Supreme Court Rules on Habeas Corpus for Attempted Murder Conviction

1 Cite as: 607 U. S. ____ (2026) Per Curiam SUPREME COURT OF THE UNI TED STATES CHRISTOPHER KL EIN, SUPERINTENDEN T, DEPARTMENT OF DETENTION FACILIT IES FOR ANNE ARUNDEL COUNTY, ET AL. v. CHARLES BRANDON MARTIN ON PETITION FOR WRIT OF CERTIORARI TO THE UNITED STATES COURT OF APPEALS FOR THE FOURTH CIRCUIT No. 25–51. Decided January 26, 2026 P ER C URIAM. Under the Antiterrorism and Effective Death Penalty Act of 1996 (AEDPA), strict standards govern the grant of fed- eral habeas relief to prisoners convicted in state court. Faithful application of those standards sometimes puts fed- eral district courts and courts of appeals in the disagreeable position of having to deny relief in cases they would have analyzed differently if they had been in the shoes of the rel- evant state court. But federal court s are dutybound to com- ply with AEDPA, and we have granted summary relief when the lower courts have departed from the role AEDPA assigns. See, e. g., Clark v. Sweeney, 607 U. S. ___ (2025) (per curiam); Dunn v. Reeves, 594 U. S. 731 (2021) (per cu- riam); Mays v. Hines, 592 U. S. 385 (2021) (per curiam); Virginia v. LeBlanc, 582 U. S. 91 (2017) (per curiam); White v. Wheeler, 577 U. S. 73 (2015) (per curiam). This is such a case. Respondent Charles Brandon Martin was convicted in a Maryland court for the attempted mur- der of one of his girlfriends, Jodi Torok. The evidence against him was strong, his conviction was affirmed on ap- peal, and an appellate court held in a state postconviction proceeding that the State’s failure to disclose certain im- peachment evidence that was favorable under Brady v. Maryland, 373 U. S. 83 (1963), did not warrant a new trial because there was no “reasonable probability that the re- sult of [the] trial would have been different” had the

2 KLEIN v. MARTIN Per Curiam evidence been turned over. App. to Pet. for Cert. 115a (App.); see Kyles v. Whitley, 514 U. S. 419, 434 (1995) (the “touchstone of materiality is a ‘reasonable probability’ of a different result”). Because t hat decision neither was “con- trary to” nor “involved an unreasonable application” of “clearly established Federal law,” AEDPA required the de- nial of Martin’s federal habeas petition. 28 U. S. C. §2254(d)(1). Yet the Court of Appeals for the Fourth Circuit affirmed the award of a new trial based on reasoning that departed from what AEDPA prescribes. We therefore grant the State’s petition for a writ of certiorari and reverse. I A Torok, the victim of the attempted murder, had been da- ting Martin for about a year when she told him she was pregnant with a baby she thought was his. Angered by this news, Martin demanded that Torok have an abortion. She refused and informed him that she intended to go to court to compel him to provide child support. She also threatened to tell his “wife or baby mama” about the child. App. 254a. A few weeks later, Martin sent Torok a text message ask- ing: “ ‘Wha t time do u work[?]’ ” Martin v. State, 218 Md. App. 1, 14, 96 A. 3d 765, 773 (2014) (alteration in original). Torok replied, “ ‘I’m off [today],’ ” thus verifying that she would likely be home. Ibid. Later that day, just before 3 p.m., Torok was alone in her apartment in Crofton, Mary- land, speaking on the phone with a close friend, Blair Wolfe, who lived in Pittsburgh. During their call, a man purport- ing to be a salesman knocked on Torok’s front door. Torok hung up to speak to the man but promised to call Wolfe back. When she did not, Wolfe called her several times, but no one answered. Growing increasingly concerned, Wolfe called Torok’s housemate, Jessica Higgs, and asked her to return home to see if Torok was safe. Higgs found Torok unconscious on the floor, bleeding from a gunshot wound to

3 Cite as: 607 U. S. ____ (2026) Per Curiam her head. Torok survived, but her unborn baby did not, and Torok suffered serious permanent injuries. When the police examined the scene of the shooting, they found no sign of forced entry and recovered several items of evidentiary value from the floor near where Torok lay: a shell casing and bullet from a .380-caliber cartridge and a peculiarly modified Gatorade bo ttle. (Photos of this bottle appear in the appendix to this opinion.) The upper part of the bottle was covered with two layers of tape, white medi- cal tape underneath and silver duct tape on top. There was a rectangular hole in the tape over the mouth of the bottle and a jagged hole at the bottom of the bottle. Circumstantial evidence strongly suggested that this ob- ject was likely brought to the apartment and left there by Torok’s assailant. Higgs testified that the bottle was not on the floor when she left for work that morning. Both Torok and Higgs testified that they did not drink Gato rade or keep it at home, and both said that they would not have left a bottle on the floor. The three items found at the crime scene played a part in the State’s case. Federal firearms records showed th at Martin owned a .380-caliber semiautomatic handgun that could have fired the bullet and ejected the casing. One of Martin’s girlfriends testified that she had seen him with a “small” “semiautomatic” gun in the weeks before the shoot- ing. App. 379a. Michael Bradley, the brother of still an- other of Martin’s girlfriends, Maggie McFadden, added that on several occasions he had seen Martin with a “small” “semi-automatic gun.” Id., at 331a–333a, 371a. As for the bottle, the State offered evidence suggesting that it was a homemade silenc er. A police sergeant said that the tape on the mouth of the bottle bore a “rectangular impression” that likely resulted from something being “stuck in there.” Id., at 292a. The bottle was available for examination by the jury, and the State argued that the opening in the tape was shaped like the muzzle of a

4 KLEIN v. MARTIN Per Curiam semiautomatic handgun. The State also noted that the edges of the hole at the botto m of the bottle were bent out- ward, suggesting that the hole had been created by a force emanating from inside the bottle. The police sergeant tes- tified that the bottle resembled homemade silencing devices he had seen in a film a nd in online videos. A detective tes- tified that he found “black soot” inside the bottle, id., at 310a–311a, but that he did not smell burnt or burning ma- rijuana in the bottle, id., at 314a–3 15a, and another witness testified that he found no “signs or evidence of controlled dangerous substances” in the bottle, id., at 309. Based on this evidence and the fact that none of Torok’s neighbors had heard a gunshot at the apartment on the afternoon of the shooting, the State argued that the bottle had been used as a silencer. Testing of the bottle revealed DNA evidence that strongly implicated Martin. A hair was found on the tape on the bottle, and one of the State’s expert witnesses, a f orensic scientist, determined that although the hair could not have come from 99.94 percent of the population of North Amer- ica, she could not rule out Martin as the source of hair. Besides the DNA taken from the hair, trace DNA from at least three persons, including at least one male, was found on the mouth of the bottle. A forensic chemist testified that she could rule out 96 percent of the black pop ulation as po- tential sources of this DNA but could not rule out Martin (who is black).* This DNA evidence linking Martin to the bottle was bol- stered by the testimony of Michael Bradley. Bradley testi- fied that on the afternoon of the shooting, he, his brother Frank, Martin, and an acquaintance named Jerry Burks smoked marijuana at McFadden’s home. At one point, ac- cording to Michael, Frank went upstairs and came back with white “medical tape.” Id., at 337a. Martin and Frank —————— *Nor could she rule out Torok.

5 Cite as: 607 U. S. ____ (2026) Per Curiam then went upstairs together. Frank came downstairs, grabbed a Gatorade bottle, and headed back up the stairs with it. Based on this testimony, the State suggested that Martin helped to make the bottle into the silencing device that the shooter used. This theory was supported by the discovery in the McFadden home of the same kind of white medical tape that was affixed to the bottle. Michael Bradley further testified that Martin acted sus- piciously at the time of and shortly after the shooting. Ac- cording to his testimony, Martin and Burks left the house together before 2 p.m. and were still out when he returned from picking up his niece at around 3 p.m. Martin and Burks eventually reappeared together some time before 6:30 p.m. Martin then handed Frank Bradley a “brown pa- per bag” and told him to “get rid of this.” Id., at 353a. The State suggested that this bag contained the hit weapon, which was never found. Neither did police ever find the.380-caliber pistol that, according to federal firearms rec- ords and witness testimony, Martin owned. In addition to all this evidence, Sheri Carter, another of Martin’s girlfriends, testified for the State. Carter said that Martin had kept at her home a laptop that he said he had obtained from a former employer. She testified that she had seen Martin “looking up gun silencers” on this laptop not long before the shooting. Id., at 377a. Martin, she as- serted, took the laptop from her apartment after the shoot- ing and “got rid of it” “in case [the] apartment got searched,” because he “didn’t want it found there.” Id., at 378a–379a. The trial judge instructed the jury that Martin could be found guilty if he had “ ‘aided ’ ” or “ ‘encouraged’ ” the at - tempted murder and assault with the intent t hat the crime succeed. Id., at 79a. The jury found him guilty as an acces- sory before the fact. The court sentenced him to life in prison, and his conviction and sentence were affirmed on direct appeal. Martin, 218 Md. App. 1, 96 A. 3d 765.

6 KLEIN v. MARTIN Per Curiam B Martin then sought postconviction relief in state court. He argued that the State had violated his right to due pro- cess by failing to disclose a forensic report that had ana- lyzed five computers found at his home. One of these ma- chines, a laptop, had been issued to Martin by a former employer. The report found no evidence that this laptop had been used since 2005, and it noted that a keyword search for words such as “ ‘Handgun,’ ” “ ‘Gatorade,’ ” “ ‘si- lencer,’ ” an d “ ‘Homemade silencer ’ ” had yielded no hits. App. 103a–104a. Martin argued that this computer was the one to which Carter had referred in her testimony and that the report tended to discredit her claim that he had used the laptop at her home to research silencer s. The postconviction court agreed and ordered a new trial, but a unanimous panel of the Maryland Court of Special Appeals reversed on the ground that the report was not ma terial. Based on its re- view of the whole record, the court concluded that even if the report had “totally discredit[ed]” Carter, the other evi- dence linking Martin to the crime was so “strong” tha t there was no “reasonable probability that the result of his trial would have been different.” Id., at 112a, 115a. The State’s high court denied review, Martin v. State, 466 Md. 554, 222 A. 3d 1075 (2020) (table), as did this Court, see Martin v. Maryland, 590 U. S. 973 (2020). C Martin sought habeas relief in federal court, and the Dis- trict Court granted his petition based on the State’s failure to disclose the forensic report. A sharply divided panel of the Fourth Circuit affirmed. The majority acknowledged that the state appellate court had correctly s tated the rule on materiality set out in our decisions and had claimed to apply that rule. Yet the majority concluded that the state court had not actually done what it said it did—had not held

7 Cite as: 607 U. S. ____ (2026) Per Curiam that there was no “reasonable probability” that the disclo- sure of the forensic report would have changed the verdict— but instead had applied the sufficiency-of-the-evidence rule that we condemned in Kyles. App. 21a–22a; see 514 U. S., at 434–435 (a Brady claimant need not show that, “after discounting the inculpatory evidence in light of the undis- closed evidence, there would not have been enough left to convict”). According to the ma jority, the state court “never engaged” with some evidence, “disregarded or miscon- strued” other evidence, and failed to assess in a “nuanced” way the evidence it did discuss. App. 24a–26a. The major- ity then ruled that no fairminded jurist could agree with the state court’s decision. Id., at 27a. Judge Niemeyer dissented, contending that the majority had defied AEDPA’s standard of review. II A As we have noted many times, AEDPA sharply limits fed- eral review of habeas claims raised by state prisoners. A federal court may grant habeas relief on a claim that a state court resolved on the merits only when the state court’s “de- cision” was “contrary to, or involved an unreasonable appli- cation of, clearly established Federal law,” or “ was based on an unreasonable determination of the facts in light of the evidence presented” in state court. 28 U. S. C. §2254(d). These standards require federal courts to give the “benefit of the doubt” to merits decision s issued by the courts of the sovereign States. Woodford v. Visciotti, 537 U. S. 19, 24 (2002) (per curiam). AEDPA review provides an important but limited safeguard: It protects against “ ‘extreme mal- functions’ ” in the state courts’ adjudication of constitutional claims. Harrington v. Richter, 562 U. S. 86, 102 (2011). So in order to obtain federal habeas relief, a state prisoner must “show far more” than “ ‘clear error.’ ” Shinn v. Kayer, 592 U. S. 111, 118 (2020) (per curiam) (quoting LeBlanc,

8 KLEIN v. MARTIN Per Curiam 582 U. S., at 94). The habeas claimant must instead estab- lish that the state court “blunder[ed] so badly that every fairminded jurist would disagree” with the decision. Mays, 592 U. S., a t 392. Only then is a decision “so lacking in jus- tification” that its error preclu des even the “possibility for fairminded” dispute. Richter, 562 U. S., at 103. “If this rule means anything,” we have said, it means that a federal court must “carefully consider all the reasons and evidence supporting the state court’s decision.” Mays, 592 U. S., at 391. That requirement is pivotal because federal courts have “no authority to impose mandatory opinion- writing standards on state courts.” Johnson v. Williams, 568 U. S. 289, 300 (2013). And a state court “need not make detailed findings addressing al l the evidence before it.” Mil- ler-El v. Cockrell, 537 U. S. 322, 347 (2003). Indeed, AEDPA requires deference even if the state court does not discuss the evidence at all. Richter, 562 U. S., at 99. What matters under §2254(d)(1)—the standard relevant here—is whether a decision is contrary to, or involves an unreason- able application of, this Court’s holdings, not whether the state court’s opinion satisfie s the federal court’s opinion- writing standards. B The panel majority contravened these well-settled princi- ples in two ways. First, it grounded its holding that the state appellate court applied the wrong legal rule on its con- clusion that the state court had not actually applied the ma- teriality test that it clearly invoked. Second, it erred i n holding that no fairminded jurist could find the forensic re- port on the computer to be immaterial. 1 The panel majority first erred in holding that the state appellate court failed to apply the right rule for Brady ma- teriality. Undisclosed eviden ce is material if it could

9 Cite as: 607 U. S. ____ (2026) Per Curiam reasonably have “ ‘put the whole case in such a different light as to undermine confidence in the verdict.’ ” S trickler v. Greene, 527 U. S. 263, 290 (1999). But when the evidence could not have reasonably had such an effect, it is not ma- terial, and its erroneous nondisclosure does not justify re- lief. Under this rule, even if undisclosed evidence “entirely discredit[s]” a prosecution witness, the failure to turn over the evidence is not material if “considerable” other evidence “link[s]” the defendant to the crime and the record provides “strong support” that the defendant would have been con- victed anyway. Id., at 292–294. The state appellate court applied these rules instead of a sufficiency-of-the-evidence test. The state court accurately summarized our Brady precedents, correctly stated the gov - erning rule on materiality, and stated unequivocally that its decision was based on that rule. It recounted salient trial evidence, acknowledged wh ere the State’s theory of the case was “attenuated,” and said that the disclosure of the forensic report would likely have eliminated any adverse inference based on the concealment of evidence. App. 114a– 115a, and n. 14. It also assumed that the jury would have “totally discredit[ed]” Carter’s testimony had the State dis- closed the report. Id., at 112a. Yet based on its review of the “ ‘entire record,’ ” t he court found that the evidence “linking” Martin to the crime was so “strong” t hat there was no “reasonable probability that the result of his trial would have been different.” Id., at 109a–110a, 112a, 115a. That standard was legally correct. And except when it was quot- ing our precedent, the state court did not use words like “sufficient,” “insufficient,” “adequate,” or “inadequate” in analyzing the Brady claim. The panel majority nonetheless held that the state appel- late court applied the wrong ru le because that court failed to discuss certain evidence that tended to undermine the State’s case and because its analysis was not sufficiently “nuanced.” App. 24a–26a. That holding was a basic

10 KLEIN v. MARTIN Per Curiam misapplication of AEDPA, which bars federal courts from imposing opinion-writing standards on state courts and de- mands that the relevant state-court decision be given the “benefit of the doubt.” Woodford, 537 U. S., at 24. The ma- jority’s “readiness to attribute error” to the state appellate court despite that court’s correct citation and synthesis of our precedent was both “inconsistent with the presumption that state courts know and follow the law” and “incompati- ble with §2254(d)’s ‘highly deferential standard for evaluat- ing state-court rulings.’ ” Ibid. (quoting Lindh v. Murphy, 521 U. S. 320, 333, n. 7 (1997)). 2 a The panel majority also went astray in holding that every fairminded jurist would find that the undisclosed forensic report about Martin’s laptop was material. On the con- trary, the record contains “strong support” f or the state court’s conclusion that Martin “would have been convicted” even if the forensic report “severely impeached” Carter. Strickler, 527 U. S., at 294. DNA evidence tied Martin tightly to the modified Ga- torade bottled that resemble d a homemade silencer and that had apparently been used in the shooting. Evidence also suggested that Martin was present when the bottle was modified. Michael Bradley test ified that shortly before the shooting Martin was in the room to which Frank Bradley brought both a Gatorade bottle and tape that matched the white medical tape on the mouth of the bottle found at the crime scene. Martin had a strong motive for aiding the attempted murder of Torok: She had refused to get an abortion, had expressed her intent to take him to court to obtain child support, and had threatened to tell his wife about the baby. On the day of the shooting, Martin had texted Torok in

11 Cite as: 607 U. S. ____ (2026) Per Curiam what a reasonable jury could think was an effort to find a time when she would be home. Martin owned the kind of gun that seemed to have been used to shoot Torok, and multiple witnesses testified that they had seen him with such a weapon. Michael Bradley also testified that Martin le ft McFadden’s house not long before the hit and that, upon returning, Martin told F rank Bradley to get rid of a brown paper bag that Martin had been holding. Because the police never found Martin’s handgun or the weapon used in the shooting, a reasonable jury might well have concluded that Martin all owed his gun to be used and then ensured its disposal. In light of all this evidence, a fairminded jurist could eas- ily conclude that the disclosure of the forensic report on the computer would not have “undercut” the relevance or force of these “item[s] of the State’s case,” Kyles, 514 U. S., at 451, or “ ‘put the whole case in such a different light as to under- mine confidence in the verdict,’ ” Strickler, 527 U. S., at 290. b The panel majority’s reasons for concluding otherwise were not consistent with the deference that AEDPA re- quires. First, the majority argued that the disclosure of the report would have bolstered the defense’s primary theory of the case: that the bottle was a device used to smoke mari- juana, not a silencer. But a fairminded jurist could find that theory farfetched. Among other things, the mouth of the bottle was covered in tape that looked like it had nestled the muzzle of a semiautomatic firearm. The bottle did not smell like burnt marijuana and bore no trace of controlled substances. The outward-punched puncture at the bottom of the bottle also looked like a bullet hole, and the defense never explained why anyone w anting to construct a bong would have poked a hole in the bottom of the bottle. The panel majority ventured that Carter’s testimony was the “only evidence connecting Martin to his potential

12 KLEIN v. MARTIN Per Curiam construction of the Gatorade bottle for use as a silencer.” App. 21a. A fairminded jurist could easily disagree based on both the bottle itself and Michael Bradley’s eyewitness testimony about what had occurred upstairs in the McFad- den house shortly before the shooting. The panel majority discounted the strength of the evi- dence against Martin because it thought that strong evi- dence linked McFadden to the commission of the crime. Yet that theory was inconsistent with the undisputed testimony that Torok ended her phone call with Wolfe to speak with a man who was at the door. Nor did any evidence tie McFad- den to the Gatorade bottle or the disposal of the firearm. And even if McFadden were somehow involved in the crime, Martin could have been an accessory anyway. After all, a fairminded jurist could decide that no evidence suggests that McFadden knew about To rok, or had any reason to want her killed, except throug h and because of Martin. See id., at 439a. Last, the panel majority voiced serious doubt about Mi- chael Bradley’s credibility, and the defense certainly had material to use in its effort to convince the jury that he should not be believed. But his testimony about Martin’s role in the creation of the makeshift silencer was supported by the DNA evidence, and the jury was able to assess his credibility firsthand. Based on all the evidence, a fairminded jurist could easily conclude that disclosure of th e forensic report on the laptop would not have made a difference. * * * We grant the State’s petition for a writ of certiorari, re- verse the judgment of the Fourth Circuit, and remand the case for further proceedings co nsistent with this opinion. It is so ordered.

13 Cite as: 607 U. S. ____ (2026) Per Curiam J USTICE J ACKSON would deny the petition for a writ of certiorari.

14 KLEIN v. MAR TIN Appendix to Per Curiam Opinion APPENDIX App. 442a

15 Cite as: 607 U. S. ____ (2026) Appendix to Per Curiam Opinion A pp. 445a

16 KLEIN v. MAR TIN Appendix to Per Curiam Opinion A pp. 444a