Opinion
No. 23-2125
04-29-2024
Allan KUSTOK, Petitioner-Appellant, v. David MITCHELL, Respondent-Appellee.
Karl A. Leonard, Attorney, Exoneration Project, Debra Loevy, Attorney, Loevy & Loevy, Chicago, IL, for Petitioner-Appellant. Garson Fischer, Attorney, Office of the Attorney General, Chicago, IL, for Respondent-Appellee.
Appeal from the United States District Court for the Northern District of Illinois, Eastern Division. No. 1:21-cv-06932 — Matthew F. Kennelly, Judge. Karl A. Leonard, Attorney, Exoneration Project, Debra Loevy, Attorney, Loevy & Loevy, Chicago, IL, for Petitioner-Appellant. Garson Fischer, Attorney, Office of the Attorney General, Chicago, IL, for Respondent-Appellee. Before Wood, Lee, and Pryor, Circuit Judges. WOOD, Circuit Judge.
In 2014, a Cook County jury convicted Allan Kustok of murdering his wife, Anita "Jeannie" Kustok. Kustok moved for a new trial, arguing that post-trial evidence cast doubt on an expert witness's testimony against him. The state trial court denied Kustok's motion and sentenced him to 60 years' imprisonment. Kustok unsuccessfully challenged the denial of a new trial on direct appeal.
Kustok then brought a state postconviction petition based on a claim of ineffective assistance of trial counsel. He argued that the jury might not have convicted him if his lawyer had discovered the exculpatory evidence before trial, and that the lawyer's failure to do so amounted to a violation of his Sixth Amendment right to counsel, as interpreted by Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). The state postconviction court, and then the Illinois Appellate Court, determined that Kustok had waived his Strickland claim by failing to raise it on direct appeal.
Kustok then brought this federal habeas corpus petition under 28 U.S.C. § 2254, presenting the same Strickland claim that the state courts found to be waived. The district court held that the state-court waiver meant Kustok had procedurally defaulted the claim for federal-court purposes. Finding that reasonable jurists could find this debatable, however, the court granted Kustok a certificate of appealability, and he appealed. We agree that Kustok procedurally defaulted his claim. Furthermore, we conclude that Kustok does not qualify for any exception to the procedural-default rules, and so we affirm the dismissal of his petition.
I
A
On the morning of September 29, 2010, Kustok arrived at Palos Community Hospital carrying the body of his wife, Jeannie. Jeannie had been shot in the head 45 to 90 minutes earlier while she was in the couple's bed.
According to Kustok, Jeannie woke up in the middle of the night after hearing a noise in the house. Kustok investigated and found nothing, but Jeannie brought a revolver to the bedroom for protection against possible intruders. Two hours later, Kustok heard a gunshot and discovered Jeannie's body in the couple's bed. He initially contemplated suicide, but then fired the remaining bullets into a dresser to unload the gun. Knowing that Jeannie was already dead, Kustok lay with her for a while before attempting to clean the blood off her body and taking her to the hospital. He claimed that Jeannie had shot herself, either by accident or to take her own life.
B
The state did not believe Kustok's story. It charged him with first-degree murder, and the case went to trial. Over the course of more than three weeks, the jury heard testimony from a number of people. The state introduced three witnesses who had spoken to Kustok after the incident; one testified that Kustok said he was in the bathroom when the gun went off, while the other two testified that he told them he had been sleeping in bed and was awoken by the gunshot. The state also called the manager of the gun store where Kustok had purchased the revolver three months before Jeannie's death, and the manager testified that Kustok said he was buying the gun for target practice. In addition, the jury heard from Jeannie and Kustok's adult daughter Sarah, who testified that she was not aware of any strife in the marriage and did not know that her parents owned a gun. Two of Jeannie's coworkers, however, said that Jeannie had told them about the gun. Several witnesses added that Jeannie appeared to be happy shortly before her death and had made plans for the future.
Over Kustok's objection, the court admitted testimony from five women who claimed to have had extramarital affairs with Kustok, one of whom testified that he told her he planned to get a divorce. The woman with whom Kustok had the longest affair (five years) testified that although Kustok never spoke ill of Jeannie, he had said that he did not feel fulfilled in his marriage.
Because there were no eyewitnesses, much of the trial centered on a battle of experts. Those experts included three forensic scientists called by the state; their testimony can be summarized briefly. The first stated that the revolver had safety features to prevent accidental discharges; the second reported finding no gunshot residue on Jeannie's hands; and the third said that blood on Kustok's clothing matched Jeannie's profile, but that she had not found Kustok's DNA on the firearm.
The testimony of the next four experts warrants more attention. They focused on two types of stains to determine where Jeannie, Kustok, and the gun were at the time of the gunshot. The first was bloodstains. The experts differentiated between "transfer stains," which show that an object came into contact with blood by touching other stained objects, and "impact spatter," which indicates that blood splashed onto the object as a spray coming from a fresh wound. The second type of stain was soot; some guns expel soot onto nearby surfaces when fired, and so experts try to use soot stains to determine where the gun was fired, relative to any stained surfaces.
The first of these four witnesses was Dr. Hiliary McElligott, a forensic pathologist. Dr. McElligott testified that, upon firing, the revolver would have left soot marks up to about six inches from its barrel. Because she did not find any soot around Jeannie's gunshot wound, Dr. McElligott believed that the gun must have been fired more than six inches from Jeannie's face. Dr. McElligott found that distance to be inconsistent with Kustok's suicide theory, reasoning that most self-inflicted gunshot wounds are contact wounds. Her conclusion was bolstered by the path of the bullet: Kustok had claimed to find the gun in Jeannie's right hand, which rested on her chest, but forensic examination showed that the bullet entered Jeannie's face on the left side and traveled downward. Dr. McElligott found it unlikely that Jeannie would have reached across her body to shoot herself, and she noted that any recoil from the gun would have pushed her hand further away from her body rather than onto her chest.
Kustok did not let this testimony go unchallenged. On cross-examination, Dr. McElligott admitted that she had not spent much time considering whether Jeannie's death might have been an accident. More importantly, Kustok offered the testimony of his own expert, Matthew Noedel. Noedel testified that it is not uncommon for the revolver that shot Jeannie to discharge accidentally; that he believed the gun was fired three to six inches from Jeannie's face; and that many self-inflicted gunshot wounds are not accompanied by gunshot residue.
That brings us to the heart of Kustok's Strickland petition: a stain on a pillowcase. The stain became relevant during the testimony of the two remaining expert witnesses. The first was Rod Englert, a retired police officer who testified as the state's expert in crime-scene reconstruction and blood-pattern analysis. The second was Paul Kish, Kustok's own expert on blood-pattern analysis.
Using photographs of the bedroom, Englert had attempted to reconstruct the scene of Jeannie's death, down to the location of each pillow on the bed. He testified that, based on the blood spatters on the gun and on Kustok's clothes, Kustok had been standing above Jeannie when she was shot, meaning that he could not have been asleep or in the bathroom at the time. Englert explained that he discounted the possibility of an accidental discharge because he had not found any soot on the pillows, which indicated that the gun was not resting on or under a pillow when it went off.
Enter the pillowcase stain. On cross-examination, Kustok asked Englert whether a dark spot on one of the pillowcases might be soot. In his reconstruction, Englert had determined that the stained pillow was beneath the pillow in which the bullet was found, meaning that, if Englert's reconstruction were correct, it would have been difficult for soot to travel from the gun to the stained pillow. Englert testified that the prosecution had asked him about that stain before trial, that he had inspected it, and that he had concluded that it was "obviously clotted blood," not soot. He admitted that he had not recommended testing the stain with sodium rhodizonate, a five-minute laboratory test that Englert acknowledged is the standard way to detect soot.
After this testimony (and ten days into the trial), Kustok's lawyer asked the court for permission to test the stain on the pillowcase. The court denied that request, reasoning that Kustok knew the pillowcase existed and should have asked to test it before trial.
That did not leave Kustok without a way to challenge Englert's testimony. Kustok called Kish to the stand, and Kish testified that he did not believe there was enough information accurately to reconstruct the scene of the shooting. He disagreed with Englert's methodology; in his view, the blood stains were inconclusive. When questioning turned to the stain on the pillowcase, Kish testified that it did not resemble a bloodstain and that it should have been tested for soot, although he admitted that he was aware of the stain before trial but had not recommended testing it.
C
The jury found Kustok guilty. His first response was to file a motion for a new trial and to renew his request for permission to test the stain on the pillowcase. This time the prosecution did not object to the latter, and the court granted his request. At the hearing on Kustok's motion for a new trial, Kustok introduced the testimony of forensic scientist Nicole Fun dell, who testified that she had tested the pillowcase and found soot on both its inside and outside. She had also fired the gun into another pillow at various distances to try to recreate the fatal gunshot, and she found that the most similar soot patterns appeared when she fired the gun three inches away from the pillow. Kustok also recalled Noedel, who testified that, based on the soot stain, Englert's reconstruction could not have been accurate.
The trial court opined that the experts' reconstructions "cancel each other out" and that, given the soot stain, "[n]o expert that testified can say how this happened." It nonetheless denied Kustok's motion for a new trial. It reasoned that the soot on the pillowcase was not newly discovered evidence because Kustok could have discovered it before trial with the exercise of due diligence; that evidence about the soot was cumulative of the defense experts' testimony contradicting Englert's reconstruction; and that other evidence overwhelmingly supported the jury's verdict.
Kustok appealed on two grounds: that the trial court should have granted the new trial, and that it erred in admitting testimony about Kustok's extramarital affairs. The Illinois Appellate Court affirmed, agreeing with the trial court that the soot stain was not newly discovered evidence. See People v. Kustok, 2016 IL App (1st) 143812-U, 2016 WL 4268717 (Ill. App. 1st Dist. 2016). The Illinois Supreme Court denied Kustok's request for leave to appeal.
D
Shortly after Kustok's direct appeal, the Supreme Court of Illinois decided a case that has taken center stage in Kustok's collateral challenges to his conviction. That case involved a defendant convicted of two counts of attempted murder. The defendant appealed, asserting that his trial lawyer provided ineffective assistance by stipulating to the admission of key evidence against the defendant. The Illinois Appellate Court concluded that the Strickland claim would be better resolved in a post-conviction petition rather than on direct appeal, and so it dismissed the claim. See People v. Veach, 401 Ill.Dec. 367, 50 N.E.3d 87, 101 (Ill. App. 4th Dist. 2016) (Veach I). The state supreme court reversed, holding that the trial record was sufficient for the appellate court to resolve the claim on direct appeal. See People v. Veach, 417 Ill.Dec. 718, 89 N.E.3d 366, 376 (Ill. 2017) (Veach II). More generally, Veach II held that when a defendant is able to raise a Strickland claim on appeal but fails to do so, she waives that claim for later petitions. Id., 417 Ill.Dec. 718, 89 N.E.3d at 375.
On December 22, 2017, Kustok filed a state postconviction petition, arguing for the first time that his trial lawyer provided constitutionally deficient representation by failing to test the pillowcase stain before trial. Citing Veach II, the post-conviction court held that Kustok had waived his Strickland claim by failing to raise it on direct appeal. The postconviction court also held that even if Kustok had not waived the claim, the court would dismiss his petition on the merits for failure to show either deficient performance or prejudice under Strickland.
Kustok appealed once again. The Illinois Appellate Court agreed that Kustok had waived his claim, but its reasoning differed slightly from that of the postconviction court. Although it acknowledged that Veach II was not decided until after Kustok's direct appeal, it concluded that the rule in Veach II had long been part of Illinois law. People v. Kustok, 2021 IL App (1st) 191899-U, ¶ 50, 2021 WL 1966758. The court also agreed that Kustok's claim failed on the merits, but, unlike the post-conviction court, it chose not to reach Strickland's performance element. It held only that Kustok could not show prejudice, because the results of the sodium rhodizonate test were cumulative evidence and even without Englert's testimony, there was enough evidence against Kustok to support the jury's verdict. Id. ¶¶ 56 -59. The state supreme court again declined to hear Kustok's case.
That brings us at last to this federal habeas corpus petition, in which Kustok raises a single claim. He alleges that his trial lawyer provided ineffective assistance by failing to test the pillowcase stain before trial—the same claim that the state courts had found to be waived. The district court disagreed with the state postconviction court's characterization of the test results as cumulative evidence. It believed that the jury would have found it significant that Englert—who testified at least 15 times that the stain was not soot—had been mistaken. The court nonetheless dismissed Kustok's petition, concluding that he had procedurally defaulted his Strickland claim by waiving it before the state courts. Kustok has appealed.
II
The procedural-default doctrine generally prevents federal courts from hearing "claims that the state court denied based on an adequate and independent state procedural rule." Davila v. Davis, 582 U.S. 521, 527, 137 S.Ct. 2058, 198 L.Ed.2d 603 (2017). To determine whether Kustok's Strickland claim is procedurally defaulted, we first must identify the contours of the state procedural rule in question. For that, we turn to the Illinois Appellate Court's decision dismissing Kustok's postconviction petition, because it was the last state court that rendered judgment on Kustok's claim. See Lee v. Foster, 750 F.3d 687, 693 (7th Cir. 2014). Citing Veach II, that decision invoked the following rule: "[W]hen the facts required to evaluate counsel's conduct appeared in the record on appeal, the traditional rules of default appl[y] if the defendant failed to raise an ineffective assistance claim on direct appeal." People v. Kustok, 2021 IL App (1st) 191899-U, ¶ 50, 2021 WL 1966758.
Our task here is only to determine whether that rule is independent and adequate—not to assess whether the Illinois court applied the rule correctly. The state court found that Kustok could have raised his claim on direct appeal, see id. ¶ 49, and that finding is conclusive on us, see Lee, 750 F.3d at 694. (Even if it were otherwise, Kustok has not disputed the critical proposition that the trial record was sufficient to resolve his Strickland claim on direct appeal.)
Illinois's rule is independent because the Illinois Appellate Court "actually relied on the procedural bar as an independent basis for its disposition of the case." Kaczmarek v. Rednour, 627 F.3d 586, 592 (7th Cir. 2010). The court also dismissed the claim on the merits, but that does not preclude a finding of independence. See Carey v. Saffold, 536 U.S. 214, 226, 122 S.Ct. 2134, 153 L.Ed.2d 260 (2002).
The harder question is whether the rule was adequate. A rule is adequate if it is "firmly established and regularly followed." James v. Kentucky, 466 U.S. 341, 348, 104 S.Ct. 1830, 80 L.Ed.2d 346 (1984). That is a question of federal law that we resolve without deference to the state court. Malone v. Walls, 538 F.3d 744, 753 (7th Cir. 2008). Kustok argues that the Illinois rule was not firmly established at the time of his appeal because it did not exist until the Illinois Supreme Court's decision in Veach II, one year after his direct appeal. If Veach II announced a new rule in 2017, that rule obviously would not have been clearly established at the time of Kustok's direct appeal in 2016.
The problem is that Veach II's rule was not new—far from it. As the Illinois Appellate Court pointed out in dismissing Kustok's petition, the rule has been around for decades. So far as we can tell, the Supreme Court of Illinois first applied its ordinary waiver rule to an ineffectiveness claim in 1968. See People v. Collins, 39 Ill.2d 286, 289, 235 N.E.2d 570 (1968). It has reaffirmed its commitment to applying the waiver rule to Strickland claims time and time again. See, e.g., People v. Owens, 129 Ill.2d 303, 308, 135 Ill.Dec. 780, 544 N.E.2d 276 (1989); People v. Tate, 366 Ill.Dec. 741, 980 N.E.2d 1100, 1103 (Ill. 2012). (The court sometimes has referred to waiver and sometimes to forfeiture. The distinction makes no difference in this case, and so we will refer to it as a waiver rule.)
Kustok acknowledges these cases, but he argues that the rule nonetheless was not regularly followed by the Illinois Appellate Court. There is some merit to that argument. The line of appellate cases to which Kustok refers begins with People v. Kunze, where the appellate court said in passing that "a claim of ineffective assistance of counsel is better made in proceedings on a petition for post-conviction relief." 193 Ill.App.3d 708, 726, 140 Ill.Dec. 648, 550 N.E.2d 284 (1990). In case after case, the Illinois Appellate Court seized on this language and cited it to dismiss Strickland claims on direct appeal, reasoning that the trial record was rarely sufficient to resolve such claims and that they were therefore better left to postconviction petitions. See Veach I, 401 Ill.Dec. 367, 50 N.E.3d at 96 (collecting cases).
The post-Kunze line of cases culminated in Veach I, where the appellate court reiterated its approach of postponing consideration of Strickland claims to postconviction petitions. Id., 401 Ill.Dec. 367, 50 N.E.3d at 98. But the Illinois Supreme Court quickly pushed back in Veach II and repeated what it had been saying for decades: "in Illinois, a defendant must generally raise a constitutional claim alleging ineffective assistance of counsel on direct review or risk forfeiting the claim." 417 Ill.Dec. 718, 89 N.E.3d at 375.
Although we recognize that the Illinois Appellate Court had been inconsistent in its adherence to the Illinois Supreme Court's waiver rule, there are two reasons why the appellate court's cases do not help Kustok. The first is our precedent. Despite the Illinois Appellate Court's approach to Illinois's waiver rule for Strickland claims, we have found that rule to be adequate. We held as much both before Veach I, see Sturgeon v. Chandler, 552 F.3d 604, 611 (7th Cir. 2009), and after Veach II, see Mata v. Baker, 74 F.4th 480, 486-87 (7th Cir. 2023).
Kustok resists this precedent by arguing that the narrow window of time between Veach I and Veach II was different—in other words, that Veach I so strongly disapproved of resolving Strickland claims on direct appeal that it made the waiver rule inadequate. We note a potential timing issue with this argument: Kustok filed the opening brief in his direct appeal months before the Illinois Appellate Court handed down Veach I, and so it is hard to see how Veach I affected his decision not to include a Strickland claim in that brief. But even putting that aside, we reject Kustok's premise. Veach I did not do anything that dozens of prior appellate decisions had not already done. Even Veach I recognized that it was simply repeating what had long been the appellate court's approach to the waiver rule. See 401 Ill.Dec. 367, 50 N.E.3d at 96 ("The Illinois Appellate Courts have widely followed the Kunze doctrine.").
The second reason for finding that the Illinois Appellate Court's practice did not render the Illinois Supreme Court's rule inadequate is the interplay between the two. The appellate court summarized its approach as follows in Veach I:
[T]he prudent and judicious course for an appellate court dealing with a defendant's claim of ineffective assistance of counsel on direct appeal is almost always to (1) decline to address the issue (while explaining its reason for doing so), (2) affirm the trial court's judgment, and (3) indicate that the defendant may raise
the ineffective-assistance-of-counsel claim in a postconviction petition.Id., 401 Ill.Dec. 367, 50 N.E.3d at 98.
The Kunze line of cases thus set forth the court's practice of refusing to hear Strickland claims on direct appeal and telling appellants to bring such claims in a postconviction petition. But nothing in this practice conflicts with the waiver rule. It does not forbid litigants from raising Strickland claims on direct appeal. We find it telling that dozens of appellants did exactly that in the cases cited by Veach I, as well as in Veach I itself. Those appellants may have expected that the court would simply dismiss their claims, but they chose to raise the claims anyway, lest the court later find that they had waived them—as it sometimes did. See, e.g., People v. Gale, 376 Ill.App.3d 344, 315 Ill.Dec. 171, 876 N.E.2d 171, 178 (2007). Nothing in the appellate court's approach to Strickland claims prevented Kustok from doing the same.
To sum up, decades of consistent precedent from Illinois's highest court required Kustok to raise his Strickland claim on direct appeal. Because he failed to do so, he waived that claim for purposes of his state postconviction petitions and accordingly procedurally defaulted his claim for purposes of this habeas corpus petition.
III
Our analysis does not end there. In the case of state prisoners seeking a federal writ of habeas corpus, the procedural-default doctrine arises from principles of comity and federalism, not from a limit on the jurisdiction of federal courts. See Coleman v. Thompson, 501 U.S. 722, 730, 111 S.Ct. 2546, 115 L.Ed.2d 640 (1991). That rationale leads to two exceptions. A federal court may hear a procedurally defaulted claim if "the prisoner can [(1)] demonstrate cause for the default and actual prejudice as a result of the alleged violation of federal law, or [(2)] demonstrate that failure to consider the claims will result in a fundamental miscarriage of justice." Id. at 750, 111 S.Ct. 2546. Kustok argued only for the second exception, but the district court found no miscarriage of justice. Kustok has not challenged that conclusion; instead, he has turned on appeal to the cause-and-prejudice exception. This skirts closely to forfeiture, if not waiver, depending on how specifically he needed to preserve his argument. But, as we now explain, he has not established actual prejudice, and that is enough to resolve this appeal.
Petitioners satisfy their burden of establishing actual prejudice if they show "not merely a substantial federal claim, such that the errors at trial created a possibility of prejudice, but rather that the constitutional violation worked to [their] actual and substantial disadvantage." Shinn v. Ramirez, 596 U.S. 366, 379-80, 142 S.Ct. 1718, 212 L.Ed.2d 713 (2022) (cleaned up) (emphasis in original). The parties assume that the analysis for actual prejudice largely mirrors the analysis for the underlying Strickland claim—that is, that if Kustok's underlying Strickland claim had merit, then procedurally defaulting that claim substantially disadvantaged him. Under Strickland's familiar two-part test, Kustok would have to show that (1) his counsel's performance was so deficient as to deny him the counsel guaranteed by the Sixth Amendment and (2) that deficient performance denied him the right to a fair trial. 466 U.S. at 687, 104 S.Ct. 2052. Further, because the Illinois Appellate Court rejected Kustok's claim on the merits, we would have to find that its decision "was contrary to[ ] or involved an unreasonable application of" Strickland to grant him relief. See 28 U.S.C. § 2254(d)(1). (We would not defer to the state court on the first part of Strickland, however, because the last state court to reject Kustok's claim did not reach the question whether his counsel was deficient. See Thomas v. Clements, 789 F.3d 760, 767 (7th Cir. 2015).)
To satisfy Strickland's prejudice requirement, Kustok would have to show that, had his attorney tested the pillowcase stain before trial, the likelihood that the jury would not have convicted him was "substantial, not just conceivable." Harrington v. Richter, 562 U.S. 86, 111, 131 S.Ct. 770, 178 L.Ed.2d 624 (2011). This burden is hard to meet, especially when we consider the "overwhelming record support" for the verdict, Strickland, 466 U.S. at 696, 104 S.Ct. 2052, layered on top of the deference we owe to the state court.
We conclude that Kustok has failed to meet that burden, even if we discount Englert's testimony entirely. The trial lasted over three weeks; Englert's testimony spanned two and a half of those days, and most of that time was occupied by Kustok's cross-examination. During the rest of the trial the state introduced, as the Illinois Appellate Court put it, "myriad additional evidence" that supports the jury's verdict. We summarize that evidence below.
Before the shooting. Kustok was having extramarital affairs, which increased in frequency before Jeannie's death and during which he reported unhappiness with his marriage and plans to divorce. When he bought the gun that killed Jeannie, he told the seller that he planned to use it for target practice. Despite Kustok's claim that the couple sometimes brought the gun to bed for protection, their adult daughter, who sometimes slept in bed with Jeannie, was not aware that the couple had a gun. Shortly before her death, Jeannie seemed happy and had scheduled appointments and social events beyond the date of her death, making suicide less plausible.
The shooting itself. The bullet entered Jeannie's face on the left side and traveled downward. Yet Kustok claimed that he had found the gun in Jeannie's right hand, as it rested on her chest. Dr. McElligott testified that it would have been difficult for a right-handed person such as Jeannie to shoot herself in that way, and that any recoil from the gun would have pushed Jeannie's hand away from her chest, not toward it. Further, Jeannie's hands tested negative for gunshot residue, while Kustok's left hand tested positive. According to Dr. McElligott, the lack of soot near the wound indicated that the gun was fired at least six inches away from Jeannie's face, which is rare for self-inflicted gunshot wounds.
After the shooting. Rather than immediately seeking help, Kustok stayed with Jeannie's body for at least 45 minutes after her death. During that time, he stacked bloody pillows on the floor, took the sheets off the bed, and wiped down areas of the bedroom with towels. All of these steps made it harder to deduce what had happened in the room. And when Kustok finally took Jeannie to the hospital, he told a nurse that he had been in the bathroom when the gun went off, but then told a police officer that he had been asleep in bed at the time. He also claimed that his glasses were in the bathroom when Jeannie was shot, but the glasses were stained with blood, and a DNA test showed that the blood contained a major female profile from which Jeannie could not be excluded.
That is not to say that Kustok did not present exculpatory evidence, but that evidence was not particularly powerful. Two of Jeannie's coworkers testified that Jeannie had told her about the couple's gun; Sarah testified that Jeannie often worried about home intrusions; and Noedel testified that the gun may have been fired fewer than six inches from Jeannie's face and that it is not uncommon for that type of gun to discharge accidentally. All told, the state court reasonably concluded that it is not substantially likely that the jury would have returned a different verdict if Kustok's lawyer had introduced evidence about the pillowcase soot-stain at trial. That omission therefore did not infect Kustok's "entire trial with error of constitutional dimensions," United States v. Frady, 456 U.S. 152, 170, 102 S.Ct. 1584, 71 L.Ed.2d 816 (1982), and so he has not shown actual prejudice from procedurally defaulting his claim.
IV
We AFFIRM the district court's dismissal of Kustok's petition for a writ of habeas corpus.