concurring in part and dissenting in part,
I am in accord with the majority on several points. I agree that this is an unusual case in which 28 U.S.C. § 2254(d) does not constrain our review of a state prisoner’s constitutional claim. See Campbell v. Bradshaw, 674 F.3d 578, 596 (6th Cir. 2012). I also share the view that Ohio law allows defendants facing a murder charge to introduce evidence of post-traumatic stress disorder (“PTSD”) in support of a voluntary-manslaughter conviction. See State v. Nemeth, 82 Ohio St.3d 202, 694 N.E.2d 1332, 1336 (1998). In light of Mr. Bruner’s decision to offer such evidence at Reddy’s sentencing hearing, I also accept the majority’s conclusion that his failure to do so during Reddy’s trial cannot be described as “sound trial strategy.” Strickland v. Washington, 466 U.S. 668, 689, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984) (quoting Michel v. Louisiana, 350 U.S. 91, 101, 76 S.Ct. 158, 100 L.Ed. 83 (1955)). At that point, however, the majority and I part ways. In my view, Reddy has not shown a “reasonable probability” that “the result of the proceeding would have been different” had the trial judge heard evidence of Reddy’s PTSD. Id. at 694, 104 S.Ct. 2052. I would therefore affirm the district court’s decision to deny Reddy’s bid for habeas relief.
As an initial matter, although the majority views Dr. Fabian’s report as clearly establishing that Reddy suffered from PTSD, the report is not quite so conclusive. Fabian did explain that he administered the Detailed Assessment of Post-Traumatic Stress (“DAPS”), a test that mental-health professionals use to detect the presence of PTSD symptoms. But the results did not conclusively show that Reddy had PTSD. Although Fabian concluded that Reddy was, “with reasonable psychological certainty, ... a mentally ill individual” who suffered from major depressive disorder and personality disorder, he never stated with that same degree of certainty that Reddy suffered from PTSD. Fabian concluded only that Reddy “presents as qualifying for many of the symptoms of PTSD,” which appear to be fairly nonspecific. Perhaps for this reason, unlike Reddy’s diagnoses of major depressive disorder and personality disorder, Fabian labeled his diagnosis of PTSD as “provisional” only.
*548Courts in several contexts have recognized that a “provisional” diagnosis is not as persuasive as an actual one, see, e.g., Lopez v. Ryan, 630 F.3d 1198, 1207 (9th Cir. 2011), overruled on other grounds by McKinney v. Ryan, 813 F.3d 798 (9th Cir. 2015) (en banc); Neumann v. Colvin, No. 3:13-CV-05723-RBL, 2014 WL 3671329, at *3 n.4 (W.D. Wash. July 22, 2014), and in adjudicating claims similar to Reddy’s, we have held that the Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), prejudice inquiry ordinarily demands more than a possibility that a petitioner has a relevant mental illness. In Campbell v. Coyle, 260 F.3d 531 (6th Cir. 2001), for example, we rejected a prisoner’s claim that he had been prejudiced by counsel’s failure to offer evidence of PTSD at trial. Although the petitioner supported his claim with reports from two psychologists who concluded that he “possessed many of the symptoms of PTSD,” id. at 549, we reasoned that without an actual diagnosis, the argument for prejudice was “much too tenuous,” id. at 555. Five years later, in Carter v. Mitchell, 443 F.3d 517 (6th Cir. 2006), we held that a doctor’s opinion, given after trial in support of a state postconviction motion, that “there [was] a ‘likelihood’ that [the prisoner] ha[d] ‘some kind of brain related difficulty”’ was not sufficiently conclusive to show that the prisoner had been prejudiced by his trial counsel’s failure to hire a neuropsychologist to evaluate him before trial. Id. at 529-30.
Although I recognize the “potential unfairness that would be created if conclusive evidence of PTSD were required for [a petitioner] to establish a viable ineffective-assistance claim,” Reddy’s claim is “no doubt weakened due to the want of a proper diagnosis.” Clarke v. Warren, 556 Fed.Appx. 396, 414 n.10 (6th Cir. 2014). This is particularly true because the probative value of Fabian’s report cannot be evaluated in isolation. The court’s psychologist evaluated Reddy as well and informed the court that Reddy suffered from major depressive disorder and acute stress disorder, but apparently not PTSD. In this context, there is reason to doubt whether Fabian’s testimony would have carried much weight had Bruner put the diagnosis at issue.
Even ignoring this concern, however, I still have difficulty concluding that Fabian’s testimony would create a “reasonable probability” of a different outcome because the trial judge’s own statements suggest that during Reddy’s trial, she already considered and rejected the possibility that the abuse that Reddy suffered caused him to react disproportionately to his mother’s knife attack. At the close of evidence, the judge explained that she did not believe “[t]he defendant’s rendition of what happened,” namely, that after years of abuse at Gloria’s hands, Reddy overreacted to her knife attack, flew into a fit of passion or rage, and choked her to death. When, at Reddy’s resentencing hearing, the court addressed Fabian’s report, which described a “nexus” between Reddy’s PTSD and Gloria’s killing, the court did not express surprise or curiosity, instead noting that “having been the trier of the facts in Mr. Reddy’s trial, I don’t think there is a lot of issue to take with” Fabian’s conclusions.
The trial judge’s muted reaction to Fabian’s report is unsurprising. In Clarke v. Warren, 556 Fed.Appx. 396 (6th Cir. 2014), we considered the habeas petition of a woman who killed two men in a robbery. Id. at 400. The prisoner claimed that she began to suffer from PTSD two weeks prior to the killings, as a result of being abducted and tortured by another individual unrelated to the robbery. Id. at 401, 414. We held that counsel’s failure to present PTSD evidence at the prisoner’s sentenc*549ing hearing was not prejudicial because the court “was aware of the trauma inflicted” on the prisoner, having heard “considerable evidence” of the abduction and assaults. Id. at 414. In other words, even though evidence of PTSD is different from evidence of abuse, see Amicus Curiae Br. 13-14, the difference is not so great as to prejudice the defense in every case in which PTSD evidence is omitted, see Clarke, 556 Fed.Appx. at 414. Tellingly, in those cases that Reddy cites, in which courts have considered trial counsel’s failure to introduce evidence of PTSD or similar disorders to be prejudicial, the trier of fact heard no evidence of the underlying abuse or trauma that gave rise to the PTSD. See Jacobs v. Horn, 395 F.3d 92, 101-02 (3d Cir. 2005); Seidel v. Merkle, 146 F.3d 750, 752 (9th Cir. 1998). The same cannot be said of Reddy’s trial.
The majority points out that in State v. Nemeth, 82 Ohio St.3d 202, 694 N.E.2d 1332 (1998), the Ohio Supreme Court held that a trial court’s refusal to admit evidence of a defendant’s diagnosis of battered child syndrome (“BCS”) was prejudicial even though the defendant testified to the abuse that he suffered at the hands of the victim. Id. at 1336, 1341. But the Nem-eth court reversed on the ground that the BCS evidence in that case was highly relevant to the defendant’s credibility. After reviewing the trial transcript, the Nemeth court observed that the jury likely assumed that the defendant had fabricated allegations of abuse because the prosecution repeatedly “implied] [at trial] that [the defendant] must have created the allegations of abuse after the fact because, otherwise, more people would have known about it.” Id. at 1337. The Ohio Supreme Court emphasized that “[t]he defense needfed] expert testimony to refute the seemingly logical conclusion that serious abuse could not be taking place if no one outside the home was aware of it.” Ibid. The same danger did not exist at Reddy’s trial, at which numerous witnesses testified that Gloria abused Reddy. Indeed, Bruner repeatedly drew the trial court’s attention to that testimony by arguing that because of the abuse, Gloria’s provocation prompted Reddy to enter a fit of passion or rage.
What is more, the reasons that the trial court gave for convicting Reddy of aggravated murder cast further doubt upon the majority’s assumption that PTSD evidence would have made a difference in this case. At the close of all proof, the trial court explained that it was rejecting Reddy’s version of events not because the record was devoid of evidence that would support. a conviction for either murder or voluntary manslaughter, but' rather because the great weight of evidence showed that the murder did not happen in the way Reddy alleged. The trial court explained that the lacerations on Gloria’s face and extensive bruising to her head, as well as the blood stains that Cleveland Police Department found throughout her home, strongly suggested that Reddy acted with “prior calculation and design” when he killed his mother:
The trouble the Court has in interpreting this evidence as being something other than prior calculation and design stems from the physical damage to the victim and the presence of blood spattering, not only in Joseph Reddy’s bedroom, but also in both walls of the hallway ... as well as the living room. Let’s not forget that we had some four or five markers when the detectives came to the home where they found blood on the living room carpeting, the plastic waste paper can, the spot on the drawing, so there were actually three different areas of the home where there are blood spatters.
*550The defendant’s rendition of what happened never mentions any of the head injuries and when you combine the evidence of the head injuries and all those blood splatters, it strikes the Court as consistent with the beating of the victim as having occurred prior to any choking, so the pattern the evidence presents to the Court and of which I’m convinced beyond a reasonable doubt is whatever the victim may have done to incite or enrage her assailant, Mr. Red-dy, the plain fact of the matter is that the sequence of events had to have occurred with the beating of the victim’s head in rooms—in various rooms, hallway, living room, Joseph’s bedroom. And that had to have ... preceded ... her choking.
The probative value that the trial court assigned to this physical evidence of a slow and drawn-out killing strongly suggests that evidence of Reddy’s PTSD would have had no impact on the trial court’s decision to convict Reddy of aggravated murder. Even assuming that the “[e]xaggerated startle response” that PTSD causes can spur individuals to kill in as methodical a manner as the trial court was convinced Gloria was murdered, Am. Psychiatric Ass’n, Diagnostic and' Statistical Manual of Mental Disorders 272 (5th ed. 2013), the result at trial would have been no different had PTSD evidence been admitted. Reddy told police that as Gloria was attacking him, he punched her in the face in self-defense, tackled her, and choked her in an attempt to render her unconscious, inadvertently killing her. At trial, Reddy’s theory of the case relied on the truth of this account; Bruner argued that Reddy began to choke his mother immediately after she attacked him, and did not realize that the choking was killing her because he was acting under a passion brought on by the combination of Gloria’s knife attack and years of abuse at her hands. But in his statement to police, Reddy made no mention of beating Gloria in multiple rooms, which, given the evidence of blood stains throughout the home and extensive injuries to Gloria’s head, convinced the- trial court that Reddy was lying about how the killing happened. Evidence of PTSD would not have changed that conclusion.
I acknowledge that Reddy’s case for ha-beas relief would be stronger if the trial court had convicted Reddy of murder after having been instructed to disregard the blood-stain evidence and reconsider its verdict. But that never happened. Despite having concluded that the trial judge’s reliance on blood-stain evidence was reversible error, the Ohio Court of Appeals did not allow Reddy a second opportunity to argue for voluntary manslaughter. See State v. Reddy, 192 Ohio App.3d 108, 948 N.E.2d 454, 460-62, 468 (2010). Although I have doubts about the appropriateness of the state court’s reasoning on that point— which is not at issue in this case—PTSD evidence would not have changed the result in the, Ohio Court of Appeals. As the majority has explained, that court applied a state-law presumption that “the trial court considered all lesser and included offenses as well as inferior degree offenses,” and then explained that it would not remand the case for consideration of voluntary manslaughter because the evidence in the record “was sufficient to convict Reddy of murder.” Id. at 468. I do not see how evidence that Reddy had PTSD would have weakened the case for deploying the presumption, or how Fabian’s testimony—which a trier of fact could easily discount—would in any way have rendered the evidence “insufficient” to convict Red-dy of murder.
In sum, I agree with the majority that if Bruner concluded that Fabian’s report was favorable enough to submit to the court for sentencing purposes, there appears to be *551no reason for why he should have thought it strategic to avoid calling Fabian as a witness at trial. But the particular facts of this case suggest that there is less than a “reasonable probability” that, but for Bruner’s decision not to call Fabian to the witness stand, the “result of the proceeding would have been different.” Strickland, 466 U.S. at 694,104 S.Ct. 2052. The majority having taken a different view, I respectfully dissent.