24, 1998), and Richter did not file a petition for certiorari with this Court. His conviction became final.
C
Richter later petitioned the California Supreme Court for a writ of habeas corpus. He asserted a number of grounds for relief, including ineffective assistance of counsel. As relevant here, he claimed his counsel was deficient for failing to present expert testimony on serology, pathology, and blood spatter patterns, testimony that, he argued, would disclose the source of the blood pool in the bedroom doorway. This, he contended, would bolster his theory that Johnson had moved Klein to the couch.
He offered affidavits from three types of forensic experts. First, he provided statements from two blood serologists who said there was a possibility Klein’s blood was intermixed with blood of Johnson’s type in the sample taken from near the pool in the bedroom doorway. Second, he provided a statement from a pathologist who said the blood pool was too large to have come from Johnson given the nature of his wounds and his own account of his actions while waiting for the police. Third, he provided a statement from an expert in bloodstain analysis who said the absence of “a large number of satellite droplets” in photographs of the area around the blood in the bedroom doorway was inconsistent with the blood pool coming from Johnson as he stood in the doorway. App. 118. Richter argued this evidence established the possibility that the blood in the bedroom doorway came from Klein, not Johnson. If that were true, he argued, it would confirm his account, not Johnson’s. The California Supreme Court denied Richter’s petition in a one-sentence summary order. In re Richter, No. S082167 (Mar. 28, 2001), App. to *97Pet. for Cert. 22a. Richter did not seek certiorari from this Court.
After the California Supreme Court issued its summary order denying relief, Richter filed a petition for habeas corpus in United States District Court for the Eastern District of California. He reasserted the claims in his state petition. The District Court denied his petition, and a three-judge panel of the Court of Appeals for the Ninth Circuit affirmed. Richter v. Hickman, 521 F. 3d 1222 (2008). The Court of Appeals granted rehearing en banc and reversed the District Court’s decision. Richter v. Hickman, 578 F. 3d 944 (2009).
As a preliminary matter, the Court of Appeals questioned whether 28 U. S. C. § 2254(d) was applicable to Richter’s petition, since the California Supreme Court issued only a summary denial when it rejected Ms Strickland claims; but it determined the California decision was unreasonable in any event and that Richter was entitled to relief. The court held Richter’s trial counsel was deficient for failing to consult experts on blood evidence in determining and pursuing a trial strategy and in preparing to rebut expert evidence the prosecution might — and later did — offer. Four judges dissented from the en banc decision.
We granted certiorari. 559 U. S. 935 (2010).
II
The statutory authority of federal courts to issue habeas corpus relief for persons in state custody is provided by 28 U. S. C. § 2254, as amended by the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA). The text of § 2254(d) states:
“An application for a writ of habeas corpus on behalf of a person in custody pursuant to the judgment of a State court shall not be granted with respect to any claim that was adjudicated on the merits in State court proceedings unless the adjudication of the claim—
*98“(1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or
“(2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.”
As an initial matter, it is necessary to decide whether § 2254(d) applies when a state court’s order is unaccompanied by an opinion explaining the reasons relief has been denied.
By its terms § 2254(d) bars relitigation of any claim “adjudicated on the merits” in state court, subject only to the exceptions in §§ 2254(d)(1) and (2). There is no text in the statute requiring a statement of reasons. The statute refers only to a “decision,” which resulted from an “adjudication.” As every Court of Appeals to consider the issue has recognized, determining whether a state court’s decision resulted from an unreasonable legal or factual conclusion does not require that there be an opinion from the state court explaining the state court’s reasoning. See Chadwick v. Janecka, 312 F. 3d 597, 605-606 (CA3 2002); Wright v. Secretary for Dept. of Corrections, 278 F. 3d 1245, 1253-1254 (CA11 2002); Sellan v. Kuhlman, 261 F. 3d 303, 311-312 (CA2 2001); Bell v. Jarvis, 236 F. 3d 149, 158-162 (CA4 2000) (en banc); Harris v. Stovall, 212 F. 3d 940, 943, n. 1 (CA6 2000); Aycox v. Lytle, 196 F. 3d 1174, 1177-1178 (CA10 1999); James v. Bowersox, 187 F. 3d 866, 869 (CA8 1999). And as this Court has observed, a state court need not cite or even be aware of our cases under § 2254(d). Early v. Packer, 537 U. S. 3, 8 (2002) (per curiam). Where a state court’s decision is unaccompanied by an explanation, the habeas petitioner’s burden still must be met by showing there was no reasonable basis for the state court to deny relief. This is so whether or not the state court reveals which of the elements in a multipart claim it found insufficient, for § 2254(d) applies when a “claim,” not a component of one, has been adjudicated.
*99There is no merit to the assertion that compliance with § 2254(d) should be excused when state courts issue summary rulings because applying § 2254(d) in those cases will encourage state courts to withhold explanations for their decisions. Opinion-writing practices in state courts are influenced by considerations other than avoiding scrutiny by collateral attack in federal court. Cf. In re Robbins, 18 Cal. 4th 770, 778, n. 1, 959 P. 2d 311, 316, n. 1 (1998) (state procedures limiting habeas are “a means of protecting the integrity of our own appeal and habeas corpus process,” rather than a device for “insulating our judgments from federal court review” (emphasis deleted)). At the same time, requiring a statement of reasons could undercut state practices designed to preserve the integrity of the case-law tradition. The issuance of summary dispositions in many collateral attack cases can enable a state judiciary to concentrate its resources on the cases where opinions are most needed. See Brief for California Attorneys for Criminal Justice et al. as Amici Curiae 8 (noting that the California Supreme Court disposes of close to 10,000 cases a year, including more than 3,400 original habeas corpus petitions).
There is no merit either in Richter’s argument that § 2254(d) is inapplicable because the California Supreme Court did not say it was adjudicating his claim “on the merits.” The state court did not say it was denying the claim for any other reason. When a federal claim has been presented to a state court and the state court has denied relief, it may be presumed that the state court adjudicated the claim on the merits in the absence of any indication or state-law procedural principles to the contrary. Cf. Harris v. Reed, 489 U. S. 255, 265 (1989) (presumption of a merits determination when it is unclear whether a decision appearing to rest on federal grounds was decided on another basis).
The presumption may be overcome when there is reason to think some other explanation for the state court’s decision *100is more likely. See, e. g., Ylst v. Nunnemaker, 501 U. S. 797, 803 (1991). Richter, however, does not make that showing. He mentions the theoretical possibility that the members of the California Supreme Court may not have agreed on the reasons for denying his petition. It is pure speculation, however, to suppose that happened in this case. And Richter’s assertion that the mere possibility of a lack of agreement prevents any attribution of reasons to the state court’s decision is foreclosed by precedent. See ibid.
As has been noted before, the California courts or Legislature can alter the State’s practices or elaborate more fully on their import. Evans v. Chavis, 546 U. S. 189, 197, 199 (2006). But that has not occurred here. This Court now holds and reconfirms that § 2254(d) does not require a state court to give reasons before its decision can be deemed to have been “adjudicated on the merits.” Richter has failed to show that the California Supreme Court’s decision did not involve a determination of the merits of his claim. Section 2254(d) applies to his petition.
Ill
Federal habeas relief may not be granted for claims subject to § 2254(d) unless it is shown that the earlier state court’s decision “was contrary to” federal law then clearly established in the holdings of this Court, § 2254(d)(1); Williams v. Taylor, 529 U. S. 362, 412 (2000); or that it “involved an unreasonable application of” such law, § 2254(d)(1); or that it “was based on an unreasonable determination of the facts” in light of the record before the state court, § 2254(d)(2).
The Court of Appeals relied on the second of these exceptions to §2254(d)’s relitigation bar, the exception in § 2254(d)(1) permitting relitigation where the earlier state decision resulted from an “unreasonable application of” clearly established federal law. In the view of the Court of Appeals, the California Supreme Court’s decision on Richter’s ineffective-assistance claim unreasonably applied the *101holding in Strickland. The Court of Appeals’ lengthy opinion, however, discloses an improper understanding of § 2254(d)’s unreasonableness standard and of its operation in the context of a Strickland claim.
The pivotal question is whether the state court’s application of the Strickland standard was unreasonable. This is different from asking whether defense counsel’s performance fell below Strickland’s standard. Were that the inquiry, the analysis would be no different than if, for example, this Court were adjudicating a Strickland claim on direct review of a criminal conviction in a United States district court. Under AEDPA, though, it is a necessary premise that the two questions are different. For purposes of § 2254(d)(1), “an unreasonable application of federal law is different from an incorrect application of federal law.” Williams, supra, at 410. A state court must be granted a deference and latitude that are not in operation when the case involves review under the Strickland standard itself.
A state court’s determination that a claim lacks merit precludes federal habeas relief so long as “fairminded jurists could disagree” on the correctness of the state court’s decision. Yarborough v. Alvarado, 541 U. S. 652, 664 (2004). And as this Court has explained, “[E]valuating whether a rule application was unreasonable requires considering the rule’s specificity. The more general the rule, the more leeway courts have in reaching outcomes in case-by-case determinations.” Ibid. “[I]t is not an unreasonable application of clearly established Federal law for a state court to decline to apply a specific legal rule that has not been squarely established by this Court.” Knowles v. Mirzayance, 556 U. S. 111, 122 (2009) (internal quotation marks omitted).
Here it is not apparent how the Court of Appeals’ analysis would have been any different without AEDPA. The court explicitly conducted a de novo review, 578 F. 3d, at 952; and after finding a Strickland violation, it declared, without further explanation, that the “state court’s decision to the con*102trary constituted an unreasonable application of Strickland.” 578 F. 3d, at 969. AEDPA demands more. Under § 2254(d), a habeas court must determine what arguments or theories supported or, as here, could have supported, the state court’s decision; and then it must ask whether it is possible fair-minded jurists could disagree that those arguments or theories are inconsistent with the holding in a prior decision of this Court. The opinion of the Court of Appeals all but ignored “the only question that matters under § 2254(d)(1).” Lockyer v. Andrade, 538 U. S. 63, 71 (2003).
The Court of Appeals appears to have treated the unreasonableness question as a test of its confidence in the result it would reach under de novo review: Because the Court of Appeals had little doubt that Richter’s Strickland claim had merit, the Court of Appeals concluded the state court must have been -unreasonable in rejecting it. This analysis overlooks arguments that would otherwise justify the state court’s result and ignores further limitations of § 2254(d), including its requirement that the state court’s decision be evaluated according to the precedents of this Court. See Renico v. Lett, 559 U. S. 766, 778-779 (2010). It bears repeating that even a strong case for relief does not mean the state court’s contrary conclusion was unreasonable. See Lockyer, supra, at 75.
If this standard is difficult to meet, that is because it was meant to be. As amended by AEDPA, § 2254(d) stops short of imposing a complete bar on federal-court relitigation of claims already rejected in state proceedings. Cf. Felker v. Turpin, 518 U. S. 651, 664 (1996) (discussing AEDPA’s “modified res judicata rule” under § 2244). It preserves authority to issue the writ in cases where there is no possibility fair-minded jurists could disagree that the state court’s decision conflicts with this Court’s precedents. It goes no further. Section 2254(d) reflects the view that habeas corpus is a “guard against extreme malfunctions in the state criminal justice systems,” not a substitute for ordinary error corree*103tion through appeal. Jackson v. Virginia, 443 U. S. 307, 332, n. 5 (1979) (Stevens, J., concurring in judgment). As a condition for obtaining habeas corpus from, a federal court, a state prisoner must show that the state court’s ruling on the claim being presented in federal court was so lacking in justification that there was an error well understood and comprehended in existing law beyond any possibility for fair-minded disagreement.
The reasons for this approach are familiar. “Federal ha-beas review of state convictions frustrates both the States’ sovereign power to punish offenders and their good-faith attempts to honor constitutional rights.” Calderon v. Thompson, 523 U. S. 538, 555-556 (1998) (internal quotation marks omitted). It “disturbs the State’s significant interest in repose for concluded litigation, denies society the right to punish some admitted offenders, and intrudes on state sovereignty to a degree matched by few exercises of federal judicial authority.” Reed, 489 U. S., at 282 (Kennedy, J., dissenting).
Section 2254(d) is part of the basic structure of federal habeas jurisdiction, designed to confirm that state courts are the principal forum for asserting constitutional challenges to state convictions. Under the exhaustion requirement, a habeas petitioner challenging a state conviction must first attempt to present his claim in state court. 28 U. S. C. § 2254(b). If the state court rejects the claim on procedural grounds, the claim is barred in federal court unless one of the exceptions to the doctrine of Wainwright v. Sykes, 433 U. S. 72, 82-84 (1977), applies. And if the state court denies the claim on the merits, the claim is barred in federal court unless one of the exceptions to § 2254(d) set out in §§ 2254(d)(1) and (2) applies. Section 2254(d) thus complements the exhaustion requirement and the doctrine of procedural bar to ensure that state proceedings are the central process, not just a preliminary step for a later federal habeas proceeding, see id., at 90.
*104Here, however, the Court of Appeals gave § 2254(d) no operation or function in its reasoning. Its analysis illustrates a lack of deference to the state court’s determination and an improper intervention in state criminal processes, contrary to the purpose and mandate of AEDPA and to the now well-settled meaning and function of habeas corpus in the federal system.
IV
The conclusion of the Court of Appeals that Richter demonstrated an unreasonable application by the state court of the Strickland standard now must be discussed. To have been entitled to relief from the California Supreme Court, Richter had to show both that his counsel provided deficient assistance and that there was prejudice as a result.
To establish deficient performance, a person challenging a conviction must show that “counsel’s representation fell below an objective standard of reasonableness.” 466 U. S., at 688. A court considering a claim of ineffective assistance must apply a “strong presumption” that counsel’s representation was within the “wide range” of reasonable professional assistance. Id., at 689. The challenger’s burden is to show “that counsel made errors so serious that counsel was not functioning as the 'counsel’ guaranteed the defendant by the Sixth Amendment.” Id., at 687.
With respect to prejudice, a challenger must demonstrate “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome.” Id., at 694. It is not enough “to show that the errors had some conceivable effect on the outcome of the proceeding.” Id., at 693. Counsel’s errors must be “so serious as to deprive the defendant of a fair trial, a trial whose result is reliable.” Id., at 687.
*105“Surmounting Strickland’s high bar is never an easy task.” Padilla v. Kentucky, 559 U. S. 356, 371 (2010). An ineffective-assistance claim can function as a way to escape rules of waiver and forfeiture and raise issues not presented at trial, and so the Strickland standard must be applied with scrupulous care, lest “intrusive post-trial inquiry” threaten the integrity of the very adversary process the right to counsel is meant to serve. Strickland, 466 U. S., at 689-690. Even under de novo review, the standard for judging counsel’s representation is a most deferential one. Unlike a later reviewing court, the attorney observed the relevant proceedings, knew of materials outside the record, and interacted with the client, with opposing counsel, and with the judge. It is “all too tempting” to “second-guess counsel’s assistance after conviction or adverse sentence.” Id., at 689; see also Bell v. Cone, 535 U. S. 685, 702 (2002); Lockhart v. Fretwell, 506 U. S. 364, 372 (1993). The question is whether an attorney’s representation amounted to incompetence under “prevailing professional norms,” not whether it deviated from best practices or most common custom. Strickland, 466 U. S., at 690.
Establishing that a state court’s application of Strickland was unreasonable under § 2254(d) is all the more difficult. The standards created by Strickland and § 2254(d) are both “highly deferential,” id., at 689; Lindh v. Murphy, 521 U. S. 320, 333, n. 7 (1997), and when the two apply in tandem, review is “doubly” so, Knowles, 556 U. S., at 123. The Strickland standard is a general one, so the range of reasonable applications is substantial. 556 U. S., at 123. Federal ha-beas courts must guard against the danger of equating unreasonableness under Strickland with unreasonableness under § 2254(d). When § 2254(d) applies, the question is not whether counsel’s actions were reasonable. The question is whether there is any reasonable argument that counsel satisfied Strickland’s deferential standard.
*106A
With respect to defense counsel’s performance, the Court of Appeals held that because Richter’s attorney had not consulted forensic blood experts or introduced expert evidence, the California Supreme Court could not reasonably have concluded counsel provided adequate representation. This conclusion was erroneous.
1
The Court of Appeals first held that Richter’s attorney rendered constitutionally deficient service because he did not consult blood evidence experts in developing the basic strategy for Richter’s defense or offer their testimony as part of the principal case for the defense. Strickland, however, permits counsel to “make a reasonable decision that makes particular investigations unnecessary.” 466 U. S., at 691. It was at least arguable that a reasonable attorney could decide to forgo inquiry into the blood evidence in the circumstances here.
Criminal cases will arise where the only reasonable and available defense strategy requires consultation with experts or introduction of expert evidence, whether pretrial, at trial, or both. There are, however, “countless ways to provide effective assistance in any given case. Even the best criminal defense attorneys would not defend a particular client in the same way.” Id., at 689. Rare are the situations in which the “wide latitude counsel must have in making tactical decisions” will be limited to any one technique or approach. Ibid. It can be assumed that in some cases counsel would be deemed ineffective for failing to consult or rely on experts, but even that formulation is sufficiently general that state courts would have wide latitude in applying it. Here it would be well within the bounds of a reasonable judicial determination for the state court to conclude that defense counsel could follow a strategy that did not require *107the use of experts regarding the pool in the doorway to Johnson’s bedroom.
From the perspective of Richter’s defense counsel when he was preparing Richter’s defense, there were any number of hypothetical experts — specialists in psychiatry, psychology, ballistics, fingerprints, tire treads, physiology, or numerous other disciplines and subdisciplines — whose insight might possibly have been useful. An attorney can avoid activities that appear “distraetive from more important duties.” Bobby v. Van Hook, 558 U. S. 4, 11 (2009) (per curiam). Counsel was entitled to formulate a strategy that was reasonable at the time and to balance limited resources in accord with effective trial tactics and strategies. See Knowles, supra, at 125-126; Rompilla v. Beard, 545 U. S. 374, 383 (2005); Wiggins v. Smith, 539 U. S. 510, 525 (2003); Strickland, 466 U. S., at 699.
In concluding otherwise the Court of Appeals failed to “reconstruct the circumstances of counsel’s challenged conduct” and “evaluate the conduct from counsel’s perspective at the time.” Id., at 689. In its view Klein’s location was “the single most critical issue in the case” given the differing theories of the prosecution and the defense, and the source of the blood in the doorway was therefore of central concern. 578 F. 3d, at 953-954. But it was far from a necessary conclusion that this was evident at the time of the trial. There were many factual differences between prosecution and defense versions of the events on the night of the shootings. It is only because forensic evidence has emerged concerning the source of the blood pool that the issue could with any plausibility be said to stand apart. Reliance on “the harsh light of hindsight” to cast doubt on a trial that took place now more than 15 years ago is precisely what Strickland and AEDPA seek to prevent. Cone, 535 U. S., at 702; see also Lockhart, supra, at 372.
*108Even if it had been apparent that expert blood testimony could support Richter’s defense, it would be reasonable to conclude that a competent attorney might elect not to use it. The Court of Appeals opinion for the en banc majority rests in large part on a hypothesis that reasonably could have been rejected. The hypothesis is that without jeopardizing Richter’s defense, an expert could have testified that the blood in Johnson’s doorway could not have come from Johnson and could have come from Klein, thus suggesting that Richter’s version of the shooting was correct and Johnson’s a fabrication. This theory overlooks the fact that concentrating on the blood pool carried its own serious risks. If serological analysis or other forensic evidence demonstrated that the blood came from Johnson alone, Richter’s story would be exposed as an invention. An attorney need not pursue an investigation that would be fruitless, much less one that might be harmful to the defense. Strickland, supra, at 691. Here Richter’s attorney had reason to question the truth of his client’s account, given, for instance, Richter’s initial denial of involvement and the subsequent production of Johnson’s missing pistol.
It would have been altogether reasonable to conclude that this concern justified the course Richter’s counsel pursued. Indeed, the Court of Appeals recognized this risk insofar as it pertained to the suggestion that counsel should have had the blood evidence tested. 578 F. 3d, at 956, n. 9. But the court failed to recognize that making a central issue out of blood evidence would have increased the likelihood of the prosecution’s producing its own evidence on the blood pool’s origins and composition; and once matters proceeded on this course, there was a serious risk that expert evidence could destroy Richter’s case. Even apart from this danger, there was the possibility that expert testimony could shift attention to esoteric matters of forensic science, distract the jury from whether Johnson was telling the truth, or transform *109the case into a battle of the experts. Accord, Bonin v. Calderon, 59 P. 3d 815, 836 (CA9 1995).
True, it appears that defense counsel’s opening statement itself inspired the prosecution to introduce expert forensic evidence. But the prosecution’s evidence may well have been weakened by the fact that it was assembled late in the process; and in any event the prosecution’s response shows merely that the defense strategy did not work out as well as counsel had hoped, not that counsel was incompetent.
To support a defense argument that the prosecution has not proved its case it sometimes is better to try to east pervasive suspicion of doubt than to strive to prove a certainty that exonerates. All that happened here is that counsel pursued a course that conformed to the first option. If this case presented a de novo review of Strickland, the foregoing might well suffice to reject the claim of inadequate counsel, but that is an unnecessary step. The Court of Appeals must be reversed if there was a reasonable justification for the state court’s decision. In light of the record here there was no basis to rule that the state court’s determination was unreasonable.
The Court of Appeals erred in dismissing strategic considerations like these as an inaccurate account of counsel’s actual thinking. Although courts may not indulge “post hoc rationalization” for counsel’s decisionmaking that contradicts the available evidence of counsel’s actions, Wiggins, supra, at 526-527, neither may they insist counsel confirm every aspect of the strategic basis for his or her actions. There is a “strong presumption” that counsel’s attention to certain issues to the exclusion of others reflects trial tactics rather than “sheer neglect.” Yarborough v. Gentry, 540 U. S. 1, 8 (2003) (per curiam). After an adverse verdict at trial even the most experienced counsel may find it difficult to resist asking whether a different strategy might have been better, and, in the course of that reflection, to magnify their own responsibility for an unfavorable outcome. *110Strickland, however, calls for an inquiry into the objective reasonableness of counsel’s performance, not counsel’s subjective state of mind. 466 U. S., at 688.
2
The Court of Appeals also found that Richter’s attorney was constitutionally deficient because he had not expected the prosecution to offer expert testimony and therefore was unable to offer expert testimony of his own in response.
The Court of Appeals erred in suggesting counsel had to be prepared for “any contingency,” 578 F. 3d, at 946 (internal quotation marks omitted). Strickland does not guarantee perfect representation, only a “ ‘reasonably competent attorney.’ ” 466 U. S., at 687 (quoting McMann v. Richardson, 397 U. S. 759, 770 (1970)); see also Gentry, supra, at 7. Representation is constitutionally ineffective only if it “so undermined the proper functioning of the adversarial process” that the defendant was denied a fair trial. Strickland, supra, at 686. Just as there is no expectation that competent counsel will be a flawless strategist or tactician, an attorney may not be faulted for a reasonable miscalculation or lack of foresight or for failing to prepare for what appear to be remote possibilities.
Here, Richter’s attorney was mistaken in thinking the prosecution would not present forensic testimony. But the prosecution itself did not expect to make that presentation and had made no preparations for doing so on the eve of trial. For this reason alone, it is at least debatable whether counsel's error was so fundamental as to call the fairness of the trial into doubt.
Even if counsel should have foreseen that the prosecution would offer expert evidence, Richter would still need to show it was indisputable that Strickland required his attorney to act upon that knowledge. Attempting to establish this, the Court of Appeals held that defense counsel should have offered expert testimony to rebut the evidence from the prose*111cution. But Strickland does not enact Newton’s third law for the presentation of evidence, requiring for every prosecution expert an equal and opposite expert from the defense.
In many instances cross-examination will be sufficient to expose defects in an expert’s presentation. When defense counsel does not have a solid case, the best strategy can be to say that there is too much doubt about the State’s theory for a jury to convict. And while in some instances “even an isolated error” can support an ineffective-assistance claim if it is “sufficiently egregious and prejudicial,” Murray v. Carrier, 477 U. S. 478, 496 (1986), it is difficult to establish ineffective assistance when counsel’s overall performance indicates active and capable advocacy. Here Richter’s attorney represented him with vigor and conducted a skillful cross-examination. As noted, defense counsel elicited concessions from the State’s experts and was able to draw attention to weaknesses in their conclusions stemming from the fact that their analyses were conducted long after investigators had left the crime scene. For all of these reasons, it would have been reasonable to find that Richter had not shown his attorney was deficient under Strickland.
B
The Court of Appeals further concluded that Richter had established prejudice under Strickland given the expert evidence his attorney could have introduced. It held that the California Supreme Court would have been unreasonable in concluding otherwise. This too was error.
In assessing prejudice under Strickland, the question is not whether a court can be certain counsel’s performance had no effect on the outcome or whether it is possible a reasonable doubt might have been established if counsel acted differently. See Wong v. Belmontes, 558 U. S. 15, 27 (2009) (per curiam); Strickland, 466 U. S., at 693. Instead, Strickland asks whether it is “reasonably likely” the result would have been different. Id., at 696. This does not require a *112showing that counsel’s actions “more likely than not altered the outcome,” but the difference between Strickland’s prejudice standard and a more-probable-than-not standard is slight and matters “only in the rarest case.” Id., at 693, 697. The likelihood of a different result must be substantial, not just conceivable. Id., at 693.
It would not have been unreasonable for the California Supreme Court to conclude Richter’s evidence of prejudice fell short of this standard. His expert serology evidence established nothing more than a theoretical possibility that, in addition to blood of Johnson’s type, Klein’s blood may also have been present in a blood sample taken near the bedroom doorway pool. At trial, defense counsel extracted a concession along these lines from the prosecution’s expert. The pathology expert’s claim about the size of the blood pool could be taken to suggest only that the wounded and hysterical Johnson erred in his assessment of time or that he bled more profusely than estimated. And the analysis of the purported blood pattern expert indicated no more than that Johnson was not standing up when the blood pool formed.
It was also reasonable to find Richter had not established prejudice given that he offered no evidence directly challenging other conclusions reached by the prosecution’s experts. For example, there was no dispute that the blood sample taken near the doorway pool matched Johnson’s blood type. The California Supreme Court reasonably could have concluded that testimony about patterns that form when blood drips to the floor or about the rate at which Johnson was bleeding did not undermine the results of chemical tests indicating blood type. Nor did Richter provide any direct refutation of the State’s expert testimony describing how blood spatter near the couch suggested a shooting in the living room and how the blood patterns on Klein’s face were inconsistent with Richter’s theory that Klein had been killed in the bedroom doorway and moved to the couch.
*113There was, furthermore, sufficient conventional circumstantial evidence pointing to Richter’s guilt. It included the gun safe and ammunition found at his home; his flight from the crime scene; his disposal of the .32-caliber gun and of Johnson’s pistol; his shifting story concerning his involvement; the disappearance prior to the arrival of the law enforcement officers of the .22-caliber weapon that killed Klein; the improbability of Branscombe’s not being wounded in the shootout that resulted in a combined four bullet wounds to Johnson and Klein; and the difficulties the intoxicated and twice-shot Johnson would have had in carrying the body of a dying man from bedroom doorway to living room couch, not to mention the lack of any obvious reason for him to do so. There was ample basis for the California Supreme Court to think any real possibility of Richter’s being acquitted was eclipsed by the remaining evidence pointing to guilt.
* * *
The California Supreme Court’s decision on the merits of Richter’s Strickland claim required more deference than it received. Richter was not entitled to the relief ordered by the Court of Appeals. The judgment is reversed, and the case is remanded for further proceedings consistent with this opinion.
It is so ordered.
Justice Kagan took no part in the consideration or decision of this case.