United States v. David Runyon

AMENDED PUBLISHED OPINION UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT

No. 17-5

UNITED STATES OF AMERICA,

Plaintiff - Appellee, v. DAVID ANTHONY RUNYON, Defendant - Appellant.

Appeal from the United States District Court for the Eastern District of Virginia at Newport News. Rebecca Beach Smith, Senior District Judge. (4:08-cr-00016-RBS-DEM-3)

Argued: September 10, 2020 Decided: December 23, 2020 Amended: February 12, 2021

Before GREGORY, Chief Judge, and WILKINSON and NIEMEYER, Circuit Judges.

Affirmed in part, vacated in part, and remanded with instructions by published opinion. Judge Niemeyer wrote the opinion, in which Chief Judge Gregory joined except as to Parts II, IV, and V and Judge Wilkinson joined except as to Part III. Chief Judge Gregory wrote a separate opinion concurring in part and dissenting in part. Judge Wilkinson wrote a separate opinion concurring in part and dissenting in part.

ARGUED: Helen Susanne Bales, FEDERAL DEFENDER SERVICES OF EASTERN TENNESSEE, Knoxville, Tennessee, for Appellant. Brian James Samuels, OFFICE OF THE UNITED STATES ATTORNEY, Newport News, Virginia, for Appellee. ON BRIEF: Michele J. Brace, VA CAPITAL REPRESENTATION RESOURCE CENTER, Charlottesville, Virginia; Dana C. Hanson Chavis, FEDERAL DEFENDER SERVICES OF EASTERN TENNESSEE, Knoxville, Tennessee, for Appellant. G. Zachary Terwilliger, United States Attorney, Alexandria, Virginia, Lisa R. McKeel, OFFICE OF THE UNITED STATES ATTORNEY, Newport News, Virginia, for Appellee.

NIEMEYER, Circuit Judge, with whom Chief Judge GREGORY joined except as to Parts II, IV, and V and Judge WILKINSON joined except as to Part III:

David Runyon shot and killed Cory Allen Voss in late April 2007 in Newport News, Virginia, pursuant to a murder-for-hire conspiracy that he entered into with Voss’s wife, Catherina Voss, and her paramour, Michael Draven. A jury convicted Runyon of conspiracy to commit murder for hire, in violation of 18 U.S.C. § 1958(a); carjacking resulting in death, in violation of 18 U.S.C. § 2119; and murder with the use of a firearm in relation to a crime of violence, in violation of 18 U.S.C. § 924(c)(1), (j)(1), and recommended that Runyon be sentenced to death. The district court accordingly entered judgment on December 4, 2009, sentencing Runyon to death. On appeal we affirmed. United States v. Runyon, 707 F.3d 475 (4th Cir. 2013), cert. denied, 135 S. Ct. 46 (2014).

Runyon has now filed this motion under 28 U.S.C. § 2255 to vacate or correct his sentence, asserting 18 grounds for relief. The district court denied his motion by order dated January 19, 2017, and denied a certificate of appealability. By order dated August 14, 2019, we granted a certificate of appealability as to four issues: (1) whether Runyon’s § 924 conviction is invalid because the offense was not committed during and in relation to a “crime of violence”; (2) whether trial counsel provided ineffective assistance by failing to investigate and present mitigating evidence of Runyon’s brain injury and potential mental illness; (3) whether the government violated Brady v. Maryland, 373 U.S. 83 (1963), in failing to disclose the codefendant’s history of sexual assault or whether, in the alternative, trial counsel’s failure to investigate that history and present it to the jury constituted ineffective assistance of counsel; and (4) whether the government exercised its *3 peremptory jury strikes in a discriminatory manner, in violation of Batson v. Kentucky, 476 U.S. 79 (1986), or whether counsel unreasonably failed to challenge the government’s strikes at trial or on direct review.

For the reasons that follow, we vacate the district court’s ruling dismissing Runyon’s claim that his counsel was constitutionally ineffective in failing to investigate mitigating evidence of brain injury and potential mental illness and remand that claim for an evidentiary hearing. Otherwise, we affirm.

I

The murder in this case was highly planned. Briefly, the facts, which are set out in more detail in our earlier opinion, 707 F.3d at 484–86, show that Catherina Voss (“Catherina”), the wife of Cory Voss (“Voss”), a U.S. Navy officer, had been engaged in an extramarital affair with Michael Draven. Catherina and Draven decided to murder Voss in the hope of gaining Voss’s Navy death benefits and life-insurance proceeds. To carry out the murder, Draven hired David Runyon, whom Draven had met as a co-participant in a drug-research study.

Shortly before the crime, Catherina opened an account at a branch of a local bank in Newport News with a five-dollar deposit. Thereafter, on the night of the murder, Catherina sent Voss to the bank’s ATM to withdraw cash. Video surveillance of the scene showed that while Voss stood at the ATM, an unidentified man — later found to be Runyon — entered Voss’s pickup truck. Voss then drove away from the ATM but returned a few minutes later and attempted another withdrawal, which was denied due to insufficient *4 funds. The next morning, Voss was found dead in his truck in a parking lot near the bank, having been shot five times at close range. Compelling evidence connected the bullets used in the murder to Runyon.

Runyon, Catherina, and Draven were ultimately arrested and charged for the murder of Voss and related offenses. Catherina pleaded guilty to all counts and was sentenced to life imprisonment. Runyon and Draven proceeded to trial, with the government seeking the death penalty against Runyon. The jury returned a verdict, finding both Runyon and Draven guilty of conspiracy to commit murder for hire, carjacking, and murder with the use of a firearm in relation to a crime of violence. Draven was sentenced to life imprisonment, while the trial continued against Runyon pursuant to the Federal Death Penalty Act, 18 U.S.C. §§ 3591–98.

In proceedings under the Death Penalty Act, the jury next found Runyon eligible for the death penalty after finding that he intentionally killed Voss and finding two statutory aggravating factors — that Runyon had committed the crime for pecuniary gain and that he committed the crime after substantial planning.

Before the next phase of trial, in which the jury was required to select the penalty, the government gave notice of four non-statutory aggravating factors for the jury to consider — in addition to the statutory factors that the jury had already found. The non-statutory aggravating factors were a lack of remorse; injury and loss to Voss and his family and friends; a history of physical abuse toward women; and use of law enforcement and military training to perpetrate the murder. The military-training aggravator was based in part on Runyon’s service as an officer in the Kansas National Guard and as an enlisted *5 member of the United States Army. The jury unanimously found each of the government’s proposed aggravating factors. It also unanimously found that Runyon had established 7 of his proposed 14 mitigators, including the mitigator that “[o]ther persons equally culpable in the crime will not be punished by death.” In addition, the jury unanimously found two non-statutory mitigators that Runyon had not proposed — that Runyon experienced domestic violence as a child and that his brother would suffer emotional harm if Runyon were executed. Ten or eleven jurors found three additional proposed mitigators, and eleven jurors agreed that Runyon had established a mitigator that he had not proposed — that Runyon was given the impression that Voss was molesting his own daughter. After making its findings on the aggravating and mitigating factors, the jury unanimously recommended the death sentence on two counts — conspiracy to commit murder for hire and murder in connection with the use of a firearm in relation to a crime of violence — and it recommended life imprisonment on the carjacking count. The district court imposed the recommended sentences, entering judgment on December 4, 2009.

In his motion under § 2255 seeking collateral review, Runyon advanced 18 claims. He sought discovery for several of the claims, as well as an evidentiary hearing. In a thorough 246-page opinion and order, the district court denied Runyon’s request for discovery and an evidentiary hearing and dismissed the § 2255 motion. It also denied a certificate of appealability. Runyon v. United States, 228 F. Supp. 3d 569 (E.D. Va. 2017).

By order dated August 14, 2019, we granted a certificate of appealability on the four issues now before us. See 28 U.S.C. § 2253.

II

With respect to the first issue certified for appeal, Runyon contends that his conviction for violating 18 U.S.C. § 924(c)(1), (j)(1) is invalid because the predicate crimes relied on for conviction — conspiracy to commit murder for hire under § 1958(a) and carjacking under § 2119 — do not qualify as “crime[s] of violence,” as defined by § 924(c)(3). The relevant portions of § 924 provide that “any person, who, during and in relation to any crime of violence” “causes the death of a person through the use of a firearm, shall — if the killing is a murder . . . be punished by death or by imprisonment.” 18 U.S.C. § 924(c)(1), (j)(1). And “crime of violence” is defined in § 924(c)(3)’s “force clause” * as any felony that “has as an element the use, attempted use, or threatened use of physical force against the person or property of another.” Id. § 924(c)(3)(A).

Runyon argues that neither conspiracy to commit murder for hire nor carjacking is a crime of violence because neither crime necessarily requires for conviction the “use of physical force.” With respect to the conspiracy predicate, he argues that the crime “requires merely an agreement to act, which cannot qualify as physical force.” And insofar as the crime might involve physical force, he contends that it lacks the requisite mens rea inherent in the “use” of physical force. Similarly, with respect to carjacking, he argues that the

* The definition of “crime of violence” also includes a “residual clause,” which defines “crime of violence” as any felony that “by its nature, involves a substantial risk that physical force against the person or property of another may be used in the course of committing the offense.” 18 U.S.C. § 924(c)(3)(B). Because the Supreme Court recently held that § 924(c)(3)(B)’s residual clause is unconstitutionally vague, see United States v. Davis, 139 S. Ct. 2319, 2323–24 (2019), the government no longer relies on that clause to argue that conspiracy to commit murder for hire and carjacking are crimes of violence.

crime “can be accomplished without strong physical force. It may be accomplished with intimidation.” Moreover, he asserts that, because the jury did not indicate which predicate it relied on to return its conviction on § 924(c)(1), (j)(1), the government must show that both predicates constitute crimes of violence.

The government does indeed contend that both conspiracy to commit murder for hire and carjacking are crimes of violence. While it acknowledges that conspiracy generally does not serve as a valid predicate under the force clause — see, e.g., United States v. Simms, 914 F.3d 229, 234 (4th Cir. 2019) (en banc) (concluding that conspiracy to commit Hobbs Act robbery is not a force-clause crime under § 924(c)(3)(A)); United States v. McCollum, 885 F.3d 300, 309 (4th Cir. 2018) (holding that conspiracy to commit murder in aid of racketeering is not categorically a crime of violence) — it argues that the elements of the conspiracy offense here are different in that “it is not possible for a conspiracy with the object of committing murder for hire to result in death without the use or threatened use of force.” And with respect to carjacking, the government notes that this court has already concluded that carjacking in violation of 18 U.S.C. § 2119 is a crime of violence. See United States v. Evans, 848 F.3d 242, 247–48 (4th Cir. 2017).

Accordingly, we are presented with the two distinct questions of whether conspiracy to commit murder for hire under § 1958(a) and carjacking under § 2119 are “crimes of violence,” as defined by 18 U.S.C. § 924(c)(3).

Because § 924(c)(3) requires us to focus on the elements of the offense — defining “crime of violence” as a felony that has “as an element” the use of force (emphasis added) — we apply the categorical approach. See Davis, 139 S. Ct. at 2327–28. Under the *8 categorical approach, we consider only the statutory definition of the offense by its elements and the fact of conviction, without considering the actual facts supporting conviction. See Taylor v. United States, 495 U.S. 575, 602 (1990); United States v. Bell, 901 F.3d 455, 468–69 (4th Cir. 2018); United States v. McNeal, 818 F.3d 141, 152 (4th Cir. 2016). And when looking at the elements of the offense, we must determine whether “there is a realistic probability — not merely a theoretical possibility — that the minimum conduct necessary for conviction . . . involves the use of physical force as defined by federal law.” United States v. Rumley, 952 F.3d 538, 548 (4th Cir. 2020) (emphasis added); see also United States v. Allred, 942 F.3d 641, 648 (4th Cir. 2019) (similar).

Federal law defines physical force to mean “violent force — that is, force capable of causing physical pain or injury to another person.” Curtis Johnson v. United States, 559 U.S. 133, 140 (2010); see also Stokeling v. United States, 139 S. Ct. 544, 553 (2019) (same); Allred, 942 F.3d at 652 (same). And that, of course, includes causing death to another person. See, e.g., United States v. Irby, 858 F.3d 231, 236 (4th Cir. 2017) (observing that “it is hard to imagine conduct that can cause another to die that does not involve physical force against the body of the person killed” (cleaned up)). But not every act that causes bodily injury or death amounts to the use of physical force as required by § 924(c)(3)’s force clause. That is because the term “use” targets action, implying a deliberate or perhaps reckless mens rea, and bodily injury or death “can result from negligent or even accidental acts.” Rumley, 952 F.3d at 549. “[T]hose acts, even if criminal, would not constitute” crimes of violence, as they do not involve a “use” of physical force. Id. (emphasis added). Thus the phrase “use of physical force” in the force *9 clause requires “a higher degree of intent than negligent or merely accidental conduct.” Leocal v. Ashcroft, 543 U.S. 1, 9 (2004); see also Allred, 942 F.3d at 652 (“[A]n offense will not have as an element the ‘use’ of force sufficient to qualify as a violent felony if it does not have the requisite level of mens rea”). By contrast, “the knowing or intentional causation of bodily injury necessarily involves the use of physical force.” United States v. Castleman, 572 U.S. 157, 169 (2014) (emphasis added); see also United States v. Battle, 927 F.3d 160, 166 (4th Cir. 2019).

Finally, when an offense includes alternative elements for conviction, it becomes divisible, and courts may then use a “modified categorical approach” to determine “which element played a part in the defendant’s conviction.” Descamps v. United States, 570 U.S. 254, 260 (2013). Under this approach, the court may look to the terms of the relevant charging document, jury instructions, plea agreement, plea colloquy, and the like. See Mathis v. United States, 136 S. Ct. 2243, 2249 (2016); Shepard v. United States, 544 U.S. 13, 26 (2005); Allred, 942 F.3d at 648 (noting that under the modified categorical approach, we “consult a limited set of record documents . . . for the sole purpose of determining ‘what crime, with what elements, a defendant was convicted of’” (quoting Mathis, 136 S. Ct. at 2249)). “[O]nce the court has [under the modified categorical approach] consulted the record and isolated the specific crime underlying the defendant’s conviction, it must then apply the categorical approach to determine if it constitutes a [crime of violence],” considering only the elements of the identified crime and the fact of conviction. Allred, 942 F.3d at 648.

In this case, the jury was not asked to indicate in its verdict form whether it was relying on conspiracy to commit murder for hire or carjacking in finding Runyon guilty under § 924(c)(1), (j)(1). Accordingly, we must assume that Runyon could have been convicted by the jury’s reliance on either predicate offense, requiring us to determine whether each predicate offense qualifies as a crime of violence. See Curtis Johnson, 559 U.S. at 137; United States v. Vann, 660 F.3d 771, 774–75 (4th Cir. 2011) (en banc) (per curiam).

We consider first, as our discussion need only be brief, whether carjacking under § 2119 is a crime of violence under § 924(c)(3)’s force clause. We recently held that a conviction under § 2119 is categorically a conviction for a crime of violence, and that holding controls here. See Evans, 848 F.3d at 245. And while Runyon invites us to overrule Evans, in this circuit it is established that “one panel cannot overrule another.” McMellon v. United States, 387 F.3d 329, 333 (4th Cir. 2004). Thus, Runyon’s carjacking offense qualifies as a crime of violence that can support a conviction under § 924(c)(1), (j)(1).

Whether conspiracy to commit murder for hire in violation of § 1958(a) is a crime of violence merits a fuller discussion. This inquiry requires us to consider the elements of the offense and whether a conviction under those elements necessarily requires the “use” of physical force within the meaning of the force clause.

Section 1958(a) provides that:

Whoever travels in or causes another (including the intended victim) to travel in interstate or foreign commerce, or uses or causes another (including the

intended victim) to use the mail or any facility of interstate or foreign commerce,

with intent that a murder be committed in violation of the laws of any State or the United States as consideration for the receipt of, or as consideration for a promise or agreement to pay, anything of pecuniary value,

or who conspires to do so,

shall be fined under this title or imprisoned for not more than ten years, or both;

and if personal injury results, shall be fined under this title or imprisoned for not more than twenty years, or both;

and if death results, shall be punished by death or life imprisonment, or shall be fined not more than $250,000, or both.

18 U.S.C. § 1958(a) (emphasis added) (spaces between clauses added).

Because § 1958(a) imposes distinct enhanced penalties in circumstances where “personal injury results” or where “death results,” those are alternative elements for conviction that must be proven to the jury beyond a reasonable doubt under Apprendi v. New Jersey, 530 U.S. 466 (2000). See Mathis, 136 S. Ct. at 2256 (“If statutory alternatives carry different punishments, then under Apprendi they must be elements”); see also Burrage v. United States, 571 U.S. 204, 210 (2014) (“Because the ‘death results’ enhancement increased the minimum and maximum sentences to which [the defendant] was exposed, it is an element that must be submitted to the jury and found beyond a reasonable doubt”). Similarly, the “conspiracy” clause requires a jury to find the alternative additional element that the defendant entered into an agreement that the underlying offense be committed. See Ocasio v. United States, 136 S. Ct. 1423, 1429 (2016); cf. Simms, 914 F.3d at 233–34 (treating conspiracy to commit Hobbs Act robbery *12 as a distinct offense from Hobbs Act robbery and holding that it requires the government to “prove . . . that the defendant agreed with another to commit actions that, if realized, would violate the Hobbs Act”). As a consequence, the multiple alternative elements of § 1958(a) define six distinct possible crimes: (1) using facilities of commerce with the intent that a murder be committed for hire; (2) conspiracy to use facilities of commerce with the intent that a murder be committed for hire; (3) using facilities of commerce with the intent that a murder be committed for hire where personal injury results; (4) conspiracy to use facilities of commerce with the intent that a murder be committed for hire where personal injury results; (5) using facilities of commerce with the intent that a murder be committed for hire where death results; and (6) conspiracy to use facilities of commerce with the intent that a murder be committed for hire where death results. In these circumstances, the modified categorical approach is necessary to determine the crime for which Runyon was convicted and which was identified as a crime of violence for his conviction under § 924(c)(1), (j)(1).

Thus, as allowed by the modified categorical approach, we review the indictment on which Runyon was convicted and the jury instructions leading up to the conviction to determine the actual crime for which Runyon was convicted. Count V of the indictment charged that Runyon “did knowingly carry and use a firearm during and in relation to a crime of violence [referring, among other things, to the crime charged in Count I] in violation of Title 18, United States Code, Section 924(c)(1), and in the course of this violation caused the death of a person through the use of a firearm, which killing was a murder . . . in that the defendants, with malice aforethought, did unlawfully kill Cory Allen *13 Voss by shooting him with a firearm,” in violation of § 924(j). And Count I charged that Runyon “did unlawfully, knowingly and intentionally conspire . . . to travel in and cause another to travel in interstate commerce . . . with intent that a murder be committed . . . as consideration for the receipt of, and as consideration for a promise and agreement to pay, something of pecuniary value, resulting in the death of Cory Allen Voss,” in violation of § 1958(a). The jury instructions likewise stated that finding guilt on Count I required the government to prove that Runyon engaged in a conspiracy to commit murder for hire resulting in Voss’s death. The jury found Runyon guilty on both Counts I and V. Thus, in finding Runyon guilty of Count I, the jury necessarily found Runyon guilty of the offense of conspiracy to use facilities of commerce with the intent that a murder be committed for hire where death results, in violation of § 1958(a). This conclusion still leaves us with the question whether that particular crime categorically qualifies as a crime of violence under § 924(c)(3)’s force clause.

We conclude that it does. While conspiracy alone does not necessarily implicate the use of force, see, e.g., Simms, 914 F.3d at 234, conspiracy in the context of the § 1958 offense at issue is different because it has heightened mens rea elements, as well as the element that “death results.” As already noted, an act that results in death obviously requires “physical force.” See Irby, 858 F.3d at 236. And the death resulting from a conspiracy to commit murder for hire has the “requisite mens rea” to constitute a use of physical force. Battle, 927 F.3d at 166. The conspiracy here has two heightened mens rea elements: (1) the intent to join the conspiracy, see Ocasio, 136 S. Ct. at 1429, and (2) the specific intent that a murder be committed for hire, 18 U.S.C. § 1958(a). While these mens *14 rea elements are not explicitly tied to the resulting-in-death element, in any realistic case, they must nonetheless carry forward to the resulting-in-death element. There is no “realistic probability” of the government prosecuting a defendant for entering into a conspiracy with the specific intent that a murder be committed for hire and for a death resulting from that conspiracy while that death was somehow only accidentally or negligently caused. Allred, 942 F.3d at 648. This means that a conspiracy to commit murder for hire where death results necessarily involves the “use of physical force.”

Runyon nonetheless argues that the death-results strain of § 1958(a) is not a crime of violence because, in United States v. Torres-Miguel, 701 F.3d 165 (4th Cir. 2012), we stated that “a crime may result in death or serious injury without involving the use of physical force.” Id. at 168. But “[t]his part of Torres-Miguel dealt with the requirement that a crime include a heightened mens rea in order to involve the ‘use’ of physical force.” Allred, 942 F.3d at 653. Or, as we put it elsewhere, this “proposition applies only where a crime does not have as an element the intentional causation of death or injury.” Battle, 927 F.3d at 166. That is why in United States v. Middleton, 883 F.3d 485 (4th Cir. 2018), we held that South Carolina involuntary manslaughter does not necessarily involve the use of force. See Allred, 942 F.3d at 653–54. “But a crime requiring the ‘intentional causation’ of injury requires the use of physical force.” Battle, 927 F.3d at 166. And that is what we have here.

Runyon also posits a hypothetical where the target of a § 1958 murder-for-hire conspiracy died from an accidental or negligent car crash while riding in a conspirator’s car and argues from this that the crime can be committed without the use of violent force. *15 While this hypothetical might be in the realm of “theoretical possibility,” there is no “realistic probability” that the government would indict the conspirator for the death-results strain of conspiracy to commit murder for hire in such a situation. See Allred, 942 F.3d at 648. Indeed, this crime satisfies the force clause just as the crime in Allred did.

In Allred, we considered whether a federal statute that prohibits “knowingly engag[ing] in any conduct and thereby caus[ing] bodily injury to another person . . . with intent to retaliate against any person for” serving as a witness satisfied the force clause. 18 U.S.C. § 1513(b)(1) (emphasis added). We explained that “[a]lthough there is no mens rea specified for the element of causation, the statute contains not one, but two heightened mens rea requirements.” Allred, 942 F.3d at 654. Specifically, the defendant must have “‘knowingly engage[d]’ in conduct with the specific ‘intent to retaliate against’ a witness.” Id. (quoting § 1513(b)). We found “it difficult to imagine a realistic scenario in which a defendant would knowingly engage in conduct with the specific intent to retaliate against a witness and thereby only recklessly or negligently cause bodily injury.” Id.

Such is the case here. Section 1958(a)’s mens rea elements cannot be limited to their individual clauses. If a defendant willingly agrees to enter into a conspiracy with the specific intent that a murder be committed for money and death results from that agreement, it follows that the defendant acted with specific intent to bring about the death of the conspiracy’s victim. And this specific intent ensures that the victim’s death was necessarily the result of a use of physical force and not merely from negligence or accident. Thus, we conclude that conspiracy to commit murder for hire where death results, in violation of § 1958(a), is a crime of violence under § 924(c)(3)’s force clause, and *16 accordingly we reject Runyon’s argument that his conviction under § 924(c)(1), (j)(1) is invalid.

III

On the second issue certified for appeal, Runyon contends that his counsel failed to provide him with effective assistance, in violation of the Sixth Amendment, by failing to investigate adequately his brain injury and potential mental illness and introduce such evidence in mitigation during the penalty phase of trial. See Wiggins v. Smith, 539 U.S. 510, 521 (2003) (requiring a petitioner to establish that his counsel’s performance was “deficient” and that the deficiency “prejudiced the defense” (citing Strickland v. Washington, 466 U.S. 668, 687 (1984))); Williams v. Stirling, 914 F.3d 302 (4th Cir. 2019) (applying this standard to a failure to investigate mitigating evidence). Runyon claims that his counsel was alerted to the evidence before trial but never followed through and that the development and presentation of the evidence would likely have swayed at least one juror from voting for death to voting for life imprisonment.

The government argues that the evidence was weak and, even if the jury credited it, “was double-edged,” as it “could have strengthened government arguments about Runyon’s dangerousness.” It adds that such evidence would also have been “in tension with [Runyon’s] claims of innocence.” It concludes, therefore, that Runyon’s counsel “made reasonable strategy calls at the penalty phase” in not presenting the evidence and that such strategic calls cannot be in violation of counsel’s duty. See Strickland, 466 U.S. *17 at 690 (noting that “strategic choices made after thorough investigation . . . are virtually unchallengeable”).

Roughly six months before the penalty phase of trial began, Stephen Hudgins was appointed to represent Runyon as lead counsel for that phase after one of Runyon’s original lawyers, Jon Babineau, developed a conflict of interest. When Hudgins took over, Babineau had already filed a motion for neuropsychological expert services after a clinical psychologist, who had examined Runyon, “strongly advised that Runyon be evaluated by a neuropsychologist for the presence of neuropsychological deficits that may bear on mitigation.” Around this time, Runyon was also examined by two government mental-health experts, both of whom reported that Runyon had sustained a series of head traumas. He had apparently been knocked out during military training after being too close to an exploding grenade, and he had suffered injuries in two serious car accidents. These doctors, however, concluded that Runyon did not “suffer from any serious mental illness, mental disorder, or brain pathology and that there [were] no mental health factors that are mitigating or aggravating to whatever sentence, if any, is determined by the court.” But both doctors did suggest that Runyon was narcissistic and demonstrated some evidence of “personality dysfunction.”

Nonetheless, Hudgins sought and obtained the appointment of his own experts. He obtained the services of neuropsychologist Dr. Allen Mirsky, who examined Runyon and submitted a preliminary report to Hudgins. Dr. Mirsky explained that Runyon’s history of head injuries was potentially relevant to his mental status. He observed that Runyon’s low scores on certain tests were “consistent with some mild, diffuse brain damage” and *18 “entirely consistent with brainstem injury, which could have resulted from” either the car accidents or the blast injury. He explained that it had become clear from a study “of wounded soldiers in Iraq and Afghanistan that blast injury can have profound effects on neurocognitive functions.” He concluded that the symptoms that he observed “merit[ed] further neurological investigation.”

In addition to this preliminary report, Dr. Mirsky also wrote Hudgins separately about his examination, explaining that while his review of the information was “not yet complete,” “it is clear from the data that there is strong evidence that [Runyon] is suffering from a neurological disorder.” (Emphasis added). “It would be essential for Mr. Runyon to be evaluated by a neurologist, and have the necessary tests to establish the nature of this disorder.” (Emphasis added).

Hudgins also obtained the services of Dr. James Merikangas, a neuropsychiatrist. After conducting a neurological examination of Runyon, Dr. Merikangas ordered brain scans. Pending receipt of those scans, however, he submitted a preliminary report to Hudgins, concluding that Runyon “presently is either in a fantasy world of grandiose wishful thinking, or suffering from delusions. He clearly has impaired executive functioning suggestive of frontal lobe brain impairment.” (Emphasis added). Dr. Merikangas also observed that the evaluations done by the government experts suggested that Runyon suffered from post-traumatic stress disorder. Dr. Merikangas promised to follow up once he received the brain scans.

The brain scans that Dr. Merikangas ordered were in fact provided to Hudgins, along with a radiologist’s report that the scans were “normal.” But Hudgins recognized, as he *19 later stated, that readings by a neuropsychiatrist, rather than a radiologist, were necessary. Yet, he never provided the scans to Dr. Merikangas, as Dr. Merikangas had requested, nor to Dr. Mirsky.

At the penalty phase of trial, neither Dr. Mirsky nor Dr. Merikangas testified, nor was the information that they provided to Hudgins presented to the jury as part of the mitigation evidence. Hudgins did, however, propose 14 other mitigation factors and offer testimony from nearly two dozen witnesses.

During closing argument to the jury, the government argued that Runyon had chosen to forsake a stable upbringing in favor of an aimless life and had ultimately chosen, of his own free will, to commit these crimes. Hudgins’s cocounsel stated during closing argument that he did not have a “glib answer” or a “pat response” as to “what caused Mr. Runyon to get involved in this and to be where we are today.” He instead focused on the inequity that Runyon’s equally culpable codefendants received only life sentences for the crime for which Runyon was facing the death penalty.

In his § 2255 motion, Runyon claimed that his counsel was deficient in failing to investigate his brain injury and potential mental illness more fully and in failing to present the evidence to the jury in mitigation. To support his motion, he presented evidence from four experts about what could have been uncovered had counsel conducted a reasonable investigation. He also presented evidence from his trial counsel in an effort to explain why the evidence had not been presented to the jury.

Dr. Merikangas, who had requested but never received Runyon’s brain scans, conducted a review of those scans in 2015. He reported that “[t]hey revealed multiple *20 white matter hyperintensities . . . consistent with [Runyon’s] history of head injuries and migraine,” which, in Dr. Merikangas’s opinion, “resulted in impaired executive functioning and decision making.” He also reported that Runyon exhibited signs of paranoia and delusions, which could be “a consequence of his brain injuries, mood disorder, and [post-traumatic stress disorder].” Dr. Merikangas stated that had he testified at trial, he “would have testified that in [his] expert opinion [Runyon] was a brain damaged individual with migrainous headaches and a disorder of executive functioning with symptoms suggestive of a psychotic thought process.”

Dr. Mirsky also reevaluated Runyon in 2015, concluding that test results “indicate the presence of significant damage to the right side of the brain, as well as a psychotic disorder.” Dr. Mirsky explained that Runyon’s history of brain injuries is “consistent” with a diagnosis of post-traumatic stress disorder.

Dr. Mark Cunningham, a psychologist, who had originally testified at the penalty phase only to Runyon’s lack of future dangerousness, also reevaluated Runyon’s records in 2015. He concluded that “Runyon suffered from a myriad of malignant formative influences that he did not choose,” meaning that “there are a number of adverse developmental and life trajectory factors . . . that singly and collectively increased the likelihood of his having adverse and criminally violent outcomes in adulthood.” The developmental factors to which he referred included Runyon’s head injuries and potential frontal-lobe damage. Dr. Cunningham explained that the literature suggested that persons with the symptoms manifested by Runyon are “impulsive” and that “once they become fixated on” a course of action, “they seem[] to forget that any other [action] [i]s possible.” *21 The literature also suggested that “frontal lobe dysfunction may result in disruption of the evaluation of personal behavior, reduced problem solving and flexibility, and marked difficulty in reprogramming an ongoing chain of behavior.”

Finally, Dr. Richard Dudley, a psychiatrist, evaluated Runyon in 2015. He stated that Runyon exhibited “considerable grandiosity and paranoia”; that “the executive functions of his brain appeared to be impaired”; and that Runyon’s insight and judgment were “poor.” He concluded that Runyon’s “cognitive deficits . . . impaired [his] decision-making capacity beyond the impairments resulting from the distorted perceptions of reality, the impairments in judgment, and the impulsivity associated with his other psychiatric disorders.”

In addition to evidence from these four experts, Runyon also presented the district court with Army records describing his treatment after the car accidents and referring multiple times to post-traumatic stress disorder. He also provided affidavits from his brother and ex-wife, both of whom stated that Runyon’s personality changed after one of the car accidents.

Finally, Runyon presented two affidavits of his lead penalty-phase counsel Hudgins for consideration by the district court. As stated in the affidavits, Hudgins remembered that in preparation for the penalty phase, he was looking for an injury in Runyon’s past that might have affected his reasoning ability but had trouble locating any medical records related to either the blast injury or the car accidents. He also remembered that Dr. Merikangas requested brain scans of Runyon, but he did not remember the results of those scans. When told that in 2009 a radiologist had reported that Runyon’s brain scans were *22 “normal,” Hudgins stated that the report “would not have supplied me with enough information to make a decision to cease investigation of David Runyon’s mental health or social history.” He stated, “I understand the difference between a radiologist’s reading and a neuropsychiatrist’s/neurologist’s reading of such scans.” When told that Dr. Merikangas in fact reviewed the 2009 scans in 2015 and identified brain damage, Hudgins responded, “This is the type of information I would have been looking for at the penalty phase.” He said the same thing about the Army records that had been uncovered. Most significantly, he stated that he could not “remember why Dr. Merikangas and Dr. Mirsky did not testify in the penalty phase.”

In dismissing Runyon’s § 2255 motion, the district court concluded that trial counsel had “thoroughly investigated the Petitioner’s mental health and found that there was not enough to present a viable mitigation argument.” “Accordingly, counsel was not deficient for deciding to forego certain mental health evidence and instead present a mitigating strategy that focused on codefendant culpability, positive prisoner evidence, and family sympathy.” The court also concluded that there was no reasonable probability that a more fulsome mental-health investigation would have changed the outcome of the penalty phase because “mental health evidence is a two-edged sword” and could have undermined the mitigation case Runyon ultimately presented.

Looking in hindsight to what decisions trial counsel made and why is, to be sure, a perilous activity, readily subject to historical revisionism. Moreover, in fulfilling his duty to investigate, it must be recognized that counsel need not have investigated “every conceivable line of mitigating evidence.” Williams, 914 F.3d at 313. But counsel do have *23 “a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” Id. (quoting Strickland, 466 U.S. at 691). Of course, “strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable; and strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.” Strickland, 466 U.S. at 690–91. In short, our focus is on “whether the investigation supporting counsel’s decision not to introduce” particular mitigating evidence “was itself reasonable.” Wiggins, 539 U.S. at 523.

In conducting this inquiry, we “consider not only the quantum of evidence already known to counsel, but also whether the known evidence would lead a reasonable attorney to investigate further.” Wiggins, 539 U.S. at 527. As for the prevailing professional norms in this regard, we look to the American Bar Association Guidelines for the Appointment and Performance of Defense Counsel in Death Penalty Cases. See Strickland, 466 U.S. at 688. They indicate “[a] ‘well-defined norm’ . . . ‘that investigations into mitigating evidence should comprise efforts to discover all reasonably available mitigating evidence.’” Williams, 914 F.3d at 313 (quoting Wiggins, 539 U.S. at 524).

In this case, the evidence that trial counsel Hudgins had in hand, as well as trial counsel’s later reflections on why further investigation was not conducted, indicates that the district court should have made a further inquiry into whether, under the applicable standards, Runyon received effective assistance of counsel. Before trial, attorney Hudgins had a preliminary report from Dr. Mirsky flagging that Runyon appeared to suffer from *24 “brainstem injury” and that his “symptoms merit[ed] further neurological investigation.” Hudgins had a follow-up email from Dr. Mirsky flagging that Runyon was likely “suffering from a neurological disorder” and that “[i]t would be essential for Mr. Runyon to be evaluated by a neurologist.” Hudgins also had in hand a preliminary report from Dr. Merikangas flagging that Runyon “suffer[ed] from delusions” and “clearly ha[d] impaired executive functions suggestive of frontal lobe brain damage.” These red flags clearly pointed to potential mitigating evidence. See Abdul-Kabir v. Quarterman, 550 U.S. 233, 261–63 (2007) (explaining that “possible neurological damage” is mitigating evidence); Rompilla v. Beard, 545 U.S. 374, 392 (2005) (noting that “organic brain damage” can be mitigating evidence). Indeed, these circumstances are much like those that we recently considered in Williams, where “evidence of [fetal alcohol syndrome] was reasonably available, but counsel failed to connect the indicators suggesting further investigation.” 914 F.3d at 315. We concluded that because evidence of fetal alcohol syndrome could be a significant mitigating factor, “reasonable counsel should have at least explored” the evidence. Id.

To be sure, placing reasonable limits on further investigation might have been sound had counsel made a strategic choice in favor of pursuing other arguments that might reasonably have been undermined by evidence of brain damage and mental illness, but the record remains unclear whether such a strategic choice was made. Hudgins averred later that he did not remember why he did not use such evidence, and he added that it was indeed the type of evidence he had been looking for.

In its present state, the record does not sufficiently support the government’s argument that Hudgins made a strategic choice not to pursue the potentially mitigating evidence after a reasonable investigation. Given that the relevant evidence is murky, we conclude that an evidentiary hearing should be conducted to resolve the issue. See 28 U.S.C. § 2255(b) (requiring an evidentiary hearing on a § 2255 motion “[u]nless the motion and the files and the records of the case conclusively show that the prisoner is entitled to no relief”); see also United States v. White, 366 F.3d 291, 297 (4th Cir. 2004) (“[I]f the parties produce evidence disputing material facts with respect to non-frivolous habeas allegations, a court must hold an evidentiary hearing to resolve those disputes”); United States v. Magini, 973 F.2d 261, 264 (4th Cir. 1992) (holding that an evidentiary hearing is necessary “where material facts are in dispute involving inconsistencies beyond the record”).

The present record also precludes us from determining whether Hudgins’s potential failure to investigate and develop this evidence “prejudiced the defense.” Wiggins, 539 U.S. at 521; Strickland, 466 U.S. at 687. The prejudice inquiry would require us to “consider ‘the totality of the available mitigation evidence — both that adduced at trial, and the evidence adduced in the habeas proceeding’ — and ‘reweigh it against the evidence in aggravation.’” Porter v. McCollum, 558 U.S. 30, 41 (2009) (quoting Williams v. Taylor, 529 U.S. 362, 397–98 (2000) (cleaned up)). That inquiry, however, cannot be undertaken until we know the evidence “adduced in the habeas proceeding.” Id. Therefore, if the district court were to conclude that the failure to investigate and develop this evidence was indeed not the product of a strategic decision made after a reasonably limited investigation, *26 the court would have to determine whether that failure “prejudiced the defense.” Wiggins, 539 U.S. at 521; Strickland, 466 U.S. at 687.

In short, we conclude that Runyon has made a colorable claim that his trial counsel’s performance was objectively unreasonable and that the material facts necessary to resolve this issue are fairly in dispute. We therefore vacate this aspect of the district court’s § 2255 order and remand for an evidentiary hearing.

IV

On the third issue certified for appeal, Runyon contends that the government violated its duty under Brady v. Maryland, 373 U.S. 83 (1963), to provide him with exculpatory evidence before trial when the government failed to disclose his codefendant Draven’s history of sexual assault. In the alternative, he argues that his trial counsel’s failure to investigate that history constituted ineffective assistance of counsel. The withheld evidence showed that Draven “sexually assaulted/abused/stalked or otherwise assaulted a number of individuals, including his younger brother and sister,” and underlying documents showed that Draven had sexually abused a developmentally delayed 15-year-old girl and at least six children under the age of seven. Runyon contends that the evidence would have been useful to him in two respects — (1) that it would have provided him with additional evidence in support of the mitigator that the equally culpable codefendant Draven faced only life imprisonment, and (2) that it would weaken the proof of the aggravator that Runyon had engaged in a recurring pattern of domestic violence, because the evidence of Runyon’s recurring pattern was, as he argues, less serious than *27 Draven’s history. Runyon’s history consisted of an ultimately dismissed 1994 charge of assaulting his girlfriend; a 2001 conviction of misdemeanor simple battery for grabbing his wife’s arm and poking her nose; and a 2007 protective order against him on the petition of another girlfriend, who claimed that he had given her a black eye, though the charges related to this incident were dismissed when the girlfriend failed to appear in court.

The district court denied this claim, holding that Runyon failed to establish a Brady violation and that counsel was not ineffective in failing to discover this evidence.

We agree. Even if the evidence could be considered exculpatory — which is not altogether clear — Runyon fails to establish that its suppression caused him prejudice. Brady only bars the suppression of material evidence. Walker v. Kelly, 589 F.3d 127, 137 (4th Cir. 2009). And 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,” and a “‘reasonable probability’ is a probability sufficient to undermine confidence in the outcome.” United States v. Bagley, 473 U.S. 667, 682 (1985); see also Juniper v. Zook, 876 F.3d 551, 567 (4th Cir. 2017) (“[S]uppressed, exculpatory evidence is ‘material’ if it ‘could reasonably be taken to put the whole case in such a different light as to undermine confidence in the verdict’” (quoting Kyles v. Whitley, 514 U.S. 419, 435 (1995))).

At the penalty phase of trial, Runyon proved and the jury found as a mitigator that equally culpable codefendants Draven and Catherina faced only life imprisonment, whereas he faced the death penalty. So the Draven evidence was not needed to establish *28 the mitigator. Runyon contends nonetheless that the evidence would have added to the weight that the mitigator carried.

Yet, it is hard to fathom in these circumstances how the jury would have acted differently if presented with the Draven evidence. Runyon’s suggestion that it would have increased the weight of the equally-culpable-codefendant mitigator and diminished the weight of the domestic-violence aggravator to the extent of producing a different outcome is most speculative. While the argument might be a theoretical possibility, we cannot conclude that there is a “reasonable probability” that the disclosure would have changed anything at sentencing. Bagley, 473 U.S. at 682. This is especially so in view of the fact that the jury included in its calculus eight other mitigating factors, as well as a total of six aggravating factors. Moreover, either 10 or 11 of the jurors found another four mitigators. Yet, in light of all the mitigation found, the jury unanimously recommended the death sentence. Given the willingness of the jury to recommend death in the face of all of these mitigating factors, there is no reasonable probability that a more detailed understanding of Draven’s criminal history would have made a difference.

At bottom, whether Runyon presses his claim as a direct Brady violation or as a claim of ineffective assistance of counsel, the prejudice standard for both is the same, see Bagley, 473 U.S. at 682, and Runyon has failed to satisfy that standard. Accordingly, we affirm the district court on this issue.

V

On the fourth and final issue certified for appeal, Runyon contends that the government exercised its peremptory strikes of potential jurors in a racially discriminatory manner, in violation of Batson v. Kentucky, 476 U.S. 79 (1986). Alternatively, he contends that counsel unreasonably failed to challenge the government’s strikes both at trial and on direct review. In making his claims, he relies principally on the fact that the government “struck 70% of the African-American jurors and did not strike similarly-situated nonblack jurors.”

Because the Batson issue was not raised at trial, the district court concluded that the claim was procedurally defaulted. In any event, the court held further that the claim — whether couched as an ineffective assistance of counsel claim under Strickland to provide cause for the default or as an independent Strickland claim — failed because Runyon failed to establish a Batson violation.

The pool of all prospective jurors in this case consisted of 243 persons, of whom 55 (22.6%) were Black. That pool were given questionnaires, and from the answers given, counsel agreed to a list of 62 persons from whom the jury would be selected. After the trial court struck 10 of those 62 persons for cause, a pool of 52 potential jurors remained, of whom 10 (19%) were Black. Against the 52-member venire, Runyon exercised 20 peremptory strikes — none against Black potential jurors — and the government exercised 19 of its 20 peremptory strikes — 7 against Black potential jurors — leaving the empaneled jury to consist of 9 White members and 3 Black members (25%). Neither counsel objected to the jury that was selected or the process pursued to select the jury, despite the trial court’s *30 repeated inquiries as to whether the parties objected to the dismissal of potential jurors during the voir dire process.

For the first time in his § 2255 motion — six years after trial — Runyon asserted a Batson claim based mainly on the government’s striking of seven Black members of the venire. In response, the government refers to the questionnaires, which asked potential jurors about their attitudes on the death penalty. According to the government, the Black members were struck because they had circled either the anti-death-penalty answer or both the anti-death-penalty and the pro-death-penalty answer. Runyon responds that the government’s explanation is pretextual because the government did not strike two nonBlack potential jurors who had also circled the anti-death-penalty answer.

At the outset, we agree with the district court that because Runyon did not bring his Batson claim on direct appeal, we cannot grant relief unless Runyon can demonstrate cause and prejudice. See United States v. Pettiford, 612 F.3d 270, 280 (4th Cir. 2010). Runyon therefore advances a Strickland claim both as cause for the default and as a freestanding claim, contending for both contexts that his counsel were ineffective in failing to raise the Batson issue during trial or on direct appeal. In the alternative, Runyon contends that his appellate counsel’s failure is excused by the fact that the strike lists were not entered in the docket when counsel filed their brief on appeal. No matter the claim’s packaging, though, we conclude that Runyon has not made out a Batson claim.

Under Batson, a defendant must carry the burden to “make out a prima facie case ‘by showing that the totality of the relevant facts gives rise to an inference of discriminatory purpose.’” Johnson v. California, 545 U.S. 162, 168 (2005) (quoting Batson, 476 U.S. at *31 93–94). Only if the defendant carries the burden does the government need to provide an explanation “by offering permissible race-neutral justifications for the strikes.” Id.

To make his case, Runyon relies first and foremost on the statistical fact that the government struck 70% of the Black potential jurors from the venire panel. In addition, Runyon relies on a statistical test known as Fisher’s Exact Test to contend that there is only a 2.6% likelihood that the strikes of the Black potential jurors were independent of their race. Runyon points also to facts from trial to suggest that the government was motivated by racial animus. For example, he argues that the government’s decision to seek the death penalty against Runyon, a person of Asian descent, and not against the White codefendants, suggests discrimination. Similarly, he suggests that the government’s plan to introduce a videotaped interrogation during trial that included discriminatory questions supports an inference of racial animus.

First, those collateral facts hardly help. We have already held that the introduction of the videotaped interrogation was harmless error, and, in any event, any connection between this video and the government’s behavior during jury selection is too attenuated to support an inference of discrimination. See Runyon, 707 F.3d at 492, 498–99. We are similarly unpersuaded by Runyon’s contention that the government manifested discriminatory animus when it chose to seek the death penalty only against him, and not against two White codefendants. This argument fails to take into account the obvious fact that Runyon was the member of the conspiracy who actually killed Voss.

Turning to the statistics, Runyon’s use of them is “both selective and uninformative.” Allen v. Lee, 366 F.3d 319, 330 (4th Cir. 2004) (en banc). At the prima *32 facie stage, we must look to the “totality of the relevant facts.” Johnson, 545 U.S. at 168 (quoting Batson, 476 U.S. at 94). And one of those facts is, as the government points out, that the seven struck Black potential jurors all expressed reservations about the death penalty. But even putting that fact aside, the statistics on which Runyon relies are misleading and overlook the larger reality. The initial jury pool of 243 persons included 55 Black potential jurors, or 22.6% of the total; the 52-person venire from whom the jury was selected included 10 Black potential jurors, or 19% of the total; and the jury ultimately empaneled included 3 Black jurors, or 25% of the total. We recognize that “the Constitution forbids striking even a single prospective juror for a discriminatory purpose,” Flowers v. Mississippi, 139 S. Ct. 2228, 2248 (2019), and that “[t]he defendant’s constitutional right is not to a specified percentage of minority jurors, but to a process in which the state considers prospective jurors wholly independent of their race,” United States v. Grandison, 885 F.2d 143, 148 (4th Cir. 1989). But the relatively steady percentage of Black potential jurors throughout jury selection — and indeed the slightly increased percentage of Black members on the empaneled jury — undermines Runyon’s reliance on a statistical pattern to create an inference of discrimination. See Allen, 366 F.3d at 330.

At bottom, we conclude that Runyon fails to establish a prima facie case of discrimination under Batson. And because Runyon cannot make such a showing, we conclude that neither trial counsel nor appellate counsel were constitutionally ineffective in failing to make such a claim. Moreover, Runyon’s inability to establish a prima facie *33 case of discrimination indicates that there can be no prejudice in appellate counsel’s failure to bring this claim. We affirm the district court’s order on this issue.

• • * *

In sum, we vacate the district court’s order dismissing Runyon’s § 2255 motion to the extent that it dismissed without a hearing Runyon’s claim of ineffective assistance of counsel for failure to investigate and present evidence of his brain damage and mental health and remand that claim for a hearing. Otherwise, we affirm.

AFFIRMED IN PART; VACATED IN PART AND REMANDED.

GREGORY, Chief Judge, concurring in part and dissenting in part:

I am pleased to concur in my colleague’s opinion remanding this case for a hearing regarding defense counsel’s mitigation investigation. The jury in this case never heard that David Runyon experienced multiple head injuries and that he may suffer from multiple mental illnesses that would significantly affect his decision-making, including post-traumatic stress disorder, delusions, and psychosis. Even at this early stage, declarations from defense counsel in the record strongly indicate that counsel failed to present this evidence not due to a reasoned strategic decision, but because the mitigation investigation was never completed. I write separately because I would expand the scope of the hearing on remand to include Runyon’s Batson, Brady, and related guilt-phase ineffective-assistance claims, and order discovery that the district court declined to allow. And while I disagree with my colleagues’ analysis of Runyon’s 18 U.S.C. § 924 claim, I would not reach that claim at this stage because the appropriate adjudication of Runyon’s Batson claim could result in a new trial.

Runyon was represented at trial by two attorneys, one of whom bore primary responsibility for the guilt phase and the other of whom bore primary responsibility for the penalty phase. As my colleagues and I noted in his direct appeal, the government’s evidence implicating Runyon in the crime was overwhelming, United States v. Runyon, 707 F.3d 475, 486 (4th Cir. 2013), increasing the importance of jury selection as well as the mitigation phase of trial.

Four months before trial began, Runyon’s penalty-phase attorney withdrew from the case because of a conflict, and another attorney was appointed in his place. Habeas *35 counsel notes that Runyon’s new attorney had just over 90 working days to prepare for trial, 57 of which he spent in court on other cases. Opening Br. 54. The district court recessed the trial for an additional four weeks between the guilt and penalty phases. After he was appointed, Runyon’s new attorney does not remember ever communicating with his predecessor. 1 JA 2803. In a declaration, the mitigation specialist remembers that Runyon’s new attorney appeared “overwhelm[ed]” by the task of preparing for the trial on a tight timeline. 2 JA 2007. Runyon’s guilt-phase attorney, from her perspective, did not seem as worried. She remembers him making an alarming comment about the penalty phase of trial: “[W]e’ll throw a couple of relatives on the stand and that will be enough.”

JA 2006.

1 This alone was a violation of the American Bar Association’s Guidelines for the Appointment and Performance of Defense Counsel in Death Penalty Cases (“ABA Guidelines”), reprinted in 31 Hofstra L. Rev. 1049, 1053–54 (2003). See ABA Guidelines 10.7(B)(1) at 1015 (“Counsel at every stage have an obligation to conduct a full examination of the defense provided to the client at all prior phases of the case. This obligation includes at minimum interviewing prior counsel and members of the defense team and examining the files of prior counsel.”). While violations of the ABA Guidelines are not by themselves sufficient to establish ineffective assistance of counsel, they represent “well-defined norms” that serve as “guides to determining what is reasonable.” Wiggins v. Smith, 539 U.S. 510, 523, 524 (2003) (internal quotation marks omitted). 2 Regardless of the trial court’s scheduling decisions, the commentary to the ABA Guidelines explicitly notes that it is

each attorney’s duty under the Model Rules of Professional Conduct neither to accept employment when it would jeopardize the lawyer’s ability to render competent representation nor to handle cases without ‘adequate preparation.’ . . . [A]n attorney whose workload threatens to cause a breach of his or her obligations under these Guidelines has a duty to take corrective action. Counsel in that situation may not simply attempt to muddle through.

ABA Guideline 10.3 Commentary at 998.

The parties spent about two days seating a jury for this capital trial. 09-11 JA 500– 04, 09-11 JA 636–39, 09-11 JA 697–99. As habeas counsel for Runyon has noted, this appears (based on other capital cases reviewed by the Supreme Court) to be an unusually short period of time. See Uttecht v. Brown, 551 U.S. 1, 10 (2007) (11 days); Miller-EI v. Dretke, 545 U.S. 231, 275 (2005) (five weeks); Penry v. Johnson, 532 U.S. 782, 801 (2001) (one month); Johnson v. Texas, 509 U.S. 350, 357 (1993) (15 days); Swindler v. Lockhart, 495 U.S. 911, 913 (1990) (five days); Press-Enterprise Co. v. Superior Court of Cal., 464 U.S. 501, 503 (1984) (six weeks); Irvin v. Dowd, 366 U.S. 717, 720 (1961) (four weeks). Also unusually, counsel principally relied on written questionnaires and answers to questions posed to the entire venire, rather than in individual voir dire, for death qualification. See, e.g., American Bar Association’s Guidelines for the Appointment and Performance of Defense Counsel in Death Penalty Cases (“ABA Guidelines”), reprinted in 31 Hofstra L. Rev. 1049, 1053–54 (2003), at 1031 (indicating that “requesting individual, sequestered voir dire on death-qualification of the jury” is a feature of death penalty representation). Counsel also agreed to strike before questioning, without any attempted rehabilitation, those jurors who indicated that imposing the death penalty would be difficult or impossible for them. JA 234.

During voir dire, the government struck seven of the ten African American members of the venire, and counsel explains on appeal that there is just a 2.6 percent chance that the government’s strikes of these jurors were unrelated to race. Opening Br. at 114. The defendant in this case is Asian and the victim is white, increasing the need for attention to race during voir dire. And “the history of capital punishment in this country is *37 intimately bound up with its history of race relations,” increasing the potential for non-white jurors to be struck from the jury pool. See ABA Guidelines 10.10.2 Commentary at 1053. But counsel was apparently neither alert to nor concerned with the government’s likely-discriminatory strikes, and never made a Batson challenge.

The defense’s mitigation evidence took as little time as voir dire. The evidence presented related almost exclusively to two mitigators: that Runyon would not pose a threat to staff or other prisoners if sentenced to life in prison instead of death, and that Runyon had other good qualities as well as family and friends who did not want him to receive the death penalty. The defense called one psychologist, Mark Cunningham, but he testified only that Runyon was unlikely to be dangerous while incarcerated. 09-11 JA 2625. Six correctional employees testified that Runyon behaved appropriately in prison. 09-11 JA 2713; 09-11 JA 2723; 09-11 JA 2731; 09-11 JA 2740; 09-11 JA 2748; 09-11 JA 2753. The rest of the defense witnesses were friends, family, and a military records specialist who offered testimony relevant to Runyon’s good character, with one witness characteristically noting that he was a “heck of a guy.” 09-11 JA 2976. Friends and family testified that the idea that Runyon had murdered someone was, as one put it, “very out of character.” 09-11 JA 2805. His mother testified that the murder must have been committed by Runyon’s “identical evil twin.” 09-11 JA 2953. None of the friends or family called by the defense to testify in mitigation had seen Runyon for years prior to the murder, and the defense offered no explanation for the lengthy time period during which Runyon lost contact with them (even though Runyon’s mental illness and brain injury would offer one). The defense did not call any other witnesses in mitigation. The overall effect from the *38 transcript is that of a series of witnesses exalting Runyon’s good character in front of a jury that had just found him guilty of murder on the basis of overwhelming evidence. Defense counsel highlighted this dissonance in closing, noting that “[c]learly there was something in Mr. Runyon that [the defense witnesses] didn’t see at the time they interacted with him.” JA 1237–38. The absence of any explanation whatsoever was glaring enough that counsel went on to muse aloud to the jury, as the majority notes: “So a legitimate question, then, would be, well, what happened? . . . And I don’t have a glib answer. I don’t have, you know, a pat response, where I stand up here and say, you know, right here is what caused Mr. Runyon to get involved in this and to be where we are today.” JA 1239.

In summary, the defense elicited very little evidence in mitigation that might help explain why Runyon could have done what he did, even highlighting this point in closing. But on the record before the Court now, it appears that such evidence—and not just a little, but a lot of it—could have been available to counsel. As the majority opinion notes, before trial, two of the government’s own experts noted Runyon’s history of head trauma and neuropsychologists hired by the defense identified potential mental illness requiring further investigation. The habeas record reveals a well of other evidence casting serious doubt on Runyon’s mental health, none of which was highlighted at trial or, at least on this record, thoroughly followed up on by counsel. See JA 1899–1916. As it is, the habeas record describes Runyon’s significant physical and emotional trauma, including brain trauma, from childhood; head injuries from a head-on collision with a drunk driver during Runyon’s time in the Army; employment and other problems consistent with brain injury; letters to trial counsel in which Runyon bragged about saving multiple lives; medical *39 records from jail during the pendency of the case indicating Runyon’s delusions; and family history of significant trauma and mental health problems. Id. Of these, only information about Runyon’s relationship with his family was even incidentally mentioned in mitigation, and defense counsel had only this to say about it in closing: “I don’t suspect that she was the perfect mother. I don’t suspect Mr. Runyon was the perfect son.” Id.

If the defense failed to complete the mitigation investigation in this case, that failure is inexcusable. Mitigation is the core focus of a capital trial where the government, as here, has “a wealth of evidence proving” that the defendant committed the crime. Runyon, 707 F.3d at 486; see, e.g., Jurywork § 23:21 (“[M]ost capital cases remain penalty[-]phase cases.”). That is not to be cynical. Evidence presented in mitigation allows jurors to “accurately gauge” a defendant’s “moral culpability” for a crime. Porter v. McCollum, 558 U.S. 30, 41 (2009). Although a direct causal connection is not a threshold requirement for the introduction of mitigation evidence, evidence about a defendant that helps explain why he committed a crime “is relevant because of the belief, long held by this society, that defendants who commit criminal acts that are attributable to a disadvantaged background . . . may be less culpable.” Penry v. Lynaugh, 492 U.S. 302, 319 (1989), abrogated on other grounds by Atkins v. Virginia, 536 U.S. 304 (2002); Eddings v. Oklahoma, 455 U.S. 104, 114 (1982). Indeed, the death penalty was only constitutionally imposed in this case if the penalty phase of trial is reliable, because this is the crucial step that narrows the penalty’s imposition on only the worst offenders. See, e.g., Zant v. Stephens, 462 U.S. 862, 876–79 (1983). Moreover, mental health mitigation evidence is the heartland of death penalty representation. The ABA Guidelines note that “mental health issues are so *40 ubiquitous in capital defense representation that the provision of resources in that area should be routine.” ABA Guidelines 4.1 Commentary at 957. Somehow, Runyon’s counsel appear to have missed the message.

Accordingly, I am pleased to join in remanding this case to the district court in order to further develop the evidence regarding counsel’s mitigation investigation. On remand, the district court will have the opportunity to consider whether “counsel chose to abandon their investigation at an unreasonable juncture.” Wiggins v. Smith, 539 U.S. 510, 527–28 (2003). If so, then making “a fully informed decision with respect to sentencing strategy” was “impossible,” id. at 527–28—and Runyon is entitled to a new sentencing proceeding.

But I would additionally remand for a hearing and discovery on Runyon’s Batson, Brady, and related ineffectiveness claims. A prisoner is entitled to a hearing “[u]nless the motion and the files and records of the case conclusively show that the prisoner is entitled to no relief.” 28 U.S.C. § 2255. We review the denial of such a hearing for the abuse of discretion. Conaway v. Polk, 453 F.3d 567, 582 (4th Cir. 2006). I would conclude that the district court abused its discretion where, as here—in Runyon’s first and only significant opportunity to address these issues—it denied an evidentiary hearing in the absence of conclusive records showing that Runyon is not entitled to relief. Runyon is not on a fishing expedition. Both his Brady and Batson claims, and the related discovery requests, are well-supported by the habeas record, and he seeks an evidentiary hearing on narrow topics arising directly from the government and defense counsel’s conduct at trial.

With regard to Brady, habeas counsel argues that the government failed to disclose evidence that could change how a jury would weigh Runyon’s desert of the death penalty in comparison to a codefendant sentenced to life, Michael Draven, who allegedly has a lengthy history of sexually assaulting children. Habeas counsel offers evidence that the government knew about this history before trial, yet failed to disclose it to Runyon. Runyon’s trial counsel writes in a declaration that, had he known at trial the same information about Draven that the government knew, he “would have used this information in support of the mitigation case” because it “supported the argument that Draven was a bad actor” and that “the co-defendants were just as culpable, if not more so” than Runyon. JA 2805, 2806. Runyon now seeks to know exactly what evidence prosecutors withheld from defense counsel and, with regard to a related ineffectiveness argument, why defense counsel did not make further inquiry into this potentially mitigating evidence after some of it was revealed by Draven’s counsel as he cross-examined a witness during the guilt phase. In a death case—which requires “heightened reliability,” see, e.g., Sumner v. Shuman, 483 U.S. 66, 72 (1987)—I would find that withheld information bearing on the weight of aggravating and mitigating factors is material for purposes of Brady. When a jury is prevented from “giving independent mitigating weight to aspects of the defendant’s character and to circumstances of the offense proffered in mitigation,” it “creates the risk that the death penalty will be imposed in spite of factors which may call for a less severe penalty.” Lockett v. Ohio, 438 U.S. 586, 605 (1978). Without an evidentiary hearing and discovery, I find the record before the Court insufficient to determine whether the government withheld evidence creating such a risk.

As for Batson, I agree with my colleagues that Runyon must establish cause and prejudice in order to obtain relief on this claim, which was not made at trial. But “if the procedural default is the result of ineffective assistance of counsel, the Sixth Amendment itself requires that responsibility for the default be imputed to the State.” Coleman v. Thompson, 501 U.S. 722, 753–54 (1991). Defense counsel’s unexplained and unusual decisions during the capital voir dire process (including, most critically, failing to object to the government’s disproportionate striking of non-white members of the venire) warrant a hearing. And on the merits of his Batson claim, Runyon seeks the government’s voir dire records in discovery only after showing, using Fisher’s Exact Test, that the government’s strikes were very likely prejudicial. 3 Runyon has established that he may be entitled to relief under Batson. The district court’s rebuke that records requested by the defendant seek “just more statistics” “involv[ing] choices by both parties” does not foreclose the possibility that the government impermissibly struck jurors on the basis of race. I would accordingly reverse the district court’s denial of Runyon’s request for discovery and a hearing related to his Batson claim.

3 As Runyon’s counsel notes, this Court has previously relied on Fisher’s Exact Test in adjudicating discrimination claims. Love v. Alamance Cty. Bd. of Educ., 757 F.2d 1504, 1510 n.4 (4th Cir. 1985).

WILKINSON, Circuit Judge, concurring in part and dissenting in part:

I readily concur in Parts II, IV, and V of Judge Niemeyer’s fine opinion. I just as readily dissent from the disposition of the ineffective-assistance-of-counsel claim in Part III. Runyon’s lawyer performed admirably throughout, and to expect that he could have produced a different result in a case featuring such a cold and calculated murder for hire is wholly unrealistic. I have great respect for those who hold principled objections to capital punishment no matter the circumstances. None of us on the bench take these weighty decisions lightly, but I am not persuaded that the rule of law is reputationally advanced when these proceedings are allowed to drag out indefinitely.

Our criminal justice system depends on finality. Trials and sentencing proceedings are often long and complicated affairs, placing strain on the parties, lawyers, judges, witnesses, and jurors involved in them. We should not lightly snub their efforts to make our justice system work. And yet contemporary collateral attacks threaten to cement trials in subordinate status. Collateral cases inevitably feature a losing defendant, and they thus provide a temptation to pin the blame on who else?—his lawyer. That temptation is especially strong in capital cases like this one, where good and able attorneys feel pressure to demean their own sound trial performance in order to assist their erstwhile client’s collateral case. See Ellen Henak, When the Interests of Self, Clients, and Colleagues Collide: The Ethics of Ineffective Assistance of Counsel Claims, 33 Am. J. Trial Advoc. 347, 348 (2009).

In Strickland v. Washington, 466 U.S. 668 (1984), the Supreme Court held that judges must not set aside sentences due to ineffective assistance of counsel absent manifest incompetence that falls outside “the wide range of reasonable professional assistance” permissible under the Constitution. Id. at 689. When a lawyer makes a strategic decision, courts must indulge a “strong presumption” that the decision was reasonable. Id. at 689. After all, “the purpose of the effective assistance guarantee of the Sixth Amendment is not to improve the quality of legal representation.” Id.

How far we have strayed from Strickland. That case provides another illustration of an all too common phenomenon. The Supreme Court acts in the belief that it is only cracking the door. Litigants then entice judges to follow their own policy predilections and fling the cracked door open wide. The trickle becomes a torrent as the ubiquitous IAC claims now attest. This case provides a good example of exactly that.

I.

Let’s first take a brief look at Strickland itself. An examination of the facts in Strickland, where the Supreme Court found counsel was not constitutionally ineffective, highlights the incorrectness of today’s decision. The defendant in Strickland, David Washington, pled guilty to a string of brutal stabbing murders and other violent crimes. His court-appointed lawyer, William Tunkey, thus faced the unenviable task of convincing the sentencing judge to spare his client the death penalty. Overcome by a “sense of hopelessness” about his client’s chances, id. at 673, counsel Tunkey did little to prepare for the sentencing hearing. He interviewed his own client to learn about his background and he spoke to Washington’s wife and mother over the telephone. Id. He tried and failed *45 to meet with family members. Id. He did not investigate psychiatric evidence. Id. He did not seek out character witnesses for Washington. Id. He did not look for new evidence about his client’s mental state. Id.

At the sentencing, counsel Tunkey made arguments based on his client’s plea colloquy and the sentencing judge’s prior statement expressing appreciation for criminals who take responsibility for their actions. He argued that Washington should be spared the death penalty because he had taken responsibility for his actions by pleading guilty, that he had a minor criminal history, and that he “was fundamentally a good person who had briefly gone badly wrong in extremely stressful circumstances.” Id. at 674. Counsel Tunkey did not introduce any evidence or even cross-examine the state’s many witnesses, including medical experts who described Washington’s gruesome acts. Id.

This lackluster strategy failed, and Washington was sentenced to death. When the Supreme Court reviewed this case, it must have been sorely tempted to give Washington another chance to avoid the death penalty. It was not difficult to imagine that a better lawyer would have put up a better fight. And the Justices would likely have slept better if Washington had merely been sentenced to imprisonment.

But the Supreme Court upheld the sentence, finding that counsel Tunkey’s uninspiring performance was constitutionally adequate. Id. at 698–700. In the process, it established fundamental rules that protect the finality of trials and sentencing proceedings. To establish ineffective assistance of counsel, a defendant must show that his lawyer provided incompetent representation and that he was prejudiced by his attorney’s errors. Id. at 687. Both requirements present high bars.

There is a “strong presumption” that lawyers rendered adequate service. Id. at 688. The Court insisted that judges must not impose specific requirements on lawyers, lest they “interfere with the constitutionally protected independence of counsel and restrict the wide latitude counsel must have in making tactical decisions.” Id. at 689. “Judicial scrutiny of a counsel’s performance must be highly deferential” and “every effort [must] be made to eliminate the distorting effects of hindsight” when assessing counsel’s adequacy. Id.

The Court further mandated that judges should rarely second-guess strategic decisions by counsel. “[C]hoices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable . . . .” Id. at 690. And this duty to investigate does not include the duty to look under every conceivable rock for potential evidence. With limited time, lawyers must make difficult decisions about what to investigate depending on the circumstances. “In any ineffectiveness case, a particular decision not to investigate must be directly assessed for reasonableness in all the circumstances, applying a heavy measure of deference to counsel's judgments.” Andrus v. Texas, 140 S. Ct. 1875, 1881 (2020) (quoting Wiggins v. Smith, 539 U.S. 510, 521 (2003)).

The prejudice requirement was similarly demanding. The defendant must “affirmatively prove” that any unprofessional assistance created a “reasonable probability” that the outcome would have been different. Strickland, 466 U.S. at 693. A reasonable probability is “a probability sufficient to undermine confidence in the outcome.” Id. at 694. Where the government’s evidence is strong, an attorney error is less likely to be prejudicial. Id. at 696 (“Moreover, a verdict or conclusion only weakly supported by the record is more likely to have been affected by errors than one with overwhelming record support.”).

II.

A.

The majority’s decision flouts Strickland. Counsel Hudgins conducted a thorough investigation and made strategic decisions that were eminently reasonable.

The majority faults counsel Hudgins for performing an inadequate investigation. That is an odd conclusion, considering counsel Hudgins did far more than the lawyer in Strickland. Unlike Washington’s lawyer, counsel Hudgins worked closely with Runyon’s family members. Unlike Washington’s lawyer, counsel Hudgins introduced character witnesses at the sentencing hearing. Unlike Washington’s lawyer, counsel Hudgins introduced a substantial amount of new evidence to prove mitigating factors. Whereas Washington’s lawyer relied almost exclusively on his client’s prior plea colloquy, counsel Hudgins introduced multiple witnesses—including family members, several prison guards, and a mental health expert.

And unlike Washington’s lawyer, counsel Hudgins did investigate Runyon’s mental health. Counsel Hudgins was assisted by seven mental health experts during the trial. As Hudgins prepared for sentencing, several of these experts advised him. As counsel Hudgins was waiting for the results of PET and MRI scans performed on Runyon, Dr. Mirsky prepared a preliminary report advising that Runyon might be suffering from mental health problems. J.A. 2131. But the MRI report indicated that results were normal, and the PET scan was non-diagnostic. See J.A. 2133. Dr. Merikangas reported that Runyon was suffering from delusions or was in a “fantasy world of grandiose wishful thinking,” and that he suffered from the effects of or withdrawal from his drug testing medication. J.A. *48 1996–97. Dr. Cunningham testified at the sentencing hearing and was presumably communicating with counsel Hudgins throughout the process. This investigation went light years beyond what Washington’s lawyer conducted.

Based on that investigation and preparation, counsel Hudgins made an eminently reasonable decision not to prioritize mental health evidence at the sentencing hearing. Before the sentencing jury, counsel Hudgins submitted evidence and presented arguments on several mitigating factors. He argued that Runyon was a decent person and that his behavior in the case was anomalous. See, e.g., J.A. 1167 (adducing testimony from Runyon’s mother that she never saw him “be unkind to anybody, anything”). He presented evidence about Runyon’s past, including lay testimony from family and friends, detailing the abuse Runyon suffered as a child and the suffering his family would bear were he executed. J.A. 1095–1128, 1131–67. Counsel Hudgins introduced testimony from several officers at the jail to argue Runyon would be well behaved in prison and would be unlikely to commit violent acts if his life were spared. J.A. 1049–94. To further bolster this argument, one of his mental health experts, Dr. Cunningham, gave a lengthy presentation to the jury on this point. J.A. 1247–48. Finally, counsel Hudgins argued that it would be unfair to execute Runyon because his co-conspirators were not going to be executed. J.A. 1235, 1239–44.

Tellingly, all of counsel Hudgins’ arguments and efforts at sentencing get boiled down into a single uninformative sentence in the majority opinion. See Maj. Op at 19 (“Hudgins did, however, propose 14 other mitigation factors and offer testimony from nearly two dozen witnesses.”). This scant analysis fails to accurately convey the substantial *49 efforts counsel Hudgins made at sentencing, misleading the reader into thinking he barely put up a fight. It is also utterly inconsistent with the Supreme Court’s command that we indulge a “strong presumption” that counsel Hudgins’ efforts were adequate and strategically sound. Strickland, 466 U.S. at 688.

Several considerations bolster the legal presumption that counsel Hudgins’ strategy was reasonable. First, it was partly successful. The jury found both statutory mitigating factors and eight non-statutory mitigating factors argued for by counsel Hudgins. This is not a case where the defense failed to put up a fight at sentencing. See Rompilla v. Beard, 545 U.S. 374, 378 (2005) (documenting the comparatively minor efforts made by counsel at sentencing before finding ineffective assistance of counsel). Unlike Washington’s lawyer, counsel Hudgins submitted a substantial amount of mitigation evidence.

Second, counsel Hudgins’ decision to focus on Runyon’s lack of dangerousness in prison was reasonable based on the government’s strategy. As an aggravating factor, the government successfully argued that Runyon had used his military and police training to execute a murder for hire. Thus, it makes sense that counsel Hudgins devoted a substantial part of his strategy to arguing Runyon would not pose a threat of violence in prison. Counsel Hudgins’ fear that the jury would see Runyon as dangerous was apparently justified, as the jury found every mitigating factor the defense argued for except those involving Runyon’s adjustment to custody and his lack of dangerousness while incarcerated. J.A. 1313.

Third, presenting evidence that Runyon was suffering from brain damage and mental health problems could have undermined counsel Hudgins’ argument that Runyon *50 would not pose a threat of violence in prison. It could have underscored Runyon’s problems with impulsiveness and anger management to the obvious detriment of the gentle and pacific individual counsel Hudgins was attempting to portray. This evidence thus would have functioned as a “two-edged sword,” Penry v. Lynaugh, 492 U.S. 302, 324 (1989), and counsel Hudgins acted reasonably in not introducing it.

Fourth, introducing more mental health evidence could have opened the door to damaging mental health evidence from the government. The government’s mental health experts could have testified, based on their examinations of Runyon, that he lacked remorse for his actions, J.A. 192–93, that he “does not suffer from any serious mental illness,” J.A. 199, that he had narcissistic features, J.A. 199, that his past head traumas and concussions were not serious, J.A. 200, that he performed above average on intelligence examinations, J.A. 224–25, and that Runyon was not experiencing any unusual level of stress or distress when he committed the crime. J.A. 231. Although conflicting evidence on mental health is to be expected in the adversarial process, counsel Hudgins could reasonably have concluded that his mental health evidence was not strong enough to justify opening the door to the government’s opposing evidence.

Fifth, presenting more mental health evidence would have clashed with Runyon’s continued insistence—to counsel Hudgins and the government’s mental health evaluators throughout the sentencing process— that he was innocent. J.A. 193, 208–09, 2016. He told the government’s mental health experts and his lawyer that he would not agree to any legal strategy that was inconsistent with his personal belief that he was innocent. J.A. 208–09, 2016. Thus counsel Hudgins had to craft a strategy that did not concede his client’s guilt, *51 and emphasizing mental health-based arguments—like Dr. Dudley’s offer, six years after the trial, to testify that Runyon’s mental health issues reduced his ability to obey the law and made him commit murder—may very well not have been acceptable to his client. “The reasonableness of counsel’s actions may be determined or substantially influenced by the defendant’s own statements or actions [because] Counsel’s actions are usually based, quite properly, on informed strategic choices made by the defendant . . . .” Strickland, 466 U.S. at 691.

What does all of this recitation demonstrate? Strategy. Pure and simple. The majority apparently disagrees with that strategy, suggesting counsel Hudgins should have looked harder and presented more mental health evidence. Frankly, I doubt the majority’s alternative strategy would have worked out better for counsel Hudgins, especially since the jury was looking at an intelligent young man who used military and police training to assassinate a Navy officer in a complex and coolly executed plot. And considering how strong the government’s own mental health evidence was, counsel Hudgins appears to have behaved wisely in not opening the door to it. See Strickland, 466 U.S. at 699 (observing that the failure of Washington’s lawyer to present evidence prevented the government from introducing damaging rebuttal evidence). In fact had counsel Hudgins done so, he would have undoubtedly been facing another Strickland claim to which the majority undoubtedly would have proved sympathetic. 1

1 A brief word as to Chief Judge Gregory’s separate opinion. It regrettably ignores the superior vantage point of the judge who presided at trial and over sentencing and who found the representation afforded Runyon to be fully effective. It then proceeds to second-

But it doesn’t really matter whether the majority’s legal strategy is superior to that of counsel Hudgins. The fact that counsel Hudgins had a strategy, and a quite reasonable one at that, should decide this case. Unlike this court, counsel Hudgins “observed the relevant proceedings, knew of materials outside the record, and interacted with the client, with opposing counsel, and with the judge.” Harrington v. Richter, 562 U.S. 86, 105 (2011). And although mental health evidence will sometimes be relevant at sentencing in death penalty cases, it is important to remember that “[n]o per se rule requires the presentment of [mental health] evidence.” Meyer v. Branker, 506 F.3d 358, 372 (4th Cir. 2007). “There are many strategically valid reasons why defense counsel . . . may decide not to offer mental health mitigation testimony: it may not be persuasive; it may appear to be a ‘flight into theory’ without proper grounding in the facts of the case . . . .” Id. Counsel Hudgins made a strategic decision not to prioritize mental health evidence, and we are required by Strickland to respect that strategic judgment.

B.

Even if I could somehow pretend counsel Hudgins did not adopt a reasonable strategy, any possible error did not prejudice Runyon. “The defendant must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the

guess counsel Hudgins’ strategic judgment, contending that a greater focus on mental health evidence would have swayed the jury in a case where Runyon’s murder for hire was altogether calculated and coolheaded. Further, presenting him as someone unable to exercise impulse control would have badly undermined counsel Hudgins’ argument to the jury that Runyon was not a dangerous person. Both legally and factually, the good Chief Judge’s fleeting stint as Runyon’s defense counsel falters badly.

proceeding would have been different.” See Strickland, 466 U.S. at 694. Each category of evidence that Runyon faults counsel Hudgins for not adducing carried sharp aggravating aspects in addition to its mitigating facets. As the district court recognized, much of the evidence Runyon claims his lawyer should have offered would have undermined the argument that Runyon was not dangerous and undercut Runyon’s continued insistence that he was innocent. Runyon v. United States, 228 F. Supp. 3d 569, 623 (E.D. Va. 2017) (“Much of the evidence that [Runyon] now suggests would have been cumulative, contrary to counsel’s strategy, or simply not of help to [Runyon’s] case.”).

The majority makes much of a letter from counsel Hudgins to Runyon’s new lawyers that notes how some of the evidence they have unearthed, many years after the sentencing, could have helped him. See Maj. Op at 21-22, 24. The only thing this letter demonstrates is the difficult ethical dilemma that an improper and non-deferential application of Strickland creates for the lawyers involved. Although counsel Hudgins no longer represents Runyon, he can still help his old client. When approached and asked to give a statement about new evidence and new arguments, counsel Hudgins undoubtedly knew that he could improve Runyon’s chance of avoiding the death penalty by casting doubt on his own performance. See Susan P. Koniak, Through the Looking Glass of Ethics and the Wrong Rights We Find There, 9 Geo. J. Legal Ethics 1, 7 (1995). Counsel Hudgins thus had to choose between hurting his old client’s legal interests and falling on his own sword. See Henak, supra, at 348. I can hardly blame this fine lawyer for whatever course he took.

Moreover, the evidence Runyon now offers would have been a drop in the bucket compared to the mountain of aggravating evidence the government introduced. As we previously recognized, “this was simply not a close case.” United States v. Runyon, 707 F.3d 475, 519 (4th Cir. 2013). The government proved to the jury that Runyon used military and police training to carry out a systematic plot to murder a Naval officer for money. As we put it on direct appeal:

There is simply no question that Runyon fired five bullets into the body of an innocent naval officer and young father—at close range and in cold blood. There is simply no question that the jury had a strong evidentiary basis for unanimously finding numerous aggravating factors beyond a reasonable doubt—including that Runyon acted with monetary motives, planned and premeditated the murder, exploited his military and law enforcement experience in committing it, and showed not a hint of remorse for inflicting inestimable human damage for a paltry sum.

Id. at 519. There is no reasonable probability that the result would have been different had counsel Hudgins made the strategic choices Runyon now argues were superior to the ones made.

III.

This decision is unfair to the deceased, unfair to his children, unfair to the sentencing jury, and unfair to the able trial judge who has done such a conscientious job with this case. And it is just wrong to diminish as ineffective the professional efforts of an attorney who plainly gave the defense his highly commendable best. Our legal system depends upon the faithful application of Strickland. Every single trial and every single sentencing proceeding can be challenged on the basis that a losing defendant had imperfect legal assistance. These proceedings can be challenged in habeas proceedings many years later, when the danger of *55 could-have/should-have is especially heightened and when every remand sets the stage for yet another post-conviction appeal. On and on it goes. There is always a losing lawyer. There is always an argument that was not made. There is always a possible winning strategy that was not pursued. As the Supreme Court warned in Strickland, 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.” Strickland, 466 U.S. at 689.

If we pick endlessly at the performance of lawyers, the integrity of trials and sentences is destroyed. Id. at 690 (“Criminal trials resolved unfavorably to the defendant would increasingly come to be followed by a second trial, this one of counsel's unsuccessful defense.”). Moreover, because an “ineffective-assistance claim can function as a way to escape rules of waiver and forfeiture and raise issues not presented at trial, [] the Strickland inquiry must be applied with scrupulous care . . . .” Harrington, 562 U.S. at 105. Although I hold no brief for capital punishment, I object to its nullification through detour and delay. Until the representative branches of our government repeal this penalty, I would follow the Supreme Court.