concurring in part and dissenting in part:
Most habeas petitioners recount, as background to their claims of ineffective assistance of counsel, stories of childhood trauma and horrific mistreatment. Not so here. Petitioner William Robert Gray, a prosperous professional with retained counsel, was the abuser, not the abused.
After stalking his estranged wife for months, Gray shot her to death in the driveway of his home as she begged a passerby, “Mister, please don’t leave. If you leave, he’ll kill me.” State v. Gray, 347 N.C. 143, 491 S.E.2d 538, 544 (1997). Gray then entered his home and conscripted his children, ages eight and eleven, in his scheme to lie to authorities about what happened.
Gray promptly retained counsel, to whom he strenuously insisted that he suffered from no psychological infirmities. Counsel nevertheless had Gray committed to Dorothea Dix Hospital for a psychological evaluation. After observing and evaluating Gray for several weeks, hospital staff found no signs that Gray was suffering from a mental illness at the time of the murder. To the contrary, a hospital physician opined that Gray’s condition at the time of the murder “did not impair his basic understanding of the nature and quality of his actions nor his ability to recognize right from wrong.” J.A. 252. Ignoring the advice of his retained counsel, Gray refused to authorize funds for additional psychiatric examinations or testing. In spite of the attempts made by counsel and Gray’s unilateral refusal, Gray later sought collateral relief from his death sentence on the basis that his counsel provided ineffective assistance by failing to adequately investigate his mental health condition in preparation for sentencing. The North Carolina MAR court, *243closely reviewing Gray’s claim, found neither deficient performance nor prejudice.
Now, based primarily on a single assessment six years after the fact, and relying in part on the legally irrelevant state of Gray’s mind after the murder, the majority indefensibly concludes that the strain of marital difficulties (the effects of which were borne in large measure by Gray’s wife) might have driven Gray inexorably towards uxoricide. In doing so, the majority accords no deference whatsoever to the reasoned decision of the MAR court, instead dismissing that court’s findings based on no more than its contrary belief, unsupported by the record, that expert mental health evidence regarding the psychological effects of marital strife upon Gray might have made the difference between sentences of life and death. AED-PA, the statute that expressly controls our review of such claims, specifically prohibits the majority’s wholesale rejection of the state court findings of fact and conclusions of law.
Because the majority opinion contravenes AEDPA’s standard of review, mis-characterizes record evidence, and ignores Supreme Court and Fourth Circuit precedent, I am compelled to dissent.
I.
Before discussing the majority’s legal errors, it is first necessary to highlight some of the liberties it takes with the facts. The majority inflates the importance, of affidavits submitted by two doctors, neither of whom evaluated Gray; relies heavily on a single psychiatric evaluation that occurred a full six years after the murder; and glides over the eminently reasonable course of action undertaken by Gray’s counsel.
It is probably fair to say, and I doubt the majority would disagree, that Gray was a boorish person. The record reflects that he abused his wife, manipulated his children, was “challenging, confrontational, arrogant and haughty” in counseling, J.A. 227, and, when the marriage disintegrated, attempted to extract negative information about his wife from her doctor for use in the couple’s ongoing divorce proceeding. What is missing on this record is any evidence that these actions, or the murder itself, were products of impaired capacity or a diagnosable mental illness.
Indeed, evidence abounds to the contrary. Following an evidentiary hearing, the state MAR court found persuasive testimony that Gray “was calm, in control, and conducted himself normally as a person in ‘their right mind’ throughout the period of the contentious divorce proceedings and up to the murder itself.” J.A. 1563. The MAR court cited the testimony of no fewer than seven witnesses who, the court found, “offered first-hand observation and experience with [Gray] in the days and months preceding the murder of [his] wife.” J.A. 1564 (emphasis added). The court also placed weight on the statement of Gray’s attorney that he would not have allowed Gray to be interviewed by the police on the night of the murder if he had harbored questions about Gray’s mental stability. J.A. 1564.
In the face of the MAR court’s clear and cogent explication of the record, the majority ignores AEDPA’s deferential standard of review and resorts to mining the record for some indication that Gray’s evils could have been explained away by psychological factors.
The contrived nature of the majority’s support for its conclusions regarding Gray’s mental condition is nowhere more apparent than in its reliance on a single sentence, taken out of context, from an affidavit sworn out by Dr. Marshall Barker, Gray’s wife’s gynecologist, in the cou-*244pie’s divorce proceeding. In his affidavit, Dr. Barker stated,
Dr. Gray persistently called me and made appearances in my office on numerous occasions to discuss my treatment of his wife. Despite my informing him that such discussion would violate physician/patient confidentiality, he continued to make calls and appearances in my office demanding information about his wife. I realized that by his persistence in making appearances and telephone calls, he was trying to gather negative information on his wife. I can see from Dr. Gmy’s behavioral pattern that the behavioral aberration and psychological problem is with him and not his wife.
J.A. 224 (emphasis added). On cross-examination, Dr. Barker admitted that his personal interactions with Gray were limited to one telephone conversation lasting approximately twenty minutes. J.A. 789-90. Despite the fact that the reason for the conversation was for Gray to gather negative information about his wife from her gynecologist, the majority would rather use Dr. Barker’s “training in psychology” to inflate the significance of his statements about Gray’s “psychological problem.” See Majority Op. at 224; J.A. 224.1
The majority also relies on an affidavit filed by Ms. Marilyn Huber, the Grays’ marital and family therapist. This affidavit, like Dr. Barker’s affidavit, is irrelevant to the issue of Gray’s mental health. The affidavit concerned Dr. and Mrs. Gray’s marital counseling sessions. In the majority’s view, it was significant that during these sessions Ms. Huber found Mrs. Gray’s behavior “entirely appropriate,” and Dr. Gray’s, by contrast, “challenging, confrontational, arrogant and haughty.” J.A. 227; see Majority Op. at 224, 233. Dr. Gray’s behavior led Ms. Huber to discontinue their counseling sessions and led her to suggest that Mrs. Gray get psychological counseling for the children “to help them deal with the dissolution of the marriage and the possible pressure ... being *245asserted upon them by their father.” J.A. 227-28.
In an effort to give these affidavits significance they do not have, the majority skews some facts and takes others out of context. For instance, it is correct that Mrs. Gray’s divorce lawyer relied on the affidavits in moving for a court order “for a mental assessment of Gray.” Majority Op. at 224. What the majority fails to disclose, however, is that the motion requested that both Dr. and Mrs. Gray, as well as their children, be required to submit to mental and physical assessments. Far from serving as evidence of Dr. Gray’s psychological condition, the motion is, on its face, no more than an attempt by the attorney for a non-custodial parent in a divorce proceeding to protect his client from attacks on her psychological well-being.
Of course, I agree with the majority that the affidavits could be probative of Gray’s counsel’s performance to the extent that they raised “red flags” that counsel should have seen, but did not. See id. at 229. But a fair reading of the affidavits, as detailed herein, leads to the conclusion that they fail to carry the weight the majority lades them with; no reasonable reading of the affidavits could support the majoritys self-serving conclusions that they “were ... indications of Gray’s mental or emotional disturbance” and that they “revealed that Gray had a psychological problem.” Id. at 233.
Due to the substantive weakness of this purported evidence of Gray’s “psychiatric impairment” prior to the murder, the majority is forced to rely on events after the murder and Gray’s subsequent arrest. Gray was placed on suicide watch; fainted during a friend’s visit; was described by-friends as “damaged emotionally,” J.A. 884, and by the jail matron as “VERY depressed,” J.A. 889-90; and was described by his own attorney as “dazed” and “appear[ing to be] in a state of shock,” J.A. 240, 1702.2 None of these reactions can legitimately be called unnatural or unexpected for a man just arrested for the brutal murder of his wife. Instead of recognizing this, however, the majority warps time in lumping these post-arrest behaviors into its standardless framework for analyzing Gray’s mental health condition prior to the murder.
In any event, far from behaving ineffectively, Gray’s counsel responded immediately to the concerns raised by these post-arrest behaviors, requesting that his client be committed to Dorothea Dix Hospital to evaluate his capacity to proceed to trial.3 Gray was subsequently evaluated over the *246course of several weeks by hospital staff. He presented originally as “tearful, upset, and anxious,” oppositional and challenging at times, and angry and frustrated when describing his wife’s past behavior. J.A. 245-47. He was resistant to testing, refused to cooperate in many procedures, and displayed arrogance and verbal aggression towards hospital evaluators. The hospital’s initial diagnostic impression of Gray was that he “presented an adjustment reaction with mixed disturbance of emotions and conduct in relation to current ongoing stress.” J.A. 246. However, after dealing with Gray for several weeks, the hospital’s clinical findings indicated “no evident psychosis nor any impairment in this patient’s ability to understand his situation and surroundings,” and “no findings to suggest evidence of ongoing medical illness.” J.A. 251. Dr. Patricio Lara, the forensic psychiatrist who completed Gray’s discharge summary, concluded his analysis and opinion as follows:
Available information describing this patient’s condition around the time of the incident in question indicate that he was under ongoing stress related to his family and marital situation. Observations reported by friends describe evidence of preoccupation and distress and no features to indicate confusion nor any serious impairment in this patient’s ability to function. Based on available information, it is my opinion that this patient’s condition at the time in question did not impair his basic understanding of the nature and quality of his actions nor his ability to recognize right from wrong. The magnitude of ongoing stressors may have contributed to regression in behavior and reductions in impulse control and control of emotions.
J.A. 252 (emphasis added). Dr. Lara then noted that Gray was, at the time the evaluation was completed, “undergoing [a] significant amount of distress and could benefit significantly from psychiatric counseling.” J.A. 252-53.4
Dr. Lara concluded the discharge summary with the following statement, which the majority attempts to imbue with significance: “If other sources of information become available to describe this patient’s functioning and condition prior to [the murder], I would be glad to review the information to expand my opinions if possible in regards to this patient’s condition at the time of the alleged offense.” J.A. 253; see Majority Op. at 225, 229-30, 230-31, 232, 232-33, 236, 239-40 (referring to this statement as, among other things, an “offer” to provide “additional assistance” and an “offer of mental health investigation assistance”). The limits of this comment should be readily apparent. It suggests only a willingness to revisit the diagnosis of “no impairment” should additional evidence be forthcoming.5
The one potential “red flag” to be found in the hospital discharge summary was *247nonspecific and equivocal: “The magnitude of ongoing stressors may have contributed to regression in behavior and reductions in impulse control and control of emotions.” J.A. 252 (emphasis added). Contrary to the majority’s repeated reference to the “many signs” of Gray’s mental or emotional instability, there were in fact few, if any, facts in this record that might have led a reasonable attorney to believe that Gray was laboring under a diagnosable mental or emotional impairment at the time that he murdered his wife.
After reviewing the Dorothea Dix discharge summary, Gray’s attorney recommended that they retain an independent psychiatrist to evaluate Gray and to be on standby for the trial and, if necessary, sentencing. J.A. 1660, 1663-64. Gray refused, telling his paid attorney “not to spend another fing penny on this trial, [and] that he didn’t need a psychiatrist.” J.A. 1666. Gray forbade his attorney to retain an independent psychiatrist and declined to authorize the expenditure of funds for a psychiatrist, even though he had authorized funds for other experts, including a surveyor and an expert on jury selection. J.A. 1665-66.
Faced with an absence of evidence from the time period leading up to the murder that Gray suffered from a diagnosable mental or emotional disorder, and counsel’s appropriate response to the circumstances at the time, the majority is forced to rely on the opinion of a doctor hired six years after the murder to support the con-elusion it seeks — that Gray suffered from a mental illness or acted with impaired capacity when he killed his wife.
Gray was sentenced to death in December 1993. The testimony and affidavit of Dr. James Bellard, a psychiatrist, were submitted at the MAR hearing in 1999. Dr. Bellard interviewed Gray for a total of four hours in January 1999. He reviewed only the records of the Grays’ family and marital therapist, the discharge summary from Dorothea Dix, and the North Carolina Supreme Court holdings in Gray’s case. (Bellard Aff., J.A. 275.)6 Based on this limited information, Dr. Bellard opined that Gray “did at the time of the offense ... and does at the current time suffer from a psychiatric illness, specifically, Paranoid Personality Disorder (DSM Code 301.0).” Id. Dr. Bellard further offered his opinion that Gray might suffer from other psychiatric disorders and that expert psychiatric testimony would have been useful to the jury in Gray’s case because it would have supported the submission of at least two statutory mitigating factors. J.A. 275-76.7
The state MAR court reasonably discounted Dr. Bellard’s testimony both because his contact with Gray occurred six years after the murder and because Dr. Bellard had not reviewed the trial testimony of Gray’s friends, relatives, and associates describing Gray’s conduct in the months leading up to the murder. J.A. 1562-63. Relying on extensive evidence and testimony of no fewer than seven wit*248nesses with knowledge of Gray’s behavior prior to and at the time of the murder, the MAR court concluded that available evidence, including both the evidence actually presented at sentencing and that which Gray contended should have been presented but was not, demonstrated that Gray was in full control of his mental faculties and his ability to reason at the time of the murder. J.A. 1566. To conclude that the result of Gray’s trial would have been different if the “missing” evidence had been presented, the MAR court held, would require “leap[s] of speculation.” Id. The court therefore rejected Gray’s ineffective assistance claim, finding that he had failed to show either that his attorney’s performance was deficient or that any deficiency resulted in prejudice. See Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).
II.
The serious errors in the majority’s cavalier recitation of the facts are compounded by its perplexing legal analysis. Though the majority sets forth the proper framework for AEDPA review, it wholly disregards its application.
Federal habeas relief is available to state prisoners only when state court proceedings (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.” 28 U.S.C. § 2254(d). A state court decision is “contrary to” established federal law if the state court “applies a rule that contradicts the governing” Supreme Court precedent or confronts facts that are “materially indistinguishable” from those presented in a relevant Supreme Court decision and arrives at the opposite result. Williams v. Taylor, 529 U.S. 362, 405-06, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000).
An “unreasonable application” of clearly established federal law “occurs when a state court identifies the correct governing legal principle from [the Supreme] Court’s decisions but unreasonably applies that principle to the facts of [a] petitioner’s case.” Rompilla v. Beard, 545 U.S. 374, 380, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005) (internal quotations omitted). To warrant the extreme remedy of federal intrusion into state criminal adjudications, it is not enough “[that] a federal court believes the state court’s determination was incorrect”; the relevant question is “whether that determination was unreasonable — a substantially higher threshold.” Schriro v. Landrigan, - U.S. -, 127 S.Ct. 1933, 1939, 167 L.Ed.2d 836 (2007) (citing Williams, 529 U.S. at 410, 120 S.Ct. 1495). In habeas cases,
[f]actual determinations by state courts are presumed correct absent clear and convincing evidence to the- contrary, ... and a decision adjudicated on the merits in state court and based on a factual determination will not be overturned on factual grounds unless objectively unreasonable in light of the evidence presented in the state-court proceeding.
Miller-El v. Cockrell, 537 U.S. 322, 340, 123 S.Ct. 1029, 154 L.Ed.2d 931 (2003). Accordingly, this court has recently reaffirmed that “federal courts are to accord considerable deference to state courts in their review of state habeas proceedings.” Meyer v. Branker, 506 F.3d 358, 365 (4th Cir.2007) (citing Williams, 529 U.S. at 412-13, 120 S.Ct. 1495).
III.
The crux of Gray’s petition is his ineffective assistance of counsel claim. To pre*249vail on such a claim under Strickland, a petitioner must establish both that counsel’s performance was deficient and that the deficiency prejudiced the defense. 466 U.S. at 687, 104 S.Ct. 2052. The majority concludes, as it must in order to grant relief, that the North Carolina MAR court engaged in an unreasonable application of Supreme Court precedent concerning both components of Strickland. The majority first holds that Gray’s counsel rendered deficient performance “by failing to investigate and develop, for sentencing purposes, evidence that Gray suffered from a severe mental illness.” Majority Op. at 223. This portion of the majority opinion is grounded in its determination that Gray’s attorneys’ investigation was “unreasonable under prevailing professional norms.” Id. at 232-33.
After arriving at its conclusion regarding deficient performance, the majority further concludes that there is a reasonable probability that this failure prejudiced the outcome at sentencing. Id. at 238. In the majority’s view, the MAR court applied the wrong standard for evaluating prejudice, and the MAR court’s decision was an “unreasonable application of the proper standard.” Id. at 235. I find each of these holdings to strain reason and confound well-established law.
A.
The precedents governing our analysis of Strickland’s deficiency prong are clear. An attorney’s performance is measured using the standard of “reasonableness under prevailing professional norms.” Strickland, 466 U.S. at 688, 104 S.Ct. 2052. “American Bar Association standards and the like ... are guides to determining what is reasonable.” Id. The Supreme Court has made clear that ABA standards such as the American Bar Association Guidelines for the Appointment and Performance of Counsel in Death Penalty Cases (“ABA Guidelines”) are “only guides,” not requirements. Id. (emphasis added); see also Rompilla, 545 U.S. at 400, 125 S.Ct. 2456; Wiggins v. Smith, 539 U.S. 510, 543, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003). Other guides might include, but are not limited to, state-specific standards and the actual courtroom practice of capital defenders in the relevant jurisdiction or nationwide. See Marshall v. Cathel, 428 F.3d 452, 467-68 (3d Cir.2005). In utilizing any guide to determine reasonableness, Strickland requires courts to be “highly deferential” to attorneys’ strategic choices to avoid the “distorting effects of hindsight.” 466 U.S. at 689, 104 S.Ct. 2052. This court recently explained,
In the sentencing context, this “highly deferential” standard means that defense counsel have the flexibility to vary their approach given their client’s unique circumstances. See Lovitt v. True, 403 F.3d 171, 179 (4th Cir.2005) (“In many cases, counsel’s decision not to pursue a particular approach at sentencing reflects not incompetence, but rather a sound strategic choice.”). This is exactly as it should be: the touchstone of effective representation must be sound, evidence-based judgment, rather than a set of mandates counsel must programmatically follow without deviation. “No particular set of detailed rules for counsel’s conduct can satisfactorily take account of the variety of circumstances faced by defense counsel or the range of legitimate decisions regarding how best to represent a criminal defendant.” Strickland, 466 U.S. at 688-89, 104 S.Ct. 2052.
Meyer, 506 F.3d at 371 (emphasis added).
The majority wisely concedes that the ABA Guidelines provide only advice and suggestions of “avenue[s] of investigation that counsel should consider,” and that in *250this context they “can be relevant” to determinations of what constitutes reasonable performance but “certainly cannot be dispositive in and of themselves.” Majority Op. at 229 (citing Meyer, 506 F.3d at 372.) (emphasis added). In doing so, the majority recognizes the Supreme Court’s admonition that programmatic rules “would interfere with the constitutionally protected independence of counsel and restrict the wide latitude counsel must have in making tactical decisions. Indeed, the existence of detailed guidelines for representation could distract counsel from the overriding mission of vigorous advocacy of the defendant’s cause.” Strickland, 466 U.S. at 689, 104 S.Ct. 2052 (internal citation omitted).
Nevertheless, by finding that an in-depth mental health investigation was required on these facts, the majority effectively calls for a per se rule that such an investigation is required in every death penalty case, a result both this court and the Supreme Court have expressly rejected. See Meyer, 506 F.3d at 371-72; Strickland, 466 U.S. at 689, 104 S.Ct. 2052.
The majority’s opinion “ignore[s] the very basic fact that different circumstances often require different strategies.” Meyer, 506 F.3d at 372; see Strickland, 466 U.S. at 689, 104 S.Ct. 2052; Frye v. Lee, 235 F.3d 897, 904 n. 7 (4th Cir.2000). Here, defense counsel, facing the circumstances presented in Gray’s case, attempted to demonstrate at sentencing that Gray, after a lifetime of good behavior, made a tragic mistake in a moment of weakness due to the dissolution of his marriage and the hospitalization of his mother. See J.A. 1439 (defense counsel’s statement during closing argument that “overall, [the jury will] be looking at a few days in 1992 versus a whole life time of 42 plus years”); J.A. 1442 (closing argument statements contrasting a lifetime of conforming to society’s rules with a few “non-conform[ing]” events).8 The mental or emotional disturbance mitigating factor played a role in this overarching theory, as elucidated by the jury instruction suggested by counsel, J.A. 1455, and the similar instruction ultimately given by the court:
For this mitigating circumstance to exist it is enough that the defendant’s mind or emotions were disturbed from any cause and that he was under the influence of the disturbance when he killed the victim.
You would find this mitigating circumstance if you find that the defendant was under mental or emotional stress as a result of his break up — the break up of the family marital unit and/or the mental or emotional stress of the hospitalization of his 82-year-old-blind mother and the uncertainty of her prognosis; and that as a result of this, the defendant was under the influence of mental or emotional disturbance when he killed the victim.
J.A. 1475-76.
Defense counsel’s approach met with some success, leading the jury to find thirteen of the twenty-one mitigating factors proffered, including that Gray was a loving father, a loving son who had cared for his mother at her time of need, a contributor to the community, and that “[f]or 42 years prior to his separation from Roslyn Gray, Defendant led an uneventful, law-abiding life posing no threat to his family or community.” J.A. 740. The jury did not find the mental health mitigating factors, and it ultimately recommended death. However, *251that defense counsel did not prevail in avoiding a death sentence does not transform its mitigation strategy into ineffective assistance. See Strickland, 466 U.S. at 689, 104 S.Ct. 2052 (“It is all too tempting for a defendant to second-guess counsel’s assistance after conviction or adverse sentence, and it is all too easy for a court, examining counsel’s defense after it has proved unsuccessful, to conclude that a particular act or omission of counsel was unreasonable.”); Lovitt, 403 F.3d at 179 (noting that, to pursue a particular theory at sentencing, though it may ultimately fail, often “reflects not incompetence, but rather a sound strategic choice”).
The diligence and cautiousness of Gray’s counsel bear repeating. They began their investigation immediately after the murder by having Gray committed for a weeks-long evaluation at Dorothea Dix Hospital. They later made the reasonable choice to forgo additional in-depth mental health investigation based on Dr. Lara’s evaluation and their overall assessment of the case, including all available evidence, their own observations of their client, and their client’s refusal to authorize funds for experts and his insistence that further psychiatric evaluation was unnecessary.9
The circumstances here did not demand that Gray’s sentencing strategy include procuring expert mental health testimony. Gray’s initial psychiatric evaluation found no evidence of impairment at the time of the murder. Gray’s attorneys nevertheless approached Gray about hiring a mental health expert to be “on standby,” though they did not necessarily intend to use psychiatric evidence. When Gray refused to authorize funds for such an expense, they did not press the matter. “[Strategic choices made after less than complete investigations are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.” Strickland, 466 U.S. at 690-91, 104 S.Ct. 2052. “Strickland does not require counsel to investigate every conceivable line of mitigating evidence no matter how unlikely the effort would be to assist the defendant at sentencing.” Wiggins, 539 U.S. at 533, 123 S.Ct. 2527. Contrary to the majority’s analysis, a fair reading of this record leads to the inescapable conclusion that the investigation undertaken by Gray’s counsel was reasonable.10
Counsel’s reasonable performance continued at sentencing. The majority is simply incorrect in asserting that “Gray’s counsel did not present (or attempt to present) any evidence at sentencing” that Gray was under mental or emotional strain when he murdered his wife.11 Majority *252Op. at 238-39. In fact, the trial court found that sufficient evidence had been presented to justify submission of the mental health mitigating factors to the jury. A number of lay witnesses, at trial and at sentencing, testified that Gray was under significant stress and acting out of character at the time of the murder.12 This testimony supported defense counsel’s sentencing theory, in particular its argument that the jury should find the mental or emotional disturbance mitigating factor because Gray “suffered from stress or a stressful situation as the result of the break up of the family marital unit and as a result of stress caused by the hospitalization of the defendant’s mother.” J.A. 1455.
The majority leaves no doubt that it would have taken a different approach at sentencing. The majority’s preferred strategy, concocted with the benefit of hindsight, grows out of its conviction that “mental health mitigating factors ... were Gray’s best hope of convincing the jury that he did not deserve the death penalty.” Majority Op. at 236. The majority opines that “[t]he missing expert evidence on Gray’s impaired mental condition would have provided the essential support for these factors.” Id. (emphasis added). In so deciding, the majority faults Gray’s counsel because, the majority suggests, “There is no indication that [defense counsel] understood that expert mental health evidence could be critical to the jury’s decision on sentencing.” Id. at 232 (emphasis added).
Regrettably, this speculation calls defense counsel to account for a failure to pursue the majority’s preferred strategy, based on the majority’s forecast of what evidence “would have” been provided and what that evidence “could” have done in the jury’s minds. Contrary to the majority’s assertion, however, there is no objective evidence that its favored strategy was the “best,” or even that its strategy offered greater promise than defense counsel’s chosen course. Therefore, the only way the majority here could rightly condemn defense counsel’s actions would be if a per se rule required the presentment of expert mental health mitigation evidence. But that is not the law. Meyer, 506 F.3d at 372 (“No per se rule requires the presentment of [mental health mitigation] evidence at trial.”).
The majority’s opinion, then, is precisely the type of overreaching condemned by Strickland and its progeny. Today, two jurists, with the benefit of hindsight, discard the high degree of deference owed to defense counsel’s strategic choices, and replace it with an attempt to mandate that sentencing strategies in death penalty cases always include presentation of expert mental health testimony. The opinion is grossly out of step with the flexible approach required under Supreme Court and Fourth Circuit precedent. See Meyer, 506 F.3d at 371-72; Strickland, 466 U.S. at 688-89, 104 S.Ct. 2052. Under AEDPA’s highly deferential standard of review, there is no basis for any conclusion other than that counsel’s approach to the investi*253gation of Gray’s mental health was marked by reasonable professional judgments, that counsel’s performance was not deficient, and that the MAR court did not unreasonably apply Supreme Court precedent in so finding.
B.
The majority’s opinion regarding prejudice is even more problematic. William Robert Gray murdered his wife in the driveway of his home after she begged a passerby for assistance. In the face of overwhelming evidence of guilt, which it wisely does not address, the majority is faced with the difficult task of showing that psychological testimony might have made a difference. By any objectively reasonable standard, it would not.
Under North Carolina law, a habeas petitioner “must demonstrate a reasonable probability that at least one juror would have found that his new mitigating evidence, combined with the existing mitigating evidence, outweighed the aggravating circumstances.” Buckner v. Polk, 453 F.3d 195, 203 (4th Cir.2006). This court applies a heavy measure of deference to the MAR court’s conclusion regarding prejudice:
Even if on de novo review we might strike a different balance concerning the relative weight of the aggravating and mitigating evidence, we may not disturb the MAR court’s conclusion that [the petitioner] did not demonstrate prejudice unless we find that conclusion to be unreasonable in light of clearly established Supreme Court precedent or in light of the evidence before the MAR court. See § 2254(d) (1) - (2). When determining whether [the petitioner] has satisfied one of these two standards, we must presume the MAR court’s factual findings to be correct unless [the petitioner] provides clear and convincing evidence to the contrary. See § 2254(e)(1).
Id.
After an exhaustive hearing, the MAR court rejected Gray’s ineffective assistance claims in a 45-page order. The court “f[ou]nd that the record contains persuasive testimony from State and defense witnesses that defendant was calm, in control, and conducted himself normally as a person in ‘their right mind’ throughout the period of the contentious divorce proceedings and up to the murder itself.” J.A. 1563. The witness testimony cited by the court “offered first-hand observation and experience with defendant in the days and months preceding the murder.” J.A. 1564. In light of this evidence, the court rejected Gray’s contention that there was a reasonable probability that the presentation of other mental health evidence would have changed the outcome of his trial. As a result, the court found that Gray failed to satisfy Strickland’s prejudice prong.
The majority first suggests that the state court applied the incorrect standard. In its order, the state court wrote, “Even with the proffered medical testimony this [c]ourt finds it to be a leap of speculation that the jury would necessarily have found the existence of the statutory and nonstat-utory mitigating circumstances.” J.A. 1566. The majority reads this statement out of context to show that the MAR court required certainty when all that was required under Strickland was a reasonable probability of a different outcome. However, the next sentence in the MAR court opinion suggests that the MAR court plainly understood the proper standard: “It is a further leap of speculation that there is a reasonable probability that the result of defendant’s trial would have been any different.” Id. Despite its inartful word choice, the MAR court applied the correct standard.
*254Next, the majority questions the MAR court’s failure to engage in an explicit reweighing of all available mitigation evidence, including expert mental health evidence, against the case in aggravation. In so doing, the majority attempts to obscure the obvious: that other than the proffered testimony of Dr. Bellard, which the MAR court fully considered but ultimately found unpersuasive, there was no additional expert mental health evidence to add to the mitigating side of the balance. Under these circumstances, an explicit discussion of the reweighing process would have been farcical.13
The majority’s substitution of its factual findings for those of the MAR court is legally impermissible, see Buckner, 453 F.3d at 203, and makes a mockery of the principles of federalism underlying habeas review. Moreover, the majority’s doe-eyed view of what Gray might have presented and its resultant conclusion notwithstanding, a dispassionate analysis of the entire case in mitigation suggests that it would have been strengthened little, if at all, by the proffered mental health evidence. Indeed, much of the omitted evidence would likely have harmed Gray’s case as much as, or more than, it would have helped.
Numerous lay witnesses, including Gray himself, testified at trial and at sentencing in support of defense counsel’s theory that Gray had been a loving father, son, and spouse, who became increasingly troubled under the strain of the break-up of his marriage and the fading health of his mother. Additional lay witness testimony to these issues would have been duplica-tive, and therefore must be discounted under our precedent. See Buckner, 453 F.3d at 207. See also Bowie v. Branker, 512 F.3d 112, 121 (4th Cir.2008). Furthermore, in evaluating the possible prejudice resulting from counsel’s failure to procure additional testimony on these topics, one must recognize that the existing testimony regarding Gray’s erratic and overly emotional behavior during the couple’s separation accompanied observations that he was in full control of his mental faculties at the time of the murder. See, e.g., J.A. 822-24, 831. Therefore, additional lay testimony in this vein likely would also have been double-edged. See Moody v. Polk, 408 F.3d 141, 152 (4th Cir.2005).
The majority fails to recognize the centrality of this abundant lay witness testimony to defense counsel’s mitigation theory. Counsel chose to focus on Gray’s mental and emotional stress and chose not to attempt to demonstrate the existence of a diagnosable mental or emotional impairment. To support its chosen theory, counsel was able to rely on numerous firsthand observations of Gray in the time period before the murder. The available evidence from medical professionals, suggesting that Gray did not suffer from a diagnosable mental illness, only bolstered counsel’s choice to focus on lay witness testimony. See J.A. 789 (Dr. Barker, commenting that “[he] felt [Gray] was in control of his mental faculties. He just *255sounded to me very obsessively compulsive about this situation.”); J.A. 252 (Dr. Lara’s evaluation, summarizing, “[I]t is my opinion that this patient’s condition at the time in question did not impair his basic understanding of the nature and quality of his actions nor his ability to recognize right from wrong,” but that “[t]he magnitude of ongoing stressors may have contributed to regression in behavior and reductions in impulse control and control of emotions.”).
Nevertheless, the majority holds that defense counsel’s failure to procure expert witness testimony prejudiced Gray because the supposedly “missing” expert testimony — in particular, the testimony of Dr. Bellard — would have demonstrated that Gray suffered from Paranoid Personality Disorder at the time of the murder.14
The majority fails to acknowledge several glaring substantive weaknesses of its assertion that expert evidence would have aided Gray. First, the expert evidence ostensibly missing from the case in mitigation would likely have been double-edged. For example, had defense counsel called Dr. Lara, the forensic psychiatrist from Dorothea Dix, to provide expert testimony in mitigation, he could have testified to the initial diagnostic impression of Gray as presenting “an adjustment reaction with mixed disturbance of emotions and conduct in relation to current ongoing stress.” J.A. 246. However, he would have also had to acknowledge that observing Gray shortly after the murder and his subsequent arrest presented “an extreme situation that may [have] exaggerate^] personality traits to a level that may [have] appeare[d] pathological.” J.A. 251. Of course, he would also have had to admit that the hospital found “no evident psychosis nor any impairment in this patient’s ability to understand his situation and surroundings,” and “no findings to suggest evidence of ongoing medical illness.” Id. In addition, he would have been constrained by his own opinion, included in the discharge summary, “that [Gray’s] condition at the time in question did not impair his basic understanding of the nature and quality of his actions nor his ability to recognize right from wrong.” J.A. 252. In the best-case scenario for Gray, testimony by Dr. Lara would have been internally inconsistent and would have also severely undercut the proffered testimony of Dr. Bellard. Our precedent demands that we discount such equivocal mental health evidence. See Bowie, 512 F.3d at 121 (holding that equivocal mental health evidence failed to demonstrate “a reasonable probability that the jury would not have recommended death”); Moody, 408 F.3d at 152.
Second, Gray failed to present any evidence to contradict the MAR court’s reasoned decision to discount the proffered expert testimony of Dr. Bellard.15 Even were the MAR court wrong to discount this testimony, the evidence the majority *256finds that the jury “would have heard” from Dr. Bellard, Majority Op. at 236, pales in comparison to the evidence presented in cases in which prejudice has been found. See Williams, 529 U.S. at 395 & n. 19, 120 S.Ct. 1495 (finding prejudice from counsel’s failure to present evidence of the defendant’s “nightmarish childhood,” which included criminal neglect and physical abuse of the defendant by his parents); Wiggins, 539 U.S. at 536, 123 S.Ct. 2527 (finding prejudice from counsel’s failure to alert the jury to the defendant’s borderline mental retardation and severe childhood physical and sexual abuse).
Neither Gray nor the majority has shown the state MAR court’s conclusions respecting potential prejudice on the available body of evidence to be unreasonable in light of clearly established Supreme Court precedent, see § 2254(d)(l)-(2), or rebuttable by clear and convincing contrary evidence, see § 2254(e)(1). Instead, finding no reasonable basis for an assault on the state MAR court’s opinion under AEDPA, the majority departs from the statutorily prescribed boundaries that should cabin its review. So situated, and free to maneuver without the nettlesome requirements of statute, precedent, and standards of review, the majority casts aside the MAR court’s opinion and conducts its own analysis in the first instance, concluding that Gray’s mitigation case, with Dr. Bellard’s opinion added, would have been “quite powerful.” Majority Op. at 237.16 In so concluding, it ignores the heavy measure of deference usually afforded defense counsel’s strategic decisions, *257instead deciding on its own what strategy presented “Gray’s best hope of convincing the jury that he did not deserve the death penalty.” Majority Op. at 236. With its preferred result in sight, the majority “weighs” a “powerful” mitigation case of its own making against the state’s evidence in support of the death penalty and concludes, unsurprisingly, that there is a reasonable probability that the jury would have returned with a different sentence.
Contrary to the majority’s analysis, given the deferential AEDPA standard of review, the substantive weakness of the new evidence, and the failure of Gray and the majority to rebut the MAR court’s adverse treatment of Dr. Bellard’s opinion, our precedent dictates that the MAR court’s findings on prejudice should have been allowed to stand. See Buckner, 453 F.3d at 207.
IV.
For the foregoing reasons, I dissent from the portion of the majority opinion reversing the judgment of the district court. I concur in the majority opinion to the extent it affirms the district court’s judgment.
. The actual facts of record put the majority's spin in proper context. Dr. Barker's "training in psychology” involved taking some number of unspecified "courses dealing with psychology” and attending a conference on domestic violence. See J.A. 780.
The majority further characterizes Dr. Barker's opinion as suggesting that Gray’s psychological issues "included obsessive compulsion.” Majority Op. at 224. Significantly, it offers no citation to any such diagnosis, perhaps because there was none. In fact, the only commentary remotely along those lines occurred in the following portion of Dr. Barker's testimony on direct examination, and it can hardly be characterized as a diagnosis:
Q. Dr. Barker, at the time you were speaking to the defendant, did you formulate any opinion to yourself whether or not he was in control of his mental and physical faculties at those times? Yes or no.
A. He was — Yes. He was in control of his faculties.
Q. Your opinion was what?
A. My opinion that he — if you're asking me is he — was this — If you’re asking me, over a telephone conversation, did this seem delusional or psychotic, I'm telling you no, it did not.
Q. Did he appear — did you formulate an opinion satisfactory to yourself, other than the agitation, he appeared and sounded to you like he was in control of his mental faculties?
A. I felt he was in control of his mental faculties. He just sounded to me very obsessively compulsive about this situation.
J.A. 789. The only objectively accurate characterization of this colloquy is that it reflects Dr. Barker's view of Gray as mentally sound, but perhaps “obsessively” focused on the marital deterioration for which he bore primary responsibility. However, as with the majority’s factual characterizations throughout, its observations here reflect the strain of its efforts to create evidence of a mental disorder that does not exist.
. Although the attorney wrote in his commitment motion that Gray appeared “unable to understand the gravity of the situation,” J.A. 240, he later clarified that he “had no doubts in my mind that [Gray] knew what was going on," but that Gray "looked haggard, worn, worn out, tired, frustrated, and kind of a realization that I'm in trouble. I mean it’s dazed like you know, I’m realizing hey, this is serious,” J.A. 1702-03.
. The majority takes Gray’s counsel to task for failing to provide affidavits from Gray's pending divorce case file to the staff at Dorothea Dix. See Majority Op. at 225. Again, a closer look at the surrounding facts demonstrates the reasonableness of counsel’s actions. The case file in question was delivered to one of Gray’s trial attorneys a mere eleven days prior to the murder, when he was retained to represent Gray in the divorce case. The attorney had no occasion to review the case file prior to the murder-he and Gray had arranged to go over the file the weekend after Thanksgiving, but Gray murdered his wife the night before Thanksgiving. Unlike the majority, I find nothing condemnable about an attorney's failure to immediately review a client's divorce case file after that client moots the case by brutally murdering his wife, or his failure to extract from such a file affidavits of dubious relevance to his client’s mental health.
. Once again, the majority conflates observations related to Gray's post-arrest behaviors with his condition prior to the murder. It should surprise no one that Dr. Lara suggested psychiatric counseling for Gray after his arrest for the brutal murder of his wife, and that Dr. Lara indicated that Gray, in his post-arrest condition, might benefit from medication.
. The majority also relies on the assertion that Dr. Lara’s assistance would have been given "free of charge.” Majority Op. at 230-31, 232. The only support for this assertion is Dr. Lara’s affidavit, executed on June 7, 2006, thirteen years after Gray’s trial and six 3'ears after the MAR proceedings. This affidavit was considered by neither the MAR court nor the district court and should not have been considered by this court on habeas review. See Bell v. Jarvis, 236 F.3d 149, 171 n. 13 (4th Cir.2000).
. The majority again engages in judicial puffing when it states that Dr. Bellard "reviewed excerpts of the trial record.” Majority Op. at 227. Dr. Bellard reviewed the testimony of only one witness — Dr. Gridley, Gray’s son’s counselor — and the testimony of Gray himself. J.A. 1752, 1756. Tellingly, Dr. Bellard's affidavit, which detailed the bases for his psychiatric opinion, did not even mention these trial record excerpts.
. The factors Dr. Bellard noted were that "the capital felony was committed while the defendant was under the influence of mental and emotional disturbance,” and that "the capacity of the defendant to appreciate the criminality of his conduct or to conform his conduct to the requirements of the law was impaired.” J.A. 276.
. Much of the testimony at the sentencing hearing supported this theory by highlighting the strength of the Grays' marriage in its early years, Gray’s excellent work in his career as a dentist, and Gray’s dedication to his children as a little league coach and a loving father.
. Of course, counsel cannot blindly accept a client's opinion regarding his own mental health as fact. But it would be equally suspect for an attorney to ignore a client whose self-assessment found abundant support and minimal contraindication in the record and comported with the attorney's own reasoned judgment. See Strickland, 466 U.S. at 691, 104 S.Ct. 2052 (“The reasonableness of counsel’s actions may be determined or substantially influenced by the defendant’s own statements or actions.”).
. Inexplicably, the majority mischaracterizes my opinion as failing to "pursu[e] the ... argument that Gray’s counsel undertook a reasonable investigation." Majority Op. at 234. But this is precisely the argument I am making-that counsel’s investigation was, under the circumstances known to him at the time, reasonable.
.The majority's statement is most directly contradicted by a comment from defense counsel Worthington's closing argument at sentencing: "You heard Charles Buchanan, a life-long friend talk about Mr. Gray's mother [whose hospitalization, Buchanan had explained, caused Gray considerable distress], and this is the mitigating factor, the felony *252was committed while the defendant was under the influence of mental or emotional distress.” J.A. 1430.
. See, e.g., J.A. 1289 (Charles Buchanan, a friend of the Grays, testifying at sentencing that the couple was very close throughout their marriage, until early 1992, and that at that point the dissolution of the marriage was “tearing [Gray] up”); J.A. 802 (Debbie Ryals, another friend of the Grays, indicating that in the period leading up to the murder Gray "didn't appear to be in his right mind” and spoke violently about his conflicts with his wife); J.A. 822 (Don Hollowell, another friend of the Grays, testifying that after the couple's separation Gray was "very disturbed” and "didn't have a grasp of what was going on”).
. The majority also ignores the significance of the Supreme Court’s recent decision in Schriro v. Landrigan, - U.S. -, 127 S.Ct. 1933, 167 L.Ed.2d 836 (2007). In Schriro, the Supreme Court emphasized that reviewing courts must assess reasonableness from the perspective of the state postconviction court at the time it made its decision. Schriro, 127 S.Ct. at 1942. When the MAR court issued its decision in this case, it was not objectively unreasonable for a state postcon-viction court to conclude that a defendant could not establish prejudice after refusing to allow his counsel to obtain mental health evidence in support of the case in mitigation. See id. The majority’s analogy to Rompilla, Majority Op. at 232, in support of the contrary conclusion is misguided; Rompilla was unavailable to the MAR court at the time it made its decision.
. The majority's star witness, Dr. Bellard, acknowledged that paranoid personality disorder requires evaluation of an individual’s long-term patterns and functioning, and manifests itself by late adolescence or early adulthood. LA. 1753. Such a diagnosis would thus cut directly against defense counsel’s mitigation theory that Gray was a normal, healthily functioning adult and a loving father, son, and husband, up until the dissolution of his marriage and his mother's hospitalization. See supra Part III.A.
. The majority suggests that the MAR court’s reasons for discounting Dr. Bellard’s testimony were "objectively unreasonable." Majority Op. at 236-37. The first factor in the MAR court's decision was the fact that Dr. Bel-lard’s opinion, based on a total of four hours spent with Gray, took place six years after the murder. I fail to see how the brevity of the evaluation and the significant time gap between the murder and the evaluation were "objectively unreasonable” factors for the MAR court to have considered. The MAR court also found it significant that "Dr. Bel-*256lard had not reviewed the trial testimony which set forth defendant's actions and conduct in the days and months leading up to the murder ... as well as the friends relatives, and associates of defendant.” J.A. 1562. Instead of recognizing the reasonableness of discounting Dr. Bellard's testimony on this basis, the majority instead highlights the sliver of the trial record that Dr. Bellard did review: the testimony of one witness, Gray’s son's counselor; the testimony of Gray himself; and two affidavits. That Dr. Bellard performed this limited review does nothing to lessen the import of the evidence Dr. Bellard failed to review and the reasonableness of the MAR court's recognition of the resulting limits of Dr. Bellard's testimony.
Finally, the majority grasps at straws in noting that "[t]he DSM-IV diagnostic criteria ... nowhere states, as the MAR court asserted, that interviewing other individuals ‘is important to making the diagnosis.’ ” Majority Op. at 237 (quoting J.A. 1562). The MAR court wrote, and Dr. Bellard agreed, that "[t]he testimony or observations of individuals with knowledge of defendant’s behavior and personality during defendant’s late adolescence and early adulthood is important to making the diagnosis arrived at by Dr. Bel-lard.” J.A. 1562; see J.A. 1753. The MAR court then commented, ‘‘This is noted by the manual Dr. Bellard himself utilizes, the ... DSM-IV.” Id. at 1562-63. The majority pounces on this line, saying that "the MAR court drastically mischaracterized the diagnostic requirements of the DSM-IV.” Majority Op. at 237. However, assuming the DSM-IV does not make the MAR court's point directly, the MAR court's statement was a fair inference from Dr. Bellard’s testimony and not remotely suggestive of any objective unreasonableness in the MAR court’s opinion. See J.A. 1753 (testimony of Dr. Bellard, admitting that the DSM-IV requires an evaluation of an individual’s long-term patterns and functioning; that paranoid personality disorder, the diagnosis given Gray by Dr. Bellard using the DSM-IV, usually manifests by late adolescence or early adulthood; and that it therefore would have been useful to talk to people who knew Gray during this time-period in his life).
. The majority speculates that Gray’s counsel "could have introduced expert testimony such as that of Dr. Bellard ...,” and then further speculates about what a jury "would have heard from such an expert.” Majority Op. at 236 (emphasis added). The majority cannot now manufacture a hypothetical expert super-witness who would have given Gray helpful testimony while somehow avoiding the deficiencies noted by the MAR court in its decision, after full consideration, to discount Dr. Bellard's opinion.