Jedidiah Isaac Murphy was convicted and sentenced to death by a Texas jury for the capital murder of Bertie Cunningham. After unsuccessfully pursuing state appellate and habeas remedies, Murphy brought a federal habeas petition, which eventually was denied. We recently granted Murphy a certificate of appealability to appeal the denial of two of his federal habeas claims. See Murphy v. Davis , 732 Fed.Appx. 249, 251-52 (5th Cir.2018) (per curiam). Murphy's first claim alleges that his trial counsel was constitutionally deficient during the penalty phase of trial by failing to correct a potentially misleading impression created by one of his experts. Murphy's second claim alleges the State suppressed *583material impeachment evidence of a pretrial conversation between a State witness and the lead prosecutor in his case. Upon full review, we agree with the district court that these claims are either procedurally barred or meritless. We AFFIRM.
I.
A.
After robbing 80-year-old Bertie Cunningham at gunpoint, Jedidiah Isaac Murphy forced her into the trunk of her own car and shot her in the head. He then drove around with her body in the trunk, using her ATM card and credit cards to buy beer and liquor. Murphy was soon arrested. He admitted to the shooting and led police to the creek where he had dumped Cunningham's body. Later at the police station, he wrote and signed a statement claiming that he accidentally shot Cunningham while forcing her into her own trunk.
In June 2001, a Texas jury convicted Murphy of capital murder. The State of Texas sought the death penalty. During the penalty phase of the trial, the sides clashed over the future threat to society Murphy would pose if allowed to live.
In particular, the severity of Murphy's history of violence was a point of contention. To demonstrate he had such a history, the State submitted Murphy's record of theft convictions. A responding officer also testified for the State about a domestic-abuse call involving Murphy and his girlfriend. The officer said that when he entered Murphy's home, the girlfriend had a bloody nose and Murphy had a knife. The officer subdued Murphy with pepper spray. Another State witness said that Murphy pulled a gun on her at a high school party. He put the gun to her head, asked if she was afraid to die, and held it there for a minute. One of Murphy's coworkers also testified for the State. She claimed that Murphy talked about having access to guns, bragged about shooting people, and threatened to "knock [her] fucking head off." The woman was so frightened that she quit her job and reported Murphy to the police.
Along with this, the State tried to implicate Murphy in a three-year-old kidnapping case. Sheryl Wilhelm testified for the State that, three years before the Cunningham killing, a man briefly kidnapped her and then stole her car. After seeing a TV news report on Cunningham's murder featuring Murphy's photo, Wilhelm called the police to report Murphy as her kidnapper. She identified Murphy during a pretrial hearing and then again at trial. Wilhelm also testified at trial that she identified Murphy in a police-constructed photo lineup. The detective who conducted the photo lineup testified that Wilhelm's "was one of the better photo" identifications he ever had. According to the detective, Wilhelm said "she was virtually sure that that was the guy who abducted her."
Murphy attacked Wilhelm's identification in a few ways. He called a psychologist who testified that Wilhelm's memory was tainted by the photo of Murphy she saw on the news. The psychologist also pointed out prominent differences between a composite sketch, made just a week after the kidnapping, and the press-released photo of Murphy. And the psychologist added that the photo lineup was unfairly constructed; obvious differences between the mugshots reduced the odds of selection from one-in-six to one-in-three. Murphy also put on an alibi defense. Wilhelm said she had been kidnapped, escaped, and had her car stolen at 11:30 a.m. in Arlington, Texas. The day after her kidnapping, Wilhelm's car was found in Wichita Falls, Texas. In the car, the police found documents belonging to another woman. That *584woman had been assaulted and had her purse stolen in Wichita Falls at 8:24 p.m. on the day of Wilhelm's kidnapping. Also on the same day, Murphy clocked in for his night shift at 11:54 p.m. in Terrell, Texas. Murphy's counsel argued that Murphy did not have time to kidnap Wilhelm in Arlington, rob the other woman in Wichita Falls, and make it to work in Terrell.
The trial did not just focus on Murphy's dangerousness. Murphy claimed that mitigating circumstances reduced his moral blameworthiness. To make his case, Murphy called, among others, a psychologist named Dr. Mary Connell. Dr. Connell had interviewed Murphy three times, interviewed Murphy's family and other acquaintances, and reviewed his records. From this, Dr. Connell was able to testify in detail about Murphy's background. She explained that Murphy's father was an abusive alcoholic. By seven, both of Murphy's parents had abandoned him and he was put in the foster-care system. In the system, Murphy went through five families. His first adoptive father hit and screamed at him. His second adoptive family broke up. As Murphy grew older, he became an alcoholic and he started to feel like he was falling into his father's pattern of abuse. He attempted suicide and sought out psychiatric treatment for depression, psychosis, and anxiety.
Drawing on what she had learned, Dr. Connell testified that Murphy "is generally described by people as a warm, outgoing, loving kind of person." Dr. Connell added that Murphy expressed remorse for his crime and that "he talked about Ms. Bertie Cunningham in almost a reverent or awed way." Based on his early childhood abuse and abandonment, Murphy became self-loathing. Per Dr. Connell, Murphy's drinking was driven by "a genetic predisposition," a desire to temporarily feel better about himself, and "his identification with his father." Like his father, he saw himself as nothing but a worthless drunk. Dr. Connell did admit, however, that Murphy is "unpredictable," "moody," and "impulsive"-behaviors that intensify when he is drinking. According to Dr. Connell, Murphy is intermittently violent. "[H]e could maintain an even keel for a period of a month or two ... but then something would set him off and he would go on another binge, get aggressive, angry, loud, belligerent, and things would spiral downward and out of control."
Dr. Connell also testified that she gave Murphy two tests: the Minnesota Multiphasic Personality Inventory-II (MMPI-2) and the Millon Clinical Multiaxial Inventory-III (MCMI-3). As background, the MMPI-2 and MCMI-3 consist of 567 and 175 true-false questions, respectively. For both tests, the subject's answers are fed through a database. A computer program, using group statistical data, then returns a profile on the subject. Upon request, an interpretative report-which supplies further hypotheses about the subject-may also be returned. Per Dr. Connell, the MMPI-2 is the "flagship" personality-assessment test and the MCMI-3 assesses the subject for character problems.
Murphy's MMPI-2 profile and interpretative report showed, according to Dr. Connell, that Murphy exhibited signs of depression, anxiety, physical ailments, and paranoid thoughts. At first, Dr. Connell thought Murphy might be exaggerating his symptoms-a fact suggested by elevated results on the MMPI-2's "lie scale." But when she looked at his answers and compared them to the interviews they had, she concluded Murphy was not lying. Turning to Murphy's MCMI-3, the test result's suggested, per Dr. Connell, that Murphy was "deeply depressed with an agitated edge." His thoughts, according to Dr. Connell, *585would shift between suicide, hopelessness, and futility "to occasional outbursts of bitter discontent or irrational demands." Murphy's results on both tests, said Dr. Connell, would normally prompt referral for psychiatric consultation and probably indicate a need for medication.
It is important here to note that neither test is designed to give a final interpretation of the test subject. Rather, as explained by the interpretative reports themselves (which were introduced by the State into evidence) and by Dr. Connell, the tests render hypotheses. It is also important to note that no psychologist, besides Dr. Connell, was directly involved in administering or interpreting Murphy's MMPI-2 and MCMI-3. The tests instead draw on computer algorithms constructed by other psychologists. As Dr. Connell explained on direct, she gave the tests to Murphy and "scored the results with a computer system" which looked at how he "compared to other people in a clinical setting."
Despite this, when cross-examined, Dr. Connell agreed that Dr. James Butcher-"probably the leading expert in the country on the interpretation of the MMPI"-had "interpreted" Murphy's MMPI-2. Strictly speaking, Dr. Butcher had only developed the computerized interpretative system which read Murphy's answers and generated an interpretative report. Dr. Butcher's name therefore appeared on the interpretative report for the MMPI-2. But Murphy had never interacted with Dr. Butcher, and Dr. Butcher never reviewed Murphy's MMPI-2 answers, profile, or interpretative report. That distinction did not stop the State from referring to the MMPI-2 interpretative report "as the report of Dr. James Butcher." Nor did it stop Dr. Connell from agreeing that the hypotheses contained in MMPI-2 report were the "statements of Dr. Butcher."
The State proceeded to read off some of the MMPI-2 interpretative report's unfavorable hypotheses, referring to them as Dr. Butcher's "statements." Per the State, Dr. Butcher stated that Murphy exaggerated his symptoms and responded to the last section of the MMPI-2 "either carelessly, randomly, or deceitfully, thereby invalidating that portion of the test." The State continued, reading off that Murphy "has serious problems controlling his impulses and temper," "loses control easily," and may be "assaultive." Murphy, according to the parts read aloud, "manipulates people" and lacks "genuine interpersonal warmth." According to the report, Murphy matches the profile of a Megargee Type H offender, a seriously disturbed inmate type. Inmates with Murphy's profile will, per the report, "not seek psychological treatment on their own" and are "poor candidates for psychotherapy."
Dr. Connell pushed back on the State's use of the interpretive report. She clarified that the report only gave hypotheses, one of which "we know ... is wrong"; Murphy had sought psychological treatment on his own. She directed the jury to the beginning of the report, which said that the offered interpretation is not meant to be final and that "interview, observation, and history should be taken into account." She also explained that based on her interviews with Murphy, she did not believe he was exaggerating his symptoms. She did not, however, clarify that it was only Dr. Butcher's computer program-not the doctor himself-that read Murphy's 567 true-false answers.
The State moved to the MCMI-3 interpretative report, calling it the "report produced by Dr. [Theodore] Millon" (again, without clarification from Dr. Connell). Dr. Millon is an authoritative figure on the MCMI-3, and his name is affixed to the MCMI-3 interpretative report. The *586prosecution elicited from Dr. Connell that through the MCMI-3, Dr. Millon himself had stated that Murphy "may have reported more psychological symptoms than objectively exist," and Murphy has "a moderate tendency toward self-deprecation and a consequent exaggeration of current emotional problems." Dr. Connell again clarified that she did not "know that these results should be considered definitive for diagnosis."
On redirect, the defense did not ask Dr. Connell if Drs. Butcher and Millon were personally involved with Murphy in any way. Instead, counsel had Dr. Connell explain what use she made of the tests and interpretative reports. Dr. Connell explained that she administered the tests to help her "gain an understanding of [Murphy's] view of his own functioning and an understanding of how he compares to other people in similar situations." Dr. Connell added, "I didn't administer them for the diagnosis or treatment of a disorder, but just to give myself a sort of objective and normative feel for who it was that I was attempting to understand." Counsel elicited that Dr. Connell's ultimate conclusions about Murphy were informed by the tests and reports, and that despite the reports, she thought Murphy was truthful, was not blaming others, and was "suffering enormously over the guilt for what he had done." The defense also had Dr. Connell read the parts of the reports the prosecution "left out." Dr. Connell read off the reports that Murphy's alcohol abuse was caused by frustration and disappointment, that his inappropriate behavior manifests when drinking, and that he has genuine feelings of guilt and remorse.
During opening summation, the State emphasized the "chilling" results of Murphy's MMPI-2 and MCMI-3, which it invited the jury to read. Specifically, the State pointed out that Murphy's profile matched that of a Megargee Type H offender-"one of the most seriously disturbed inmate types" for whom "[a]djustment to prison appears to be difficult." During its summation, the defense did not mention the MMPI-2 or MCMI-3. During the State's rebuttal summation, it returned to the tests:
You know, all you have to do if you really have a question about what this guy's going to do in prison, if you look at the defense's own expert-now, these are not people that the State of Texas hired on his behalf, but you look at Dr. James Butcher who the defense hired and this report was brought out on cross-examination. And you remember what Dr. Butcher said? You know, he's the doctor, I suppose, who's not opposed to the death penalty. Here is what he says about the defendant. He says this man right here is a poor candidate for psychotherapy. Individuals with his profile are not very amenable to changing their behavior. You have a litany illustrating his behavior. He goes on to say this, they tend to be quite aggressive. And finally, if you have any question about what this man is all about in a confined setting, adjustment to prison appears to be difficult for them. Those aren't my words, ladies and gentlemen. That's not some expert that we hired. That's Dr. James Butcher hired by the defense to look at the tests administered to this man over here. So even if I look at that set alone, which each and every one of you told us you weren't going to do, but even if you do that, I mean, their own expert says, that ain't going to fly in this case. This man is going to be a danger wherever he's going to be.
The defense did not object to this line of argument.
The jury found that Murphy was a continuing threat to society and there were *587insufficient mitigating circumstances to justify life in prison. See Tex. Code Crim. Proc. Ann. art. 37.071, § 2. Based on these two findings, Murphy was sentenced to death.
Murphy's conviction and sentence were affirmed on direct appeal. Murphy v. State , 112 S.W.3d 592, 595 (Tex. Crim. App. 2003). Likewise, his first state habeas application was denied. Ex parte Murphy , No. WR-70,832-01, 2009 WL 766213, at *1 (Tex. Crim. App. Mar. 25, 2009) (per curiam) (not designated for publication).
B.
In 2009, Murphy's new lawyer cold called Sheryl Wilhelm and asked her what happened during the photo lineup where she identified Murphy as her kidnapper. Wilhelm said that when she identified Murphy, she told the detective: "This is him. This looks a lot like him, and I'm pretty sure it's him." She also stated that: "You know, nobody's ever 100 percent sure.... I'm talking about anything in this life but, I mean, to me, it was him. I mean, 95 to 100 percent it was him." Wilhelm also disclosed to Murphy's lawyer that the lead prosecutor in Murphy's case came to her house before trial. At her house, Wilhelm said that she asked the prosecutor "was that the guy?" He responded, "Yes." Following up on what she meant, Murphy's lawyer asked Wilhelm: "when you asked [the prosecutor] if you got the right guy, and he said that you did, it confirmed in your mind the accuracy of your identification?" Wilhelm responded, "Right." Murphy's lawyer then asked if the prosecutor "hadn't told you that, would you have retained any, perhaps, uncertainty about it?" Wilhelm responded, "No." "I mean, I knew it was him. Yeah I knew it was him. I think that happened. I swear it's been so long ago, but I'm pretty sure that I asked him, 'Was this the guy?' ... And he nodded his head so, I don't know."
Soon after this phone exchange, Murphy filed a federal habeas petition. The district court stayed the proceedings to give Murphy time to exhaust three sets of claims in the state system: (1) suppression of evidence impeaching Wilhelm's identification,1 (2) ineffective assistance of trial counsel at the guilt phase, and (3) ineffective assistance of trial counsel at the penalty phase.
Following the stay, Murphy filed a second state habeas application. The Texas Court of Criminal Appeals (TCCA) dismissed Murphy's two ineffectiveness claims as abuses of the writ. Ex parte Murphy , No. WR-70,832-02, 2010 WL 3905152, at *1 (Tex. Crim. App. Oct. 6, 2010) (per curiam) (not designated for publication). With respect to Murphy's suppression claim, the TCCA remanded to the convicting trial court with instructions to determine whether the claim was procedurally barred and, if not, whether it had merit. Id.
The state trial court held an evidentiary hearing where it heard testimony from Sheryl Wilhelm, the detective who administered the photo lineup, the lead prosecutor, and Murphy's lead trial counsel. At the hearing, Murphy's counsel elicited from Wilhelm that, at the pretrial interview, the following occurred: Wilhelm asked the lead prosecutor whether she "had identified the right man in the photo spread," the prosecutor told her she had, and this confirmed in her mind the accuracy of her identification of Murphy. On cross-examination, Wilhelm testified that, in fact, she had asked the lead prosecutor *588whether she had identified the same person who was arrested for Bertie Cunningham's murder, not whether her identification of the person who kidnapped her was correct. Wilhelm said she understood the distinction between the two concepts. She added that when she made her in-court identification she was thinking about the person that who kidnapped her.
When examined at the evidentiary hearing, the lead prosecutor denied that he confirmed the accuracy of Wilhelm's photo-lineup identification. He said that as best he could recall, he told Wilhelm that he was "there to prosecute a capital murder" and that he believed "that the same person who had kidnapped [Wilhelm] had abducted and killed [Cunningham]." He swore that he "would have done everything he could not to inform her about the validity of her photo identification." He maintained that he did not believe that he ever told Wilhelm "I believe you have the right person." When asked if he thought he might have affirmatively responded to the question "did I pick the right guy out of the photo spread?" he responded, "No. I don't recall it being in response to that. I think that would've been in response to do you think that you have the same person who kidnapped me."
After the hearing, the trial court found that Murphy's claim should be dismissed as an abuse of the writ and alternatively denied as meritless. To do so, it entered two sets of factual findings: one related to the procedural bar and the other related to the merits. With respect to the procedural bar, the trial court found that Murphy reasonably could have ascertained the factual basis for his suppression claim before his first state habeas application was filed by asking Wilhelm questions during her pretrial or trial cross-examination. With respect to the merits, the court found that Wilhelm was not certain what was said at the pretrial interview, that the "prosecutor purposefully avoided doing things that might influence or taint Wilhelm's identification," and that the pretrial conversation "would not have altered Wilhelm's courtroom identification." The court added that the defense presented more credible attacks on the identification than the pretrial interview and that the "best evidence" of future dangerousness "was the primary offense"-that is, Cunningham's murder. Based on the trial court's findings, the TCCA concluded that Murphy's application was an abuse of the writ and dismissed his application "without considering the merits of the [his] claims." Ex parte Murphy , No. WR-70,832-02, 2012 WL 982945, at *1 (Tex. Crim. App. Mar. 21, 2012) (per curiam) (not designated for publication).
Murphy returned to federal court and raised the three sets of now exhausted claims, among others. The district court denied Murphy relief on all of his claims, finding them procedurally barred and alternatively meritless. It also denied his request for an evidentiary hearing. Murphy then sought a certificate of appealability (COA) from us to appeal all three sets of claims. We obliged him on parts of two separate claims:
(1) Ineffective assistance of trial counsel at the penalty phase arising from his counsel's failure to correct a false impression created by Dr. Connell; and
(2) Suppression of evidence, specifically the pretrial conversation where the lead prosecutor allegedly confirmed to Wilhelm that she got the right *589guy.2
Murphy , 732 Fed.Appx. at 255-56.
Our task now is to decide whether either claim justifies reversal of the district court. We conclude that neither does.
II.
Murphy claims that his trial counsel was constitutionally ineffective at the penalty phase. The specific defect in counsel's representation was, according to Murphy, not eliciting from Dr. Connell on redirect that it was Drs. Butcher's and Millon's computer algorithms-not the doctors themselves-that generated the damaging MMPI-2 and MCMI-3 interpretative reports.3 He asserts that the State weaponized any jury confusion over the reports' origins by stressing in summation that leading psychologists, hired by the defense, had concluded that Murphy could not be rehabilitated and would be dangerous in prison. Murphy also claims he needs an evidentiary hearing to develop whether counsel knew of the reports' origins and, if so, what reason they may have had for not exposing them. We agree with the district court that no evidentiary hearing is needed and this claim is meritless.4
A.
Murphy's ineffectiveness claim is governed by the familiar Strickland standard. To prevail, Murphy must show: (1) that his trial counsel's performance was deficient, and (2) that the deficient performance resulted in prejudice. See Strickland v. Washington , 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).
Strickland 's first prong "sets a high bar." Buck v. Davis , --- U.S. ----, 137 S.Ct. 759, 775, 197 L.Ed.2d 1 (2017). "To demonstrate deficient performance, the defendant must show that, in light of the circumstances as they appeared at the time of the conduct, 'counsel's representation fell below an objective standard of reasonableness' as measured by 'prevailing professional norms.' " See Rhoades v. Davis , 852 F.3d 422, 431-32 (5th Cir.2017)
*590(quoting Strickland , 466 U.S. at 687-88, 104 S.Ct. 2052 ). We "must apply a 'strong presumption' that counsel's representation was within the 'wide range' of reasonable professional assistance." Harrington v. Richter , 562 U.S. 86, 104, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011) (quoting Strickland , 466 U.S. at 689, 104 S.Ct. 2052 ).
To satisfy Strickland 's second prong, "[t]he defendant must show that there is 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." Strickland , 466 U.S. at 694, 104 S.Ct. 2052. "The likelihood of a different result must be substantial, not just conceivable." Richter , 562 U.S. at 112, 131 S.Ct. 770. We must "consider all the relevant evidence that the jury would have had before it if [trial counsel] had pursued the different path." Wong v. Belmontes , 558 U.S. 15, 20, 130 S.Ct. 383, 175 L.Ed.2d 328 (2009) (per curiam).
We will consider both Strickland prongs, reviewing the district court's conclusions of law de novo and factual findings for clear error. See Woodfox v. Cain , 609 F.3d 774, 788-89 (5th Cir.2010). But before we do so, we must decide whether the district court's refusal to grant Murphy an evidentiary hearing was proper. We determine that it was.
B.
Federal courts faced with habeas petitioners' discovery requests have, in some circumstances, a duty "to provide the necessary facilities and procedures for an adequate inquiry." See Gibbs v. Johnson , 154 F.3d 253, 258 (5th Cir.1998) (quoting Harris v. Nelson , 394 U.S. 286, 299, 89 S.Ct. 1082, 22 L.Ed.2d 281 (1969) ); see also Bracy v. Gramley , 520 U.S. 899, 908-09, 117 S.Ct. 1793, 138 L.Ed.2d 97 (1997). This is the case when "specific allegations before the court show reason to believe that the petitioner may, if the facts are fully developed, be able to demonstrate that he is ... entitled to relief." See Murphy v. Johnson , 205 F.3d 809, 814-15 (5th Cir.2000) (alteration in original) (quoting Gibbs , 154 F.3d at 258 ). But we need not allow "fishing expeditions." See Ward v. Whitley , 21 F.3d 1355, 1367 (5th Cir.1994).
When a habeas petitioner requests an evidentiary hearing, district courts have discretion over whether to grant one. See Schriro v. Landrigan , 550 U.S. 465, 468, 127 S.Ct. 1933, 167 L.Ed.2d 836 (2007).5 A district court does not abuse its discretion when denying an evidentiary hearing if it had "sufficient facts before it to make an informed decision on the merits." See McDonald v. Johnson , 139 F.3d 1056, 1060 (5th Cir.1998). For instance, no abuse of discretion occurs when, assuming "the truth of all the facts" the petitioner seeks "to prove at the evidentiary hearing," we are confident that "he still could not be granted federal habeas relief." Schriro , 550 U.S. at 481, 127 S.Ct. 1933.6
Murphy argues that the district court should have granted him an evidentiary *591hearing. Such a hearing would, according to Murphy, reveal whether trial counsel were aware that Drs. Butcher and Millon did not evaluate his test answers and did not directly author the reports. If counsel were not aware, then, per Murphy, they could not have strategically decided not to correct any false impression. And if they were aware, then exploration of their decision not to expose the prosecutor's misleading framing would be needed to determine whether their decision making was strategically sound.
Contra Murphy, we conclude that an evidentiary hearing is not necessary. To reach our conclusion, we proceed based on the assumption that a hearing would show what Murphy wants-that trial counsel did not know the MMPI-2's and MCMI-3's origins.7 See id. Proceeding with this favorable assumption, we are still confident Murphy could not show deficiency or prejudice.
1.
Counsel did not render deficient performance, whether they knew of the MMPI-2's and MCMI-3's origins or not. Broadly stated, counsel "has a duty to bring to bear such skill and knowledge as will render the trial a reliable adversarial testing." Strickland , 466 U.S. at 688, 104 S.Ct. 2052. In preparing for the penalty phase of a death penalty trial, "counsel must either (1) undertake a reasonable investigation or (2) make an informed strategic decision that investigation is unnecessary." See Charles v. Stephens , 736 F.3d 380, 389 (5th Cir.2013). "In other words," counsel must "make reasonable investigations" or "a reasonable decision that makes particular investigations unnecessary." Strickland , 466 U.S. at 691, 104 S.Ct. 2052. When the omission alleged is failing to investigate something in particular, we look at "the known evidence" and whether it "would lead a reasonable attorney to investigate further." Wiggins v. Smith , 539 U.S. 510, 527, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003). Working through counsel's investigation, what they learned, and choices they made demonstrates their competence.
In an affidavit submitted at the behest of Murphy's current counsel, Murphy's lead trial counsel identified a key constraint she faced: "regardless of the interpreter of the tests, the results were what they were." Counsel understood that most of their evidence was going to be "double-edged"-meaning it was both mitigating and showed Murphy's dangerousness. Counsel calculated that Dr. Connell's testimony about Murphy's upbringing and problems it caused in his life might lead the jury to "weigh in more heavily on mitigation instead of aggravation."
At trial, the defense executed this mitigation-focused strategy. Dr. Connell's testimony painted a sympathetic picture of Murphy. She established that, as a child, Murphy was abused, abandoned, and mistreated. To cope with feeling unlovable, Murphy drank. Doing so, he started to fall into the same pattern of abuse his father *592displayed. This, admittedly, meant that Murphy was intermittently violent when drinking and could "spiral downward and out of control." But after, Murphy would express genuine remorse and guilt for his wrongs, including killing Bertie Cunningham. On cross-examination, counsel watched as their own expert-who had administered the tests and reviewed the interpretive reports-agreed with the State's characterization of the interpretative reports as Dr. Butcher's and Millon's. Counsel also watched Dr. Connell push back in a different way-pointing out that the reports only generated hypotheses. On redirect, counsel pressed forth with their mitigation-focused strategy while also countering the prosecution's use of the MMPI-2 and MCMI-3. Counsel elicited that the tests' reports had a limited purpose and drew out the reports helpful parts. Counsel had Dr. Connell point out that the reports indicated that Murphy's alcoholism was a product of self-resentment and guilt. And counsel elicited that Dr. Connell's sympathetic conclusions about Murphy were drawn, in part, from the reports.
In light of counsel's selected mitigation-focused strategy and their known counter to the interpretative reports, it was reasonable not to investigate the extent of Drs. Butcher and Millon's involvement. Figuring out this single detail about the reports might "distract[ ] from more important duties," see Bobby v. Van Hook , 558 U.S. 4, 11, 130 S.Ct. 13, 175 L.Ed.2d 255 (2009) (per curiam), especially because, as counsel explained, the reports "were what they were." We must also be wary of "the distorting effects of hindsight." See Strickland , 466 U.S. at 689, 104 S.Ct. 2052. Looking backward, we can see the precise way the State weaponized the reports in rebuttal summation. It certainly would have been optimal for the defense to have preempted the State's eventual spin. But it was not obvious beforehand that the State would go down this path-especially in light of Dr. Connell's repeated emphasis that the reports only gave hypotheses. Cf. Maryland v. Kulbicki , --- U.S. ----, 136 S.Ct. 2, 4-5, --- L.Ed.2d ---- (2015) (per curiam) (noting that lawyers generally need not "go 'looking for a needle in a haystack,' " especially "when they have 'reason to doubt there is any needle there' " (quoting Rompilla v. Beard , 545 U.S. 374, 389, 125 S.Ct. 2456, 162 L.Ed.2d 360 (2005) ) ). And in any event, counsel's performance need not be optimal to be reasonable. See Richter , 562 U.S. at 104, 131 S.Ct. 770. Murphy was entitled to "reasonable competence, not perfect advocacy." See Yarborough v. Gentry , 540 U.S. 1, 8, 124 S.Ct. 1, 157 L.Ed.2d 1 (2003) (per curiam).
Even more vitally, counsel was entitled to rely on Dr. Connell to alert her to the potential misapprehension over Drs. Butcher and Millon's involvement. Of course, hiring an expert and having her testify does not give counsel license to "completely abdicate ... responsibility." See Turner v. Epps , 412 F. App'x 696, 704 (5th Cir.2011) (per curiam).8 But "counsel should be able to rely on that expert to alert counsel to additional needed information or other possible routes of investigation." See id. Here, Dr. Connell repeatedly failed to challenge the State's framing regarding Drs. Butcher and Millon's involvement. Dr. Connell's failure in this regard stands out, given her push back on other issues. In light of this, counsel was entitled to "rely upon the objectively reasonable *593evaluations and opinions of" Dr. Connell. See Segundo v. Davis , 831 F.3d 345, 352 (5th Cir.2016) (quoting Smith v. Cockrell , 311 F.3d 661, 676-77 (5th Cir.2002), overruled on other grounds by Tennard v. Dretke , 542 U.S. 274, 124 S.Ct. 2562, 159 L.Ed.2d 384 (2004) ). "Without a red flag ... it is too much to insist that counsel second-guess her" expert's testimony. See (Patrick) Murphy v. Davis , No. 17-70030, --- Fed.Appx. ----, ----, 2018 WL 2945900, at *12 (5th Cir. June 11, 2018) (per curiam).9
Assuming what Murphy seeks to prove, trial counsel still would not have rendered deficient assistance. Counsel made an informed strategic choice to adopt a mitigation-focused strategy, lessening the need to counter the State's potential spin on the reports. Counsel also already discovered a convincing counter to any such spin. Plus, counsel's expert, whom they were entitled to reasonably rely upon, did not alert them to any issues. And it was not prospectively apparent what any further investigation would turn up or what its value would be. Whether counsel knew of the test's origins or not, their overall performance would still fall "within the 'wide range' of reasonable professional assistance." See Richter , 562 U.S. at 104, 131 S.Ct. 770.
2.
Murphy also falls short on the prejudice prong-a problem an evidentiary hearing cannot change. To show prejudice, Murphy must perform a delicate balancing act. He must show that the jury gave a good deal of weight to the reports' hypotheses. He must also show that exposing the reports' true origins would then meaningfully change that assessment. This is a tough balance to strike. The jury heard that the reports only gave hypotheses not conclusions. The jury heard that one of those hypotheses-that Murphy would not seek psychological treatment on his own-was conclusively wrong. And it heard that another hypothesis fell apart after Dr. Connell's interviews. Assuming, as we must, that the jury fairly weighed this evidence, it would be difficult for it to place much faith in the reports, regardless of how they came to be. Cf. Strickland , 466 U.S. at 695, 104 S.Ct. 2052 (holding that the prejudice inquiry assumes "that the decisionmaker is reasonably, conscientiously, and impartially applying the standards that govern the decision" and that the inquiry does "not depend on the idiosyncra[s]ies of the particular decisionmaker").
But if, despite all this, the jury was still inclined to give weight to the reports, revealing the degree of Drs. Butcher and Millon's involvement (or lack thereof) would not change much. It is unlikely that the jury would put less stock in the reports based on the realization that a computer, not a person, scored a true-false exam. The jury still would have heard that computer programs devised by leading psychologists had generated disturbing hypotheses about Murphy. And the reports would retain an imprimatur of neutrality because Murphy's expert administered the tests *594and computer programs returned the results. Given all this, Murphy cannot prevail. It is hard to conclude that the jury gave the reports much weight at all. It strains credulity to believe that, had the jury given the reports weight, merely clearing up that the tests were graded by computer, not hand, would change its opinion.
In any event, no matter how the jury viewed the reports, the historical evidence of Murphy's dangerousness was far more damaging. The jury heard that Murphy kidnapped an elderly woman, shot her point blank in the head, drove around with her body in the trunk as he used her credit cards, and eventually dumped her in a creek. It heard that Murphy had put a gun to the head of a woman at a high school party and asked if she was afraid to die. It heard that a police officer had to subdue him while he was wielding a knife after he hit his girlfriend. It heard that he had repeatedly threatened to shoot his coworker. It heard from Dr. Connell that Murphy was intermittently violent. Of course, it might be "conceivable" that the jury was swayed by a misapprehension over Drs. Butcher and Millon's involvement. See Richter , 562 U.S. at 112, 131 S.Ct. 770. But looking over "all the relevant evidence that the jury would have had before it," see Belmontes , 558 U.S. at 20, 130 S.Ct. 383, the chances of a different result are not "substantial," see Richter , 562 U.S. at 112, 131 S.Ct. 770.
* * *
As it stands, the record does not show deficient performance or prejudice. And assuming what Murphy seeks to prove at an evidentiary hearing-that counsel was unaware of the reports' origins-Murphy would still fail to make out a case for deficient performance. Plus, Murphy does not even argue that a hearing would save his case for prejudice. Thus, the district court did not abuse its discretion by denying Murphy an evidentiary hearing and it properly denied Murphy's Strickland claim on the merits.
III.
Murphy's second claim is that the State suppressed evidence that could have been used to impeach Sheryl Wilhelm, the State witness who identified Murphy as her kidnapper. Specifically, he complains that the State did not disclose a pretrial interview between the lead prosecutor and Wilhelm where the prosecutor allegedly confirmed the accuracy of her identification. This conversation, according to Murphy, could have been used to impeach Wilhelm's identification, knocking out a key part of the State's case for future dangerousness. We agree with the district court that this Brady v. Maryland , 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), claim is procedurally barred and meritless.
A.
Before considering the validity of Murphy's Brady claim, we must determine our standard for reviewing the state trial court's relevant factual conclusions. This requires recitation of some procedural history.
Murphy's Brady claim was first raised in his second state habeas application. The TCCA remanded his Brady claim to the convicting state court. Murphy , 2010 WL 3905152, at *1. The TCCA instructed the trial court to first decide if Murphy's Brady claim should be dismissed as an abuse of the writ. Id. Specifically, it tasked the trial court with deciding whether the factual basis for the Brady claim was ascertainable through the exercise of reasonable diligence before or on the date the original state habeas application was filed. Id. ; see *595Tex. Code Crim. Proc. Ann. art. 11.071, § 5(a)(1), (e) (setting forth the abuse of the writ standard). If the Brady claim was not abusive, the TCCA instructed the trial court to consider its merits. Murphy , 2010 WL 3905152, at *1.
On remand, the state trial court held an evidentiary hearing and eventually issued written findings of fact and conclusions of law. It found Murphy's claim was procedurally barred and meritless. Within the trial court's written ruling, it made two sets of factual findings. The first set supported the conclusion that Murphy's Brady claim was available before his first application was filed and so his claim was abusive. The second set supported the conclusion that Murphy's Brady claim failed on the merits. Based on these findings, the TCCA, in a brief order, dismissed Murphy's application "as an abuse of the writ without considering the merits of the claims." Murphy , 2012 WL 982945, at *1.
Before us, the sides now quarrel over whether both sets of factual findings are entitled to a presumption of correctness under 28 U.S.C. § 2254(e)(1).10 The State contends that both sets are owed deference; Murphy contends that only the first is. We side with the State.
Under 28 U.S.C. § 2254(e)(1), "a determination of a factual issue made by a State court shall be presumed to be correct" and the "applicant shall have the burden of rebutting" this presumption "by clear and convincing evidence." Unlike § 2254(d), no adjudication on the merits is needed for § 2254(e)(1) to apply. And by § 2254(e)(1) 's terms, it applies to factual determinations "made by a State court," making no distinction between trial and appellate courts. Cf. Craker v. Procunier , 756 F.2d 1212, 1213-14 (5th Cir.1985) (making a similar point about a comparable provision in § 2254 's predecessor). We thusly held in Williams v. Quarterman that a presumption of correctness attaches to a state trial court's factual findings "even if the state appellate court reached a different legal conclusion when applying the law to those facts." See 551 F.3d 352, 358 (5th Cir.2008) (first citing Sumner v. Mata , 449 U.S. 539, 546-47, 101 S.Ct. 764, 66 L.Ed.2d 722 (1981) ; then citing Craker , 756 F.2d at 1213-14 ). But there are some circumstances where a state trial court's factual findings will not "survive review." Id. We have held that trial courts' findings do not survive an appellate court's review "where they were neither adopted nor incorporated into the appellate court's peremptory denial of relief, but instead were directly inconsistent with the appellate court's decision." Id. (citing Micheaux v. Collins , 944 F.2d 231, 232 (5th Cir.1991) (en banc) (per curiam) ).
In this case, both sets of the state trial court's findings "survive[d] review." See id. Regarding the first set-which went to the procedural bar-the TCCA expressly "[b]ased" its abuse of the writ holding on the state court's factual findings. Murphy , 2012 WL 982945, at *1. And the second set-which went to the merits-also survived as it was not "directly inconsistent" with the TCCA's dismissal. See Williams , 551 F.3d at 358. Indeed, many of the factual findings on the merits bore on whether Murphy's Brady claim was previously available and thus was an abuse of the writ.
Murphy tries to distinguish Williams . He argues that the TCCA's dismissal without *596consideration of the merits was directly inconsistent with the trial court's merits findings. Plus Williams , per Murphy, states that a presumption of correctness only attaches when the state appellate court "appl[ies] the law to those facts," something the TCCA did not do for the trial court's merits findings. See id.
Murphy misreads Williams . Neither Williams nor any of our caselaw indicates that an appellate court's failure to consider an issue makes its ruling "directly inconsistent" with the trial court's relevant finding. Plus in this case, inconsistency is particularly difficult to infer; the TCCA did not state which factual findings it based its dismissal upon, and the factual findings on the merits of the Brady claim are relevant to whether the claim was an abuse of the writ. And the full passage of Williams Murphy cites states that a "trial court's factual findings are entitled to a presumption of correctness even if the state appellate court reached a different legal conclusion when applying the law to those facts." Id. (emphasis added). It does not say that the appellate court must apply law to a set of facts for those facts to survive.
Murphy also cites Jones v. Davis for support. See 890 F.3d 559, 565 (5th Cir.2018). In Jones , we reviewed a fair-trial claim without deference to the state trial court's relevant factual findings. Id. There, the TCCA had "expressly rejected all findings and conclusions made by the lower habeas court and decided the case on procedural grounds." Id. "[B]ecause the TCCA decided the case on procedural grounds, there was no 'determination of a factual issue made by a State court' to which the federal court could have deferred under § 2254(e)(1)." Id.
Jones involved an express rejection of factual findings. Id. It therefore aligns with our cases holding that direct inconsistency between a trial court's fact finding and an appellate court's ruling removes the presumption of correctness from the trial court's finding.11 Jones therefore does not avail Murphy. In this case, unlike in Jones , the TCCA did not "expressly reject[ ]" the trial court's findings on the merits. See id. Instead, it *597simply did not consider the merits of Murphy's claim.
Accordingly, all the state trial court's express factual findings are owed a presumption of correctness, a presumption Murphy may rebut only with clear and convincing evidence. See 28 U.S.C. § 2254(e)(1). Not only that, this presumption "also applies to those unarticulated findings which are necessary to the state court's conclusions of mixed law and fact." Valdez v. Cockrell , 274 F.3d 941, 948 n.11 (5th Cir.2001). This standard "is demanding but not insatiable." Miller-El v. Dretke , 545 U.S. 231, 240, 125 S.Ct. 2317, 162 L.Ed.2d 196 (2005).
Resolved that § 2254(e)(1) constrains our review of all the state trial court's factual findings, we next turn to the Brady claim itself.
B.
To prove his Brady claim, Murphy must show three things: (1) the evidence at issue is favorable to the defense, either because it is exculpatory or impeaching, (2) the prosecution suppressed the evidence, and (3) the evidence is material. See United States v. Brown , 650 F.3d 581, 587-88 (5th Cir.2011). Recall, however, that Murphy's Brady claim was dismissed as an abuse of the writ by the TCCA, Murphy , 2012 WL 982945, at *1, an independent state ground that ordinarily will bar our review, Moore v. Quarterman , 534 F.3d 454, 463 (5th Cir.2008). To overcome his default, Murphy invokes Banks v. Dretke , 540 U.S. 668, 124 S.Ct. 1256, 157 L.Ed.2d 1166 (2004). "A federal court may consider the merits of a procedurally defaulted claim if the petitioner shows 'cause for the default and prejudice from a violation of federal law.' " Canales v. Stephens , 765 F.3d 551, 562 (5th Cir.2014) (quoting Martinez v. Ryan , 566 U.S. 1, 10, 132 S.Ct. 1309, 182 L.Ed.2d 272 (2012) ). Under Banks , a petitioner shows "cause" if "the reason for his failure to develop facts in state-court proceedings was the State's suppression of the relevant evidence." 540 U.S. at 691, 124 S.Ct. 1256. To show prejudice, Murphy must demonstrate that "the suppressed evidence is 'material' for Brady purposes." See Rocha v. Thaler , 619 F.3d 387, 394 (5th Cir.2010). In effect then, to overcome his procedural default and prevail on the merits, Murphy must establish a valid Brady claim. He fails to do so. Even assuming the evidence was suppressed, we conclude it was not material.12
Suppressed evidence is material "if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different." United States v. Bagley , 473 U.S. 667, 685, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985). We consider evidence's materiality in light of other suppressed evidence, not merely the probative value of the suppressed evidence standing alone. See Kyles v. Whitley , 514 U.S. 419, 436, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995).
*598"The materiality of Brady material depends almost entirely on the value of the evidence relative to the other evidence mustered by the state." Rocha , 619 F.3d at 396 (quoting United States v. Sipe , 388 F.3d 471, 478 (5th Cir.2004) ). "If the evidence provides only incremental impeachment value, it does not rise to the level of Brady materiality." Miller v. Dretke , 431 F.3d 241, 251 (5th Cir.2005).
Murphy claims he could have used the conversation to impeach Wilhelm's two identifications: her initial photo identification and her in-courtroom identification. He argues that the jury would believe that the conversation confirmed for Wilhelm the accuracy of her prior photo identification and influenced her later in-courtroom identification. This revelation, according to Murphy, would therefore undermine a key piece of evidence of his future dangerousness.
We disagree. The pretrial conversation was of marginal value to the defense and was cumulative with already presented impeachment evidence. See Miller , 431 F.3d at 251. The pretrial conversation could not impeach Wilhelm's photo identification-an identification which occurred before the conversation. This is particularly detrimental to Murphy's case for materiality, as the detective who conducted the lineup called it "one of the better photo" identifications he ever had. So, best case for Murphy, the conversation could only impeach Wilhelm's in-courtroom identification. What is more, the conversation would not be very potent impeachment material. The conversation could be shown to have bolstered Wilhelm's confidence when giving her in-courtroom identification; it could not demonstrate that her memory was tainted or altered. As the state trial court found, the prosecutor did not do things during the pretrial interview that might taint Wilhelm's memory, like showing Wilhelm photos of Murphy, the police composite sketch, or where Murphy would be in the courtroom. Plus whatever confidence boost the pretrial conversation gave Wilhelm would be slight relative the other suggestive circumstances the defense identified at trial. The defense drew out that: Wilhelm saw Murphy's photo on TV and in the newspapers; Wilhelm's mother told Wilhelm that Murphy looked like the man in the composite sketch; and the photo lineup was unfairly constructed. "[W]hile the defense surely could have used [the pretrial conversation] in its cross-examination of [Wilhelm], it would not have significantly added to the impeachment ammunition that [Murphy]'s counsel already had." See Cobb v. Thaler , 682 F.3d 364, 380 (5th Cir.2012). Finally, as we detailed on Murphy's ineffectiveness claim, the State put on significant other evidence to show Murphy's future dangerousness besides the Wilhelm kidnapping.
The pretrial conversation was not material. Thus, the district court correctly held that Murphy's Brady claim was both procedurally defaulted and meritless.
IV.
Before we finish, we must address one more argument from Murphy. Murphy claims that the prejudice from his Strickland and Brady claims should be considered cumulatively if, considered individually, neither is sufficiently prejudicial. This argument does not avail Murphy.
As an initial matter, Murphy did not raise this argument below, and thus he may not present it on appeal. See United States v. Scroggins , 599 F.3d 433, 446-47 (5th Cir.2010) ; see also Fed. R. App. P. 28(a)(8)(A). In any event, Murphy cannot cumulate the prejudice from his claims because his Brady claim is, as we held, procedurally barred. See Derden v. McNeel , 978 F.2d 1453, 1458 (5th Cir.1992)
*599(en banc) (establishing, as a condition for showing cumulative error, that "the error complained of must not have been procedurally barred from habeas corpus review").13 Moreover, because we have rejected both his alleged errors, "there is nothing to accumulate" and the "doctrine has no applicability." See United States v. Delgado , 672 F.3d 320, 344 (5th Cir.2012) (en banc). Finally, even if we assumed he established error on his claims and amassed the prejudice from both, we would find no cumulative error. For both claims, we concluded that the prejudice arising from the alleged violation itself was slight. Neither violation interacts with the evidence of Murphy's dangerousness the state trial court found most potent-his unprovoked murder of an elderly woman.
V.
For the foregoing reasons, we AFFIRM the judgment of the district court.
Murphy also raised a related prosecutorial-misconduct claim based on use of false testimony. He no longer presses this claim.
Originally, Murphy requested a COA based on suppression of evidence Brady v. Maryland , 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), as well as use of false testimony under Giglio v. United States , 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972). We previously noted that Napue v. Illinois , 360 U.S. 264, 79 S.Ct. 1173, 3 L.Ed.2d 1217 (1959), "is a better fit for Murphy's claims," as he alleged "the use of false testimony, not merely the failure to disclose contradictory evidence." Murphy , 732 Fed.Appx. at 256 n.1. On appeal, Murphy presses neither Giglio nor Napue as a basis for relief. Any such argument is therefore forfeited. See United States v. Scroggins , 599 F.3d 433, 446-47 (5th Cir.2010) ; see also Fed. R. App. P. 28(a)(8)(A).
Murphy does not separately claim that counsel was deficient for failing to object to the State's summation. The argument is therefore forfeited. See Scroggins , 599 F.3d at 446-47 ; see also Fed. R. App. P. 28(a)(8)(A).
Murphy did not raise an ineffectiveness claim in his original state habeas application. When he raised this ineffectiveness claim in his second application, the TCCA dismissed it as an abuse of the writ-an adequate and independent state ground. See Canales v. Stephens , 765 F.3d 551, 566 (5th Cir.2014). Murphy's claim is therefore procedurally defaulted, see id. , and not subject to the strictures of § 2254(d), which requires an adjudication on the merits, see 28 U.S.C. § 2254(d). But instead of deciding if Murphy can overcome his procedural default via Martinez v. Ryan , 566 U.S. 1, 132 S.Ct. 1309, 182 L.Ed.2d 272 (2012), and Trevino v. Thaler , 569 U.S. 413, 133 S.Ct. 1911, 185 L.Ed.2d 1044 (2013), we will cut straight to the merits to deny his claim. See Bell v. Cone , 543 U.S. 447, 451 & n.3, 125 S.Ct. 847, 160 L.Ed.2d 881 (2005) (declining to consider whether the court of appeals correctly held that the petitioner had not defaulted and citing § 2254(b)(2) for the proposition that a habeas application "may be denied on the merits, notwithstanding a petitioner's failure to exhaust in state court").
This of course assumes that the availability of an evidentiary hearing is not barred by 28 U.S.C. § 2254(e)(2). As we conclude the district court did not abuse its discretion in denying an evidentiary hearing, we do not consider whether it would be barred from doing so by statute.
Murphy argues that a federal evidentiary hearing is required because the state courts did not allow him to develop the facts needed to support his ineffectiveness claim. But he cites nothing showing that a lack of record development before the state courts entitles him to a federal evidentiary hearing. See Segundo v. Davis , 831 F.3d 345, 351 (5th Cir.2016) ("[W]e decline to hold that Martinez mandates an opportunity for additional fact-finding in support of cause and prejudice.").
If the hearing revealed that counsel knew of the MMPI-2's and MCMI-3's origins, Murphy's case would be even weaker. If counsel was aware, we must presume that the omission on redirect was done "for tactical reasons rather than through sheer neglect." See Yarborough v. Gentry , 540 U.S. 1, 8, 124 S.Ct. 1, 157 L.Ed.2d 1 (2003). Indeed, we must "affirmatively entertain possible 'reasons ... counsel may have had for proceeding as they did.' " Cullen v. Pinholster , 563 U.S. 170, 196, 131 S.Ct. 1388, 179 L.Ed.2d 557 (2011) (quoting Pinholster v. Ayers , 590 F.3d 651, 692 (9th Cir.2009) (Kozinski, C.J., dissenting) ). In this case, counsel could reasonably narrow to scope of her redirect to avoid "shift[ing] attention to esoteric matters" or "distract[ing] the jury" from other, stronger points. See Richter , 562 U.S. at 108, 131 S.Ct. 770.
As Turner is unpublished, we cite it as "persuasive authority." See United States v. Torres-Jaime , 821 F.3d 577, 582 (5th Cir.2016).
Murphy also appears to argue that counsel failed him by not preparing Dr. Connell to testify. Certainly, failing to prepare a witness may constitute deficient assistance. See, e.g. , Bemore v. Chappell , 788 F.3d 1151, 1163 (9th Cir.2015) (holding that it was deficient performance when counsel's only preparation of a crucial alibi witness was meeting with him the night before the witness took the stand); Harris v. Thompson , 698 F.3d 609, 640 (7th Cir.2012) (deficient to totally fail to prepare a crucial, six-year-old witness for his testimony); Alcala v. Woodford , 334 F.3d 862, 890 (9th Cir.2003) (deficient to call a witness without even interviewing him). But Murphy cites no case even remotely analogous to the one before us. We cannot conclude that counsel was deficient because they did not fully prepare an expert to testify on the intricacies of tests the expert administered.
The parties also clash over whether the strictures of 28 U.S.C. § 2254(d) apply. We do not resolve this issue. See Berghuis v. Thompkins , 560 U.S. 370, 390, 130 S.Ct. 2250, 176 L.Ed.2d 1098 (2010) (holding that federal courts can "deny writs of habeas corpus under § 2254 by engaging in de novo review when it is unclear whether AEDPA deference applies"). Assuming § 2254(d) 's strictures do not apply, Murphy's claim still fails.
Compare Craker , 756 F.2d at 1213-14 (applying a presumption of correctness to the trial court's findings when the TCCA "did not reject the factual findings of the state trial court; it merely held that the facts as found did not entitle [the petitioner] to relief"), with Micheaux , 944 F.2d at 232 n.1 (declining to apply a presumption of correctness in favor of the trial court's findings and distinguishing Craker on the ground that there the TCCA "did not reject the factual findings of the state [trial] court" (alterations in original) ); cf. also Williams , 551 F.3d at 357-59 (holding that review was de novo when the TCCA expressly found that "some" findings lacked support because it would be "pure speculation" to determine which facts TCCA concluded were and were not supported).
Indeed, Jones cited Williams , which in turn held that "a state habeas trial court's factual findings do not survive review by the [TCCA] where they [are] neither adopted nor incorporated into the appellate court's peremptory denial of relief, but instead were directly inconsistent with the appellate court's decision."Williams , 551 F.3d at 358. In its parenthetical quoting Williams , however, Jones omitted the second part of this sentence, making the sentence read only that "a state habeas trial court's factual findings do not survive review by the [TCCA] where they [are] neither adopted nor incorporated into the appellate court's peremptory denial of relief." 890 F.3d at 565 n.24 (alterations in original). As we explained previously, the facts of Jones do not bring it into conflict with our longstanding precedent. And even if we assume that Jones 's half-complete quotation was intended to change the law, one panel "may not overturn another panel's decision, absent an intervening change in the law, such as by a statutory amendment, or the Supreme Court, or our en banc court." Jacobs v. Nat'l Drug Intelligence Ctr. , 548 F.3d 375, 378 (5th Cir.2008). We are therefore bound by our prior statement of the law in Williams .
The parties quarrel over whether the pretrial conversation could impeach Wilhelm's photo and in-courtroom identification. Murphy argues that exposing the conversation could explain why Wilhelm was confident about her photo identification. And exposing the conversation, according to Murphy, could show that her in-courtroom identification was tainted. The State counters, arguing that the state trial court's factual findings indicate that the conversation did not confirm her photo identification or taint her in-courtroom identification. The State adds that the conversation lacked impeachment value as the jury would naturally suspect that a prosecutor spoke with Wilhelm beforehand and that Wilhelm knew she was called because she had previously identified Murphy. We assume without deciding that the evidence was impeaching, and nevertheless conclude that the evidence was not material.
Murphy does not argue that we may cumulate errors to satisfy the prejudice prong to excuse procedural default, and thus we do not consider any such argument here.