Katie Garding v. Montana Department of Corrections

FOR PUBLICATION UNITED STATES COURT OF APPEALS FOR THE NINTH CIRCUIT KATIE GARDING, Nos. 23-35272 23-35327 Petitioner-Appellee / Cross-Appellant, D.C. No. v. 9:20-cv-00105- DLC-KLD MONTANA DEPARTMENT OF CORRECTIONS, OPINION Respondent-Appellant / Cross-Appellee. Appeal from the United States District Court for the District of Montana Dana L. Christensen, District Judge, Presiding Argued and Submitted September 15, 2023 Seattle, Washington Filed June 28, 2024 Before: William A. Fletcher, Ryan D. Nelson, and Daniel P. Collins, Circuit Judges. Opinion by Judge R. Nelson; Dissent by Judge W. Fletcher 2 GARDING V. MONTANA DEP’T OF CORR. SUMMARY * Habeas Corpus On cross-appeals from the district court’s partial denial and partial grant of Katie Garding’s habeas petition, the panel affirmed the district court’s order denying Garding’s claims under Brady v. Maryland and reversed the district court’s grant of Garding’s ineffective-assistance-of-counsel claim. A Montana jury convicted Garding of vehicular homicide while under the influence, failure to stop immediately at the scene of an accident involving an injured person, and driving without a valid driver’s license. The panel rejected Garding’s jurisdictional arguments. The panel explained that the state court’s vacatur of her conviction pursuant to the district court’s habeas decision, and her release from custody, did not moot this case. As the new trial against Garding has not yet begun, this court can provide Montana with relief by reversing the district court’s order. Because Garding was “in custody” under the underlying state conviction when she filed her habeas petition, jurisdiction attached at that time; binding precedent forecloses her argument that AEDPA does not give this court power to hear the case because she is no longer in “custody.” The panel held that the Montana Supreme Court’s determination that Garding’s counsel’s performance was not * This summary constitutes no part of the opinion of the court. It has been prepared by court staff for the convenience of the reader.  GARDING V. MONTANA DEP’T OF CORR. 3 deficient was reasonable. The Montana Supreme Court reasonably held that Garding’s counsel’s decision not to hire an accident reconstruction expert was within the wide range of professionally competent assistance and reasonably concluded that Garding’s claim would require the court to engage in second-guessing with 20/20 hindsight her counsel’s choices, which Strickland v. Washington forbids. The Montana Supreme Court’s determination of the facts supporting its holding was also reasonable. The panel held that the Montana Supreme Court reasonably rejected Garding’s Brady claims, and thus deferred to the Montana Supreme Court as 28 U.S.C. § 2254(d) requires. The Montana Supreme Court reasonably held that the state had not in any way suppressed evidence concerning x-rays of the victim, reasonably held that Garding did not show that the non-disclosure of photos from a different car crash was material, and reasonably concluded that the photos did not establish that Garding was not involved in the accident. Dissenting, Judge W. Fletcher wrote that it is clear from the trial and postconviction record that Garding is innocent, but her innocence is not the legal basis for his agreement with the district court, which held that trial counsel provided ineffective assistance in failing to present evidence from an accident reconstruction expert. Judge Fletcher agreed with the district court because Garding established both deficient performance and prejudice under Strickland and is entitled to relief under AEDPA. 4 GARDING V. MONTANA DEP’T OF CORR. COUNSEL E. Lars Phillps (argued), Crowley Fleck PLLP, Bozeman, Montana, for Petitioner-Appellee. Roy Brown (argued), Assistant Attorney General; Austin Knudsen, Montana Attorney General; Office of the Montana Attorney General, Helena, Montana; Brad Fjeldheim, Assistant Attorney General, Montana Department of Justice, Agency Legal Services, Helena, Montana; for Respondent- Appellant. OPINION R. NELSON, Circuit Judge: We review on cross-appeals the district court’s partial denial and partial grant of Katie Garding’s habeas petition. We hold that the Montana Supreme Court reasonably determined that Garding’s trial counsel was not constitutionally deficient and that her Brady claims lacked merit. We thus affirm the district court’s order denying the Brady claims and reverse its grant of the ineffective assistance claim. I A Early New Year’s Day 2008, a vehicle hit and killed Bronson Parsons. Parsons and his friend, Daniel Barry, were walking westbound on the righthand side of Highway 200 in East Missoula. The two planned to stop by Ole’s Convenience Store and then go to last call at The Reno, a casino and bar across the street.  GARDING V. MONTANA DEP’T OF CORR. 5 At around 1:40 am, a vehicle struck Parsons from behind. Barry stated that he “felt . . . a rush of wind,” and then Parsons was gone. Parsons “stuck to the front of the car,” and then “came to rest off [of it.]” The vehicle, described as a dark-colored SUV or truck, fled. Trooper Novak of the Montana Highway Patrol (MHP) responded. He found Parsons “lying . . . sideways on his back.” He investigated, including by collecting evidence and interviewing Barry. He did not find any of the striking vehicle’s debris. Later that day, two other MHP officers—Troopers Hader and Wolfe—stopped Garding’s vehicle, a dark Chevrolet S- 10 Blazer. At the time, they were looking for a car with heavy front-end damage. Trooper Hader testified that Garding’s windshield was visibly cracked. After stopping Garding, the officers saw that her car did not have full-front- end damage, so the officers let her go. Later that week, however, while examining Parsons’s body, Trooper Hader realized Parsons’s injuries did not suggest a “full-frontal impact.” The State then changed its investigation to look for a minimally damaged car. Around that time, MHP received a tip about Garding. A man reported a dark Blazer with front-end damage. MHP ran a registration check, identifying it as Garding’s car. Trooper Novak contacted Garding’s father, whom he knew personally, but did not speak with Garding. The case went cold for about a year. Then an inmate in Missoula, Teuray Cornell, claimed to have information about the crash. Trooper Hader met with Cornell, who thought Garding was involved. He divulged that he had “taped up” Garding’s bumper’s turn indicator light right after the crash, suggesting that it had recently been damaged. 6 GARDING V. MONTANA DEP’T OF CORR. Trooper Hader then interviewed Garding. Based on further investigation, Garding was charged with Vehicular Homicide While Under the Influence or Negligent Homicide, Failure to Stop Immediately at Accident Scene, Tampering With or Fabricating Physical Evidence, and Driving Without a Valid Drivers License based on a “totality of the evidence.” B Garding’s criminal trial was in June 2011. A public defender represented Garding. Garding maintained her innocence. What matters for this appeal is the State’s crash theory, or how Garding’s car caused Parsons’s injuries. Garding claims that her counsel was not able to effectively push back against the State’s theory because her counsel did not use an accident reconstruction expert and that the State kept evidence from her. Several State witnesses testified about the crash, including the three investigating MHP Troopers— Strauch, Hader, and Novak—and expert witness Dr. Gary Dale, who medically examined Parsons’s body. We discuss the salient parts of the trial. 1 Each of the three Troopers testified about the crash, including how Garding’s vehicle was involved. Trooper Strauch testified about how the crash might have happened. He used a method called “total station,” relying on “an electronic distance measuring instrument,” to help him gauge how far Parsons might have traveled from impact. He estimated this to be about ninety feet. That said, he could not identify the location where Parsons had been hit and could not estimate the vehicle’s speed. He said that tire  GARDING V. MONTANA DEP’T OF CORR. 7 marks might have helped him estimate, but didn’t recall if any were found. Trooper Hader testified about how the scene pointed to Garding’s vehicle. He thought Parsons’s injuries, which differed from full-frontal impact injuries, fit the Blazer’s minimal damage. He reasoned that, if Parsons’s full body had struck the vehicle, there should have been some greater evidence of impact, such as broken ribs or more bruising, but that there was not. Trooper Hader thought that the crash was likely a “swerving-type impact,” consistent with minimal damage. He also thought Garding’s big, steel, aftermarket bumper could explain the minimal front-end damage. Trooper Novak testified about his interview with Barry. He stated that Barry described seeing Parsons “on the hood . . . by the windshield” after he was struck. He stated that Barry also described Parsons being “carried” by the car and falling onto the road. The Troopers did not provide a comprehensive theory of how the crash happened. None of them claimed to be an expert in accident reconstruction, nor were they offered as experts. 2 Dr. Dale’s autopsy identified the cause of death as blunt force head injuries, resulting from when Parsons hit the asphalt. He testified that, in his opinion, Parsons’ other upper body injuries resulted from impact with the asphalt as well. Parsons also suffered faint bruising and crushed calf muscles, which Dr. Dale thought Garding’s bumper could have caused as well. That said, he admitted that any bumper of a similar height could have caused Parsons’s injuries. 8 GARDING V. MONTANA DEP’T OF CORR. 3 Garding’s counsel pushed back against the State’s crash testimony. She called Dr. Thomas Bennett, an expert witness in forensic pathology to rebut the State’s theory. During voir dire, Dr. Bennett clarified that he did not “do accident reconstruction,” but “usually work[ed] with other accident reconstructionists” in similar types of cases. In his opinion, the bruises on the back of Parsons’s legs “would not [have been] caused by a bumper like” Garding’s but were “more consistent with a more rounded bumper.” He thus concluded that Garding’s “bumper could not have caused [Parsons’s] injuries.” Garding’s counsel extensively critiqued the State’s theory of the crash during closing argument. She noted the inconsistencies with the State’s theory presented during Trooper Novak’s testimony and argued that it was “not possible” that “Parsons [was] struck from behind going backwards,” but “g[ot] forward 150 feet.” She also mentioned that “[Garding’s] vehicle d[id] not have heavy front-end damage.” The jury found Garding guilty on June 10, 2011. Garding was sentenced to forty years in prison. She was released on parole on February 3, 2022. C Garding moved for habeas relief in state court. She alleged ineffective assistance, Brady violations, and newly discovered evidence. We discuss the evidence supporting those claims still on appeal—the ineffective assistance and Brady claims. As for the ineffective assistance claim, Garding’s counsel represented that she had been “ineffective.” On her Brady claims, Garding argued that the  GARDING V. MONTANA DEP’T OF CORR. 9 State did not disclose exculpatory evidence: (1) photographs of a 2005 hit-and-run collision and (2) x-rays of Parsons’s lower legs. In 2018, the state court granted the State’s motion for partial summary judgment on Garding’s Brady claim related to the x-rays and her newly discovered evidence claim. The court scheduled a hearing for the ineffective assistance claim. 1 The hearing lasted two days. The court listened to evidence on whether Garding’s counsel was ineffective for not securing an accident reconstruction expert or conducting a reasonable investigation. Several witnesses testified, including Garding’s counsel, two concurring attorney witnesses, and accident reconstruction experts. Garding’s counsel claimed that she was ineffective because she did not take “necessary steps” to consult and secure an accident reconstruction expert. She claimed to be isolated, overwhelmed, and without adequate help. That said, she admitted that she had used such an expert in a similar case and knew they could offer “valuable insight.” She also admitted that she had help, including co-counsel and investigators. Two expert attorney witnesses concurred that she was ineffective. That said, both acknowledged that defense counsel can prefer cross-examination over expert testimony, and that this can be an effective strategy. Accident reconstruction experts also testified. One claimed that he could “[a]bsolutely” “refute the . . . theories presented at trial.” But he admitted that other data, which was unavailable, would be needed for a “precise 10 GARDING V. MONTANA DEP’T OF CORR. reconstruction.” Another admitted that the state usually provides a “counter expert” who typically reaches different conclusions. The State offered a rebuttal accident reconstruction expert, Trooper Smart. He explained that there usually is not enough data to do a “full accident reconstruction” when the car flees the scene or the speed or impact point are unknown. He said that Garding’s experts used “[g]arbage data,” including an illogical assumed speed. 2 In 2019, the state court denied all Garding’s habeas claims. The state court held that Garding’s counsel’s trial performance was not constitutionally deficient because, among other things, she “effectively cross examined the State’s witnesses.” The court rejected her contradictory testimony, characterizing it as “self-serving” and “not credible.” Instead, the court thought Garding’s counsel’s choice was strategy, not error. The state court found that Garding’s counsel made a strategic decision to not use an accident reconstructionist and that this decision was “reasonable.” The court based this conclusion on several considerations. For example, the state court found a lack of evidence to precisely determine the speed of the vehicle. So, according to the state court, Garding’s experts relied on faulty assumptions. Concluding that not enough data justified use of an accident reconstruction expert, the court found that Garding’s counsel made a “calculated decision” to rely instead on cross- examination. The court also rejected the Brady claim. It held that there was not enough information about the crash photos to assess  GARDING V. MONTANA DEP’T OF CORR. 11 their “relevancy” or “exculpatory value” and that they were “not material.” 3 The Montana Supreme Court affirmed. Garding v. State, 466 P.3d 501 (Mont. 2020). It first analyzed Garding’s ineffective assistance of counsel claims under the first part of the two-prong test in Strickland v. Washington, 466 U.S. 668 (1984)—whether counsel’s performance was deficient. Garding, 466 P.3d at 506–09. The court found that Garding’s counsel’s performance was adequate. First, it rejected Garding’s counsel’s “self- proclaimed inadequacies,” as those “do not hold great persuasive value with this Court.” Id. at 507. It then determined that Garding’s counsel provided an “extensive and strong defense.” Id. She “countered or sought to undermine virtually every evidentiary contention introduced by the State, and the jury was left with the unenviable task of making numerous credibility determinations in order to resolve evidentiary conflicts necessary to reach a verdict.” Id. The court identified several ways Garding’s counsel performed adequately. For example, Garding’s counsel retained a forensic pathologist, Dr. Bennett, to counter the State’s only expert testimony. He testified extensively that Garding’s bumper could not have caused Parsons’ injuries. Id. Garding’s counsel also “elicited multiple concessions” from the State’s expert, Dr. Dale, that “any other vehicle with a bumper the same height as Garding’s could have caused Parsons’ injuries.” Id. The court also squarely rejected Garding’s argument that failure to hire an accident reconstruction expert was 12 GARDING V. MONTANA DEP’T OF CORR. deficient. Id. at 508. “Notably,” it pointed out, “the State did not pursue [one] either.” Id. The court also stated that Garding’s counsel “presented a strong defense.” Id. To otherwise find for Garding, the court concluded, it “would [be] require[d] . . . to engage in second guessing with ‘20/20 hindsight’ of the choices made by her counsel,” even though Strickland does not allow this analysis. Id. The court thus affirmed the denial of habeas relief without reaching Strickland’s second prong. Id. The court also affirmed the denial of Garding’s Brady claims. As to the x-rays, their existence was disclosed, the state’s expert referenced them, and Garding’s expert noted that reference. Given that Garding’s “expert referenced” the x-ray result and her counsel “examined witnesses based on it,” the state court held that “Garding was not only aware of the evidence . . . she . . . actively used it.” Id. at 510. Thus, no Brady violation could be found. Id. As to the crash photos, the court disagreed that the prosecution suppressed them, given that they were independently obtained by the expert after his testimony and “placed within his own file.” Id. at 510–11. Thus, “it is unlikely Garding could have used the photos to directly impeach Dr. Dale at all.” Id. at 511. Moreover, even if he had, “the many distinctives between the photographs and this case” would have likely made them inadmissible. Id. As a result, they were neither “suppressed, material nor exculpatory.” Id. Garding unsuccessfully sought review in the United States Supreme Court. Garding v. Montana, 141 S. Ct. 1076 (2021). D Garding next sought federal habeas relief. She argued that the Montana Supreme Court unreasonably applied  GARDING V. MONTANA DEP’T OF CORR. 13 Strickland and that habeas relief was therefore available under 28 U.S.C. § 2254(d)(1). The district court partially granted the habeas petition and partially denied it. Garding v. Montana Dep’t of Corr., No. CV 20-105-M-DLC, 2023 WL 3086883 (D. Mont. Mar. 27, 2023). On the Strickland claim, the district court held that there was ineffective assistance. Id. at *10. It claimed that “there [was] no scenario under which” Garding’s counsel could have thought an accident reconstruction expert “could have inculpated her client.” Id. at *9. Thus, her failure to use such an expert was constitutionally deficient, failing to satisfy Strickland’s objectively reasonable requirement. Id. at *10. The district court denied the Brady claims. Id. at *17– 19. It determined that “[t]he Montana Supreme Court reasonably rejected [them],” and so the court “must afford deference under . . . § 2254(d).” Id. at *17. Garding filed an appeal in 2023. Montana timely cross- appealed. II Garding raises two jurisdictional issues, which we address from the start. See, e.g., Great S. Fire Proof Hotel Co. v. Jones, 177 U.S. 449, 453 (1900) (“On every writ of error or appeal, the first and fundamental question is that of jurisdiction.”). First, Garding argues that this appeal is moot because we cannot reinstate her criminal conviction, and so cannot give relief to the State. Second, she argues we do not have statutory jurisdiction under AEDPA. We reject both arguments. 14 GARDING V. MONTANA DEP’T OF CORR. A We assess mootness by whether there is “a present controversy” for which we can grant relief. All. for the Wild Rockies v. Petrick, 68 F.4th 475, 485–86 (9th Cir. 2023) (citation omitted). The party claiming mootness has a heavy burden of proof. Forest Guardians v. Johanns, 450 F.3d 455, 461 (9th Cir. 2006). And the remedy need not be “fully satisfactory.” Calderon v. Moore, 518 U.S. 149, 150 (1996) (per curiam). If some relief can be granted, the case is not moot. Forest Guardians, 450 F.3d at 461. Garding argues that the state trial court’s release of her from custody and the vacatur of her conviction deprives this court of jurisdiction over her habeas appeal. This is because, Garding claims, this court “has [no] power to alter [the] state court order.” Thus, Garding claims we can grant no effective relief, and the case is moot. Garding relies on Brown v. Vanihel, 7 F.4th 666 (7th Cir. 2021)—an out of circuit case. There, a federal district court granted Brown habeas relief. Brown, 7 F.4th at 668. The State then asked to vacate Brown’s conviction and retry him. Id. The state court vacated the conviction. Id. at 668–69. Brown asked to dismiss the appeal, arguing that the vacatur order mooted the State’s appeal. Id. at 669. The Seventh Circuit agreed, holding that the vacatur of the conviction took away its power to hear the case because the State’s appeal concerned a nonexistent judgment. Id. Thus, it dismissed the case as moot. Id. The problem is that Brown is contrary to the Supreme Court’s decision in Moore, and thus wrong. In Moore, the Court held that a factually similar habeas appeal was not moot. 518 U.S. at 149–50. The petitioner challenged his conviction, and the district court granted relief, directing that  GARDING V. MONTANA DEP’T OF CORR. 15 he be released, or that the State have a new trial. Id. at 149. “The State . . . set Moore for retrial.” Id. at 150. We held this mooted the case. Id. The Supreme Court reversed, holding that the new trial order did not amount to a situation in which, “by virtue of an intervening event, a court of appeals cannot grant ‘any effectual relief whatever.’” Id. Although “the administrative machinery necessary for a new trial ha[d] been set in motion, that trial ha[d] not yet even begun, let alone reached a point where the court could no longer award any relief in the State’s favor.” Id. At a minimum, “a decision in the State’s favor would release it from the burden of [a] new trial.” Id. Thus, at least some relief was available. Id. (citing Mills v. Green, 159 U.S. 651, 653 (1895)). Here, just as in Moore, the state court judgment was set aside only because of the district court’s habeas decision. This started a process for a new trial in state court. True, the district court below did not set aside the judgment directly. But that does not justify ignoring Moore. The State here moved for a new trial in state court only under compulsion of the habeas order, which otherwise barred retrial. That was when the state court vacated the conviction and set a new trial. Indeed, the state court order vacated conviction “[p]ursuant to the Order in the United States District Court for the District of Montana, Missoula Division, cause number CV-20-105-M-DLC and based upon the State’s Motion to Renew Proceedings filed in compliance with that order.” (emphasis added). Reversal of the district court’s order would remove the current federal court impediment to any state court reinstatement of the judgment and cancellation of the new trial. Reversal would, as in Moore, “release [the state of] the burden of the new trial itself.” 518 U.S. at 150. 16 GARDING V. MONTANA DEP’T OF CORR. Brown conflicts with Moore and did not consider Moore. In both Brown and Moore, the underlying conviction was vacated. Moore, 518 U.S. at 149; Brown, 7 F.4th at 668–69. Brown suggests that this was enough to take away our power to hear the case, because “[i]f the state court vacates the underlying judgment, there is usually nothing more for the federal courts to do.” Id. at 669. But Moore held the opposite; federal courts can relieve the state of the burden of a new trial. 518 U.S. at 150. Brown also conflicts with the Supreme Court’s holding in Eagles v. U.S. ex rel. Samuels, 329 U.S. 304, 307–08 (1946). There, the Court held that even where “the writ has been granted and the prisoner released,” an appellate court can still “affect the litigants in the case before it” because “[r]eversal undoes what the habeas corpus court did and makes lawful a resumption of the custody.” Eagles, 329 U.S. at 307–08. Brown sought to distinguish Eagles because the latter did not involve a vacatur of an underlying conviction. 7 F.4th at 672. But that is a difference without a distinction. Garding was formerly “in custody” as a state parolee before the district court’s grant of habeas relief. See Thornton v. Brown, 757 F.3d 834, 841 (9th Cir. 2013) (“A state parolee is ‘in custody’ for purposes of the federal habeas statute.”). Thus, just as in Eagles, a reversal would allow a “resumption of the custody” that had been challenged in habeas corpus. The state court’s vacatur of Garding’s conviction did not moot this case. The new trial against Garding has not yet begun, and by reversing the district court’s order, we can provide Montana with relief. B Garding also argues that AEDPA does not give us power to hear this case because she is no longer in “custody.”  GARDING V. MONTANA DEP’T OF CORR. 17 Binding precedent, however, forecloses this statutory interpretation. The statute asks whether the petitioner was “in custody” under the “judgment of a State court” when the petition was filed. Stow v. Murashige, 389 F.3d 880, 885 (9th Cir. 2004) (quotations omitted); see Maleng v. Cook, 490 U.S. 488, 490–91 (1989) (“We have interpreted the statutory language as requiring that the habeas petitioner be ‘in custody’ under the conviction or sentence under attack at the time his petition is filed.”); Carafas v. LaVallee, 391 U.S. at 234, 238 (1968) (“The federal habeas corpus statute requires that the applicant must be ‘in custody’ when the application for habeas corpus is filed.”). All agree that Garding was “in custody” under the underlying state conviction when she filed her habeas petition. Jurisdiction attached at that time. III We turn to the merits. We review a district court’s grant or denial of a habeas petition de novo. Earp v. Davis, 881 F.3d 1135, 1142 (9th Cir. 2018); Wilkinson v. Gingrich, 806 F.3d 511, 515 (9th Cir. 2015). We apply “AEDPA’s standard of review to the ‘last reasoned state-court decision.’” Noguera v. Davis, 5 F.4th 1020, 1034 (9th Cir. 2021) (quoting Martinez v. Cate, 903 F.3d 982, 991 (9th Cir. 2018)). That standard is “highly deferential.” Davis v. Ayala, 576 U.S. 257, 269 (2015). As relevant here, by AEDPA’s terms, we can reverse a state court decision only if the “decision . . . was contrary to, or involved an unreasonable application of, clearly established” Supreme Court precedent. 28 U.S.C. § 2254(d)(1). Guided by these principles, we defer to the Montana Supreme Court’s application of Strickland and Brady. 18 GARDING V. MONTANA DEP’T OF CORR. A In reviewing an ineffective assistance claim, we ask “whether the state court’s application of the Strickland standard was unreasonable.” Harrington v. Richter, 562 U.S. 86, 101 (2011). “This is different from asking whether defense counsel’s performance fell below Strickland’s standard.” Id. We first evaluate whether the Montana Supreme Court reasonably applied Strickland when it held that Garding’s counsel’s performance was not deficient. Strickland, 466 U.S. at 687. Because we hold that it did, we do not reach the second part of the Strickland test. Id. at 697 (“[T]here is no reason for a court deciding an ineffective assistance claim . . . to address both components of the inquiry if the defendant makes an insufficient showing on one.”). The Montana Supreme Court’s determination that Garding’s counsel’s performance was not deficient was reasonable. First, the state court reasonably held that Garding’s counsel’s decision not to hire an expert was within the “wide range of professionally competent assistance.” Garding, 466 P.3d at 508 (citing Whitlow v. State, 183 P.3d 861, 866 (Mont. 2008) (quoting Strickland, 466 U.S. at 690)). It also reasonably held that Garding’s counsel’s defense was strong, and that she effectively countered the State’s case. Id. at 507. It further reasonably concluded that Garding’s claim would require “the Court to engage in second guessing with ‘20/20 hindsight’” her counsel’s choices, which Strickland forbids. Id. at 508. The district court held that because no reasonable defense attorney would have failed to use an accident reconstruction expert here, the Montana Supreme Court unreasonably held that Garding’s counsel acted within the  GARDING V. MONTANA DEP’T OF CORR. 19 range of professional competence. See Garding, 2023 WL 3086883, at *8–10. We disagree. The Montana Supreme Court reasonably applied Strickland to the facts as found by the Montana Supreme Court. These facts included that at trial, the State elected not to present any expert. Garding, 466 P.3d at 508 (“Notably, the State did not pursue an accident reconstruction [expert] either.”) And, the state high court concluded, Garding’s counsel “countered or sought to undermine virtually every evidentiary contention introduced by the State.” Id. at 507. The state trial court also rejected Garding’s counsel’s representations as “self-serving statements” contradicted by other testimony. And then, holding that counsel’s testimony was not credible, the state trial court reviewed the total record, and concluded that counsel made a “strategic decision” not to use an accident reconstruction expert. The state trial court’s analysis is reasonable under our highly deferential review. The dissent faults the Montana Supreme Court for relying too much on Garding’s counsel’s representations while not discussing Garding’s post-conviction accident reconstruction evidence. Dissent at 41. But the dissent’s analysis is flawed. The postconviction accident reconstruction experts’ evidence was considered by the trial court but only related to the prejudice issue, not the deficiency issue. The Montana Supreme Court’s decision to deny Garding’s claims because her counsel’s performance was not deficient was reasonable. Thus, the Montana Supreme Court did not separately address the prejudice issue. See Strickland, 466 U.S. at 697. The dissent collapses these two inquiries by concluding that Garding’s counsel’s performance was deficient because an accident reconstruction expert’s “testimony would have been 20 GARDING V. MONTANA DEP’T OF CORR. devastating to the State’s case.” Dissent at 40. Put differently, the dissent argues that Garding’s counsel was necessarily deficient because Garding was prejudiced. The dissent’s argument violates Strickland’s very dictates that “[a] fair assessment of attorney performance requires every effort be made to eliminate the distorting effects of hindsight.” See Strickland, 466 U.S. at 689. Moreover, the Montana Supreme Court reasonably concluded that Garding’s counsel mounted a strong defense. And, as part of this defense, Garding’s counsel relied on the State’s disjointed presentation to cast doubt on the State’s case. The Montana Supreme Court’s conclusion that this was a strategic decision over using an accident reconstruction expert was reasonable, especially given that Garding’s counsel had used an accident reconstruction expert before. This conclusion follows Richter. There, the petitioner claimed his counsel was constitutionally deficient because he failed to secure expert testimony on blood evidence. Richter, 562 U.S. at 96. The Supreme Court disagreed, holding that “[i]t was at least arguable that a reasonable attorney could decide to forgo [the] inquiry.” Id. at 106. This is because “making a central issue out of blood evidence would have increased the likelihood of the prosecution’s producing its own evidence on the blood pool’s origins and compositions,” and “there was a serious risk that expert evidence could destroy Richter’s case.” Id. at 108. The state courts reasonably concluded that a similar risk was present here. As the state trial court noted, there was “a counter-analysis” presented at the post-conviction hearing that argued for a conclusion consistent with Garding’s guilt.  GARDING V. MONTANA DEP’T OF CORR. 21 The State’s expert presented a crash theory that tracked the minimal injuries to Parsons, minimal damage to Garding’s vehicle, and reflected the eyewitness testimony. The dissent objects to the State’s use of the Troopers’ testimony about their investigation of the accident, claiming that the Montana Supreme Court was wrong in stating that the State did not pursue an accident reconstruction. Dissent at 41. But the Troopers were never offered or formally qualified as experts, and the Montana Supreme Court reasonably concluded that whatever limited opinions they offered did not amount to the sort of “accident reconstruction” that Garding now contends that her counsel should have done. See Garding, 466 P.2d at 504. This reasonable finding is one that we may not second-guess on AEDPA review. See 28 U.S.C. § 2254(d). The dissent’s citation to Duncan v. Ornoski, 528 F.3d 1222, 1235 (9th Cir. 2008), Dissent at 41–42, does not change this. There, we simply said it was possible that when the prosecution puts up an expert witness, a defense counsel’s failure to put up their own rebuttal expert may constitute deficient performance. Ornoski, 528 F.3d at 1235. But the State did not offer expert testimony. The dissent’s reinterpretation of the facts to suggest they did is inappropriate under AEDPA review and undermines Richter’s holding that it is sometimes strategic for defense counsel not to pursue expert testimony. See 562 U.S. at 106. The district court wrongly held that the Montana Supreme Court unreasonably applied Strickland. Garding, 2023 WL 3086883, at *8. The Montana Supreme Court’s decision was a reasonable application of Strickland. See § 2254(d). Likewise, its determination of the facts supporting this holding was also reasonable. See id. The 22 GARDING V. MONTANA DEP’T OF CORR. district court misapplied the law and misconstrued the record in holding otherwise. 1 B Garding claims that her constitutional rights were violated because the prosecution failed to disclose evidence. See Brady v. Maryland, 373 U.S. 83, 87 (1963). She has two theories. First, the prosecution did not disclose x-rays of Parsons’s leg. Second, they did not disclose unrelated crash scene pictures. Garding claims that if she had had either, she might have been found not guilty. “A Brady violation occurs when the government fails to disclose evidence materially favorable to the accused.” Youngblood v. W. Virginia, 547 U.S. 867, 869 (2006). Evidence is “material only if there is a reasonable probability that, had [it] been disclosed to the defense, the result of the proceeding would have been different. A ‘reasonable probability’ is [one] sufficient to undermine confidence in the outcome.” Amado v. Gonzalez, 758 F.3d 1119, 1139 (9th Cir. 2014) (quoting United States v. Bagley, 473 U.S. 667, 682 (1985)). 1 The dissent concludes that, in light of the new evidence developed in the state habeas proceedings, Garding’s showing of prejudice is strong enough to conclude that she is “innocent.” Dissent at 43. But we cannot reach the issue of prejudice unless we are able first to conclude, applying the deference required by AEDPA, that the state court unreasonably applied the “highly deferential” Strickland standard for assessing “counsel’s performance.” Cullen v. Pinholster, 563 U.S. 170, 190 (2011). The dissent fails to apply this “doubly deferential” standard of review. Id. It also relies extensively on “the distorting effects of hindsight,” rather than assessing counsel’s performance “from counsel’s perspective at the time.” Strickland, 466 U.S. at 689. We thus reiterate that we cannot and do not reach the issue of prejudice.  GARDING V. MONTANA DEP’T OF CORR. 23 We agree with the district court that the Montana Supreme Court reasonably rejected Garding’s Brady claims. We thus defer to the Montana Supreme Court’s conclusions as § 2254(d) requires. 1 Garding alleged that the State violated Brady by failing to turn over x-rays of the victim. Dr. Dale, the State’s expert and the medical examiner who performed the autopsy, created an x-ray of Parsons’s injuries. The x-ray was never provided to Garding’s counsel. That said, Garding’s counsel received a summary of the x-ray, which she used effectively at trial. Garding relies on her expert, Dr. Bennett, to show that the x-rays were “impeaching and exculpatory.” Dr. Bennett explained that the x-rays showed a “slight hairline fracture,” which would have “cast[] doubt upon and undermine[d] the State’s case.” Dr. Bennett concluded that the x-ray confirms that Garding’s Blazer was not involved because its custom bumper would have caused more damage to Parsons’s leg. Similarly, Garding argues that the x-ray would have undermined Dr. Dale’s testimony that the injuries pointed to the Blazer. The Montana Supreme Court reasonably concluded that Garding’s theory does not show a Brady violation. Brady requires the disclosure of exculpatory or impeaching evidence that, “if disclosed and used effectively, . . . may make the difference between conviction and acquittal.” Bagley, 473 U.S. at 676. As the Montana Supreme Court noted, the existence of the x-rays was disclosed, a summary of what was shown by the x-rays was discussed by both experts, and defense counsel “examined witnesses based on it.” Garding, 466 P.3d at 510. The Montana Supreme Court 24 GARDING V. MONTANA DEP’T OF CORR. reasonably held that the state had not “in any way suppressed the evidence.” Id. 2 Garding also argues that the State violated Brady by not disclosing exculpatory pictures from a different car crash. Three days after he testified, Dr. Dale discovered photos of a victim and vehicle connected to a different crash. Garding, 466 P.3d at 510. These showed similar injuries to the victim, but different damage to the vehicle. Dr. Dale explained that he thought they might be helpful if he was called as a rebuttal witness. But he never was. Id. He did not use the photos to form his testimony. And after reviewing them, he did not change his mind. Id. The Montana Supreme Court reasonably held that Garding did not show that the non-disclosure of these photos was material. The photos were from a crash with “many distinctives” from this case—differences that made the Montana Supreme Court question the likelihood of the photos’ admissibility. Id. at 511. More importantly, the Montana Supreme Court reasonably concluded that the photos did “not establish that Garding was not involved in the accident.” Id. IV The Montana Supreme Court was objectively reasonable in determining that Garding failed to establish an ineffective assistance of counsel claim under Strickland or any Brady violations. AFFIRMED IN PART AND REVERSED IN PART.  GARDING V. MONTANA DEP’T OF CORR. 25 W. Fletcher, J., dissenting. The district court granted petitioner Katie Garding’s federal habeas petition, holding that her trial counsel provided ineffective assistance in failing to present evidence from an accident reconstruction expert. The majority concludes that the district court erred. I disagree and would affirm the district court. This case is a miscarriage of justice. It is clear from the trial and postconviction record that Garding is innocent. I. Background On January 1, 2008, at about 1:40 a.m., Bronson Parsons was walking beside his friend Daniel Barry on the side of Highway 200 in East Missoula, Montana. A vehicle struck Parsons from behind. Barry told state troopers who arrived on the scene that the vehicle had been a rounded, dark colored SUV or truck. He testified at trial that the vehicle had been traveling “extremely fast,” “too fast” for someone to survive. Barry recounted, “[A]ll of a sudden [Parsons] was gone. I felt like a rush of wind.” He told a trooper who arrived at the scene that the vehicle had been traveling at about 60 miles per hour, and that Parsons had been “on the hood and up by the windshield.” He testified that when the vehicle slowed down, Parsons slid off the hood onto the ground. Another eyewitness, Deborah Baylor, was driving in the opposite direction on Highway 200 when Parsons was hit. Baylor testified that she saw a dark colored vehicle hit Parsons. “I think they were going regular speed.” The vehicle, a “little bit smaller” than a Cadillac Escalade, had “rounded edges.” “[I]t was so fast. . . . I saw something get 26 GARDING V. MONTANA DEP’T OF CORR. hit and—and then I hear a—it’s like a pop, like a quick bang.” When State Trooper Andrew Novak arrived at the scene, Parsons was on the ground and “agonally breathing.” Parsons had blood coming “from his stomach area” and “from his head, the back of his head, and his mouth.” Based on Barry’s description of what had happened, Novak believed that the striking vehicle would have sustained “heavy, front-end damage.” Parsons was taken to the hospital and was later pronounced dead. Windshield glass was recovered from Parsons’ clothing. Later that morning, Montana Highway Patrol troopers were on the lookout for vehicles with broken windshields. About twelve hours after Parsons was hit, Trooper Richard Hader stopped Garding in East Missoula because she had a cracked windshield. She was quickly released because the crack in her windshield was old and there was no observable damage to her vehicle. About a year later, after the case had gone cold, a jail inmate named Teuray Cornell contacted Trooper Hader, saying he had information about who had hit Parsons. Cornell made it clear that in exchange for his testimony he wanted to get out of jail. Cornell’s call rekindled interest in Garding. Garding was ultimately charged with having killed Parsons. According to the Montana Innocence Project, Garding was offered an extremely favorable deal under which, in return for a guilty plea, she would receive a suspended sentence and no prison time. Montana Innocence Project, Katie Garding, https://mtinnocenceproject.org/katie- garding-2/ [https://perma.cc/NY4Y-BG5P]. Garding, who has consistently said she was innocent, rejected the deal.  GARDING V. MONTANA DEP’T OF CORR. 27 The case was tried to a jury in June 2011. Garding was represented by Jennifer Streano, a Montana Public Defender. Streano had four-and-half years of criminal defense experience and had previously been lead counsel in only one homicide case. The jury found Garding guilty of vehicular homicide, failure to stop, and driving without a license. She was sentenced to a term of 30 years for the homicide, a consecutive term of 10 years for failure to stop, and a concurrent term of 6 months for driving without a license. The Montana Supreme Court affirmed. State v. Garding, 315 P.3d 912 (Mont. 2013). Montana trial court denied Garding’s petition for postconviction relief, and the Montana Supreme Court again affirmed. Garding v. State, 466 P.3d 501 (Mont. 2020). Garding timely filed a federal habeas petition under 28 U.S.C. § 2254. After giving the deference required by the Anti-Terrorism and Effective Death Penalty Act (“AEDPA”), the district court granted habeas relief, holding that in denying Garding’s ineffective assistance of counsel claim, the Montana Supreme Court had unreasonably applied Strickland v. Washington, 466 U.S. 668 (1984). A. Trial Court Evidence The State’s case against Garding relied heavily on testimony from James Bordeaux, Garding’s ex-boyfriend. Bordeaux had been one of two passengers in Garding’s vehicle on the night Parsons was killed. In exchange for his testimony, Bordeaux obtained a favorable plea deal on an unrelated burglary charge. The State had indicated that it intended to pursue a persistent felony offender designation against Bordeaux, exposing him to a potential sentence of up 28 GARDING V. MONTANA DEP’T OF CORR. to 100 years. In exchange for Bordeaux’s testimony, the State agreed to recommend a five-year suspended sentence. Bordeaux testified at trial that on the night of the accident, Garding had been driving and that he had been in the front passenger seat. He testified that after he turned to argue with Paul McFarling, who was in the back seat, about McFarling’s handgun, he felt an impact and saw “[a] person flying through the air.” Bordeaux was asked, “How do you know it’s a person?” He answered, “I mean you can tell. Two feet. Two arms.” Before he testified at trial, Bordeaux’s story had changed several times. After he agreed to a plea deal, Bordeaux was unable to locate where the fatal accident had occurred. Trooper Hader asked Bordeaux about the sequence of events and the location of the accident six different times. Bordeaux consistently denied traveling east from The Reno, the bar where he, McFarling, and Garding had been drinking before getting into Garding’s vehicle. Parsons was east of The Reno when he was killed. After Trooper Hader told Bordeaux where the accident occurred, Bordeaux changed his story to match that location. Bordeaux also changed his narrative of the evening several times. For example, Bordeaux originally claimed that Garding’s vehicle had “rolled over” something, and that “Garding would have stopped if she knew she hit something.” At trial, Trooper Novak was asked whether it was his opinion that Bordeaux had made inconsistent statements. Novak responded, “That would not be my opinion. That would be fact.” Bordeaux’s testimony conflicted with the testimony of Barry, who had been walking beside Parsons. According to Barry, Parsons had not flown through the air, with his arms  GARDING V. MONTANA DEP’T OF CORR. 29 and legs visible. Rather, Barry testified that after Parsons was struck, he was carried by the vehicle across its hood and windshield and that he slid off the hood when the vehicle stopped. Bordeaux’s testimony also conflicted with the testimony of Paul McFarling, Garding’s back-seat passenger. McFarling testified that he had spent the evening drinking at The Reno with Garding and Bordeaux, and that the three of them had left The Reno in Garding’s vehicle in search of cocaine. He testified that when they were driving near the I- 90 underpass he had argued with Bordeaux about a handgun. The underpass is a considerable distance from where Parsons was killed. When Trooper Novak told McFarling that Bordeaux had said that Garding had hit Parsons, McFarling told Novak that Bordeaux’s story was “ridiculous” and “pure fiction.” He told Novak that “there was not one cell or molecule in his body that believed Katie Garding hit anything that night.” The county attorney offered McFarling an immunity deal on an unrelated charge if he testified against Garding, but McFarling refused the deal. When McFarling testified at trial, he was asked, “And without a doubt, while you were in the vehicle with them, she hit nothing that night[?]” He answered, “She hit nothing.” Cornell, the jailhouse inmate who had rekindled interest in Garding, was not called to testify by the State because his story was replete with inconsistencies and contradictions. Instead, Cornell was called by Garding to underline the weakness of the State’s case. Cornell had initially told Trooper Hader that he had taped a light back onto the front of Garding’s vehicle the day after Parsons was killed. Cornell’s statement conflicted with Hader’s own observations. Hader had stopped Garding’s 30 GARDING V. MONTANA DEP’T OF CORR. vehicle in the late morning the day of the accident because of the crack in her windshield. Damage to the light, as described by Cornell, would have been easily and immediately visible. Yet Hader testified at trial that he had observed no damage to the front of Garding’s vehicle. Cornell had originally told the authorities that Bordeaux was the driver and that Garding was performing a sexual act on him when Parsons was hit. Later, after Bordeaux was placed in a pod with Cornell at the Missoula County Detention Center, Cornell changed his story to say that Garding had been driving. Michael Crawford, Cornell’s cellmate at the time, testified that Cornell had told him that he was going to lie and say that Garding had been driving. The prosecution called the state medical examiner, Dr. Gary Dale, to testify about Parsons’ injuries. Dr. Dale testified that the cause of death was a skull fracture caused by contact with asphalt. He testified that the location of injuries to both of Parson’s calves and a fracture of his left fibula was consistent with the height of the bumper on Garding’s vehicle. The prosecutor asked only about the height of the bumper. He did not ask whether Garding’s bumper, which was an unusual square-edged after-market front bumper, could have caused the injuries to Parsons’ calves. Garding called Dr. Thomas Bennett, a forensic pathologist, who testified that the unusual bumper on Garding’s vehicle could not have caused the injuries to Parsons’ calves. Dr. Bennett testified, “This is not the mark a square bumper like this would leave.” Rather, “these bruises are more consistent with a rounded bumper.” The State relied on Troopers Strauch, Hader, and Novak to reconstruct the accident. Trooper Strauch testified that he  GARDING V. MONTANA DEP’T OF CORR. 31 had received over 160 hours of crash investigation training, 16 hours of training on forensic mapping software, and another 80 hours of training in “reconstruction school.” Strauch had drawn a map of the scene of the crash that was introduced into evidence. Trooper Hader testified that he had training as a “technical crash investigator,” had completed over 240 hours of “crash reconstruction” course work, and had responded to 1,600 crashes over sixteen years. Hader testified that he had initially searched for a vehicle with heavy front-end damage caused by a “full-frontal impact,” based on Barry’s description of the crash. He testified that he changed his mind about the nature of the impact after he personally inspected Parsons’ body at the funeral home two days after crash: “Upon examining the body, it was evident to me that we didn’t have a full-frontal impact with the injuries that the body showed to us. . . . Basically all I saw on Mr. Parsons was a bruise on his left calf [in addition to] his head injury that happened when he hit the pavement.” Based on what he perceived as a minor injury only to Parsons’ left calf, Hader concluded that the vehicle had not hit Parsons with “full- frontal impact.” Rather, in Hader’s opinion, the vehicle had swerved and merely clipped Parsons on his left side. Hader discounted Barry’s eyewitness testimony that Parsons had been on the vehicle’s hood as “pretty much . . . impossible.” “I feel what he saw was Mr. Parsons being flipped by the vehicle.” Trooper Novak testified that he had worked for the Montana Highway Patrol for about five years and that he had been trained at the Advanced Traffic Enforcement Academy and had received additional field training. Novak had been the first trooper to arrive at the scene. He estimated the distance between the point of impact and where Parsons was 32 GARDING V. MONTANA DEP’T OF CORR. found as somewhere between 90 and 150 feet. He testified that he had originally believed that the striking vehicle would have sustained “heavy, front-end damage.” Novak testified that he accompanied Trooper Hader to the funeral home to examine Parsons’ body. After that visit and after gathering “more information,” he concluded that he “should be looking for a vehicle with minor front-end damage on the right side.” Novak said that he described the impact to Dr. Dale as “more of a clip.” Streano was not prepared to refute the troopers’ accident reconstruction testimony. She did not object to any of their testimony on the ground that they had not been qualified as experts. She had not consulted an accident reconstruction expert and offered no expert testimony of her own. B. Postconviction Evidence In 2015, Garding sought postconviction relief in state court. She was represented by the Montana Innocence Project. Garding presented evidence from three accident- reconstruction experts: Keith Friedman, an expert in pedestrian impact crash reconstruction with over thirty-five years of experience; David Rochford, an expert in a crash reconstruction with over forty years of experience; and Dr. Harry W. Townes, an expert in crash reconstruction with over fifty years of experience. All three experts concluded that the State’s theory of the accident was impossible. In Friedman’s words, the State’s theory “violates the laws of physics.” The experts identified critical flaws with the State’s theory. Most important was the fact that there was no damage to Garding’s vehicle. Given the nature and extent of Parsons’ injuries, the experts each concluded that the impact would have caused significant damage to the bumper  GARDING V. MONTANA DEP’T OF CORR. 33 and the windshield. In addition to the injury to Parsons’ legs, they pointed to an abrasion on his left shoulder consistent with the size and shape of a windshield wiper and shards of windshield glass on Parsons’ clothing. Further, if the accident had occurred in the manner posited by the State, Parsons would have struck a radio antenna at the base of the windshield on the passenger side of Garding’s vehicle. The antenna was undamaged. The experts also all concluded that Parsons’ leg injuries could not have been caused by Garding’s square-edged bumper. Her bumper had no shock absorbing capacity. Friedman concluded that Parsons’ leg injuries were instead consistent with a modern rounded bumper with shock- absorbing technology. Friedman’s simulations showed there would have been “catastrophic fractures” to both of Parsons’ legs if Garding’s bumper had hit him, even if her vehicle had only been going 15 mph. Rochford similarly concluded that Garding’s bumper would have caused far more damage to Parsons’ legs. Dr. Townes concluded that, given the nature of the front bumper, if Garding’s vehicle had struck Parsons, the tibias and fibulas in both of his legs would have been broken. KARCO Engineering LLC, an automotive and safety testing firm, conducted a physical crash test using a nearly exact replica of Garding’s vehicle, including her customized bumper. The test vehicle traveled at 35 miles per hour (the speed limit on Highway 200) and hit a stationary 198-pound dummy. The dummy victim was hit in the legs by the front of the test vehicle. The dummy’s head then struck the hood and windshield. The vehicle’s grille and trim around the passenger side headlight were broken in several places; the hood was badly dented; and the windshield was broken and dented by several inches. Two photographs were put into 34 GARDING V. MONTANA DEP’T OF CORR. evidence, illustrating the difference between Garding’s undamaged vehicle and the damaged test vehicle: Garding’s Vehicle KARCO Test Vehicle: Post- Crash The three experts unanimously concluded that Garding’s undamaged vehicle could not have possibly struck Parsons. According to Friedman, the State’s theory of the case was a “physical impossibility.” He concluded categorically, “Systems analysis proves that Ms. Garding’s vehicle was not involved in the death of Mr. Parsons.” Dr. Townes wrote that it was “beyond a reasonable doubt” that Garding’s vehicle did not strike Parsons. Rochford wrote that Garding’s “Blazer was obviously not the vehicle that struck Mr. Parsons.” The State changed its theory in response to Garding’s expert evidence. The State’s theory at trial had been that the vehicle had been traveling somewhere between a high and normal rate of speed, as Barry and Baylor had testified, and that the right side of the vehicle had “clipped” Parsons, as Troopers Hader and Novak had testified. Now, on  GARDING V. MONTANA DEP’T OF CORR. 35 postconviction review, in an unsigned and undated report by Trooper Philip Smart, the State advanced an entirely new theory. Trooper Smart recounted in his report that he had received “over 300 hours of instruction in crash investigation” and was “an instructor a[t] the Montana Law Enforcement Academy on the subject.” His report concluded, contrary to the evidence the State had presented and relied upon at trial, that Garding’s vehicle had been traveling “below 20 mph,” perhaps as low as 12–16 mph. Smart speculated, “If driven by a distracted and/or impaired driver who may have intended to pull over, it might have been going slowly and sneaked up behind Mr. Parsons and Mr. Barry.” Trooper Smart wrote, “Mr. Barry never said the vehicle was going fast.” “[T]he collision was a low-speed carry. This matches Barry’s description of the collision[.]” Trooper Smart’s description of Barry’s evidence was flat wrong. Barry had described the vehicle as traveling “extremely fast.” He estimated its speed as 60 mph. Smart’s description was also inconsistent with Baylor’s testimony. Baylor, who had been on Highway 200 driving the other direction, estimated the vehicle’s speed as “regular speed.” Garding’s three experts disagreed vehemently with Trooper Smart. Dr. Townes wrote that Smart had misapplied scientific methods. Friedman wrote that Smart provided “a deeply flawed analysis.” “His whole premise relied on his misunderstanding of the injuries received and then misusing a table in a paper with the erroneous injury information.” Rochford criticized Smart for failing to observe the appropriate professional standards of crash reconstruction procedure, and accused Smart of failing to 36 GARDING V. MONTANA DEP’T OF CORR. read or understand the articles cited in his report. Rochford concluded, “It is astonishing that the state proceeded on a vehicular manslaughter case, without first having an analysis and reconstruction performed by an expert qualified in the field of auto [v.] pedestrian crashes.” Streano, Garding’s trial lawyer, testified during postconviction proceedings that she had failed “to take necessary steps to consult with an accident reconstruction expert and secure appropriate testing,” “failed to request funding to secure testing,” “failed to request more time to secure testing,” and failed to make an investigation into the use of accident reconstruction in support of Garding’s defense. She testified that her “failure to take these steps had nothing to do with strategy.” Defense investigator Mori Woods also testified that she and Streano never discussed the possibility of consulting with or procuring an accident reconstructionist during the investigation leading up to Garding’s trial. Two experienced criminal defense attorneys testified that Streano had provided ineffective assistance. David Ness, who had been a criminal defense attorney for over 30 years, and Wendy Holton, who had practiced law in Montana since 1989, testified that an accident reconstruction was the most critical aspect of the case, and that Streano’s failure to consider employing an expert in the field fell below an objective standard of reasonableness under prevailing professional norms. II. Discussion The Sixth Amendment guarantees a criminal defendant’s right to effective assistance of counsel. This right is “beyond question a fundamental right.” Kimmelman v. Morrison, 477 U.S. 365, 374 (1986). “[I]t assures the fairness, and thus  GARDING V. MONTANA DEP’T OF CORR. 37 the legitimacy, of our adversary process.” Id. at 378. “[A]ccess to counsel’s skill and knowledge is necessary to accord defendants the ‘ample opportunity to meet the case of the prosecution’ to which they are entitled.” Strickland, 466 U.S. at 685. To prevail under Strickland, a defendant must demonstrate, first, that counsel’s performance fell below an objective standard of reasonableness and, second, that the defendant was prejudiced by reason of counsel’s actions. Kimmelman, 477 U.S. at 375. The Montana Supreme Court denied Garding’s ineffective assistance claim under Strickland’s first prong, holding that Streano’s performance was professionally competent. Garding, 466 P.3d at 507. The Court wrote, “Our examination of the trial record ‘in light of all the circumstances,’ leads us to the conclusion that Garding’s trial counsel presented an extensive and strong defense.” Id. (internal citation omitted). “Against the entirety of the trial record, Garding claims ineffective assistance because her counsel did not pursue another possible defense tactic—the hiring of an accident reconstructionist. Notably, the State did not pursue an accident reconstruction either. . . . [T]he trial record here proves convincingly that Garding’s counsel provided a strong defense.” Id. at 508. The Court did not reach Strickland’s second prong. Justice Gustafson dissented. She wrote, “[E]ffective cross-examination did not and could not counter officer testimony about the mechanics of the collision. Expert testimony to explain why the scenario offered by the officers violated the laws of physics and could not have occurred was required.” Id. at 517 n. 4 (Gustafson, J., dissenting). “[I]t was constitutionally deficient to allow the State to put on non-expert opinions about the mechanics of the impact without any counter. The officer[s’] testimony likely carried 38 GARDING V. MONTANA DEP’T OF CORR. much weight with the jury and trial counsel failed to provide expert evidence to support an alternative scenario or to explain that the State’s theory violated the laws of physics and was not physically possible.” Id. at 517. Garding sought federal habeas under 28 U.S.C. § 2254. Applying the deferential standard of AEDPA, the district court granted relief, holding that Garding had received ineffective assistance of counsel. Applying Strickland, the court found, first, that Garding’s attorney performed deficiently, and, second, that Garding was prejudiced. I agree with the district court. Garding is innocent, but her innocence is not the legal basis for my agreement with the district court. Rather, I agree with the district court because Garding has satisfied both steps of the Strickland analysis and is entitled to relief under AEDPA. In reaching both steps, I consider not only evidence showing deficient performance but also evidence showing prejudice. A. Deficient Performance To establish deficient performance under Strickland, a petitioner must demonstrate that counsel’s representation “fell below an objective standard of reasonableness.” Strickland, 466 U.S. at 688. Counsel are constitutionally deficient when their “unprofessional errors so upset the adversarial balance between defense and prosecution that the trial was rendered unfair and the verdict rendered suspect.” Kimmelman v. Morrison, 477 U.S. 365, 374 (1986). It is well established that Strickland imposes a “duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.” 466 U.S. at 691; Kimmelman, 477 U.S. at 384; Hinton v.  GARDING V. MONTANA DEP’T OF CORR. 39 Alabama, 571 U.S. 263, 274 (2014). An attorney’s “performance at trial, while generally creditable enough,” cannot justify the “apparent and pervasive failure to” uphold this duty. Kimmelman, 477 U.S. at 386; see also United States v. Cronic, 466 U.S. 648, 657, n. 20 (1984). The Supreme Court has consistently held that a single, serious error is sufficient to support a claim of ineffective assistance of counsel. Kimmelman, 477 U.S. at 383; Murray v. Carrier, 477 U.S. 478, 496 (1986); Cronic, 466 U.S. at 657 n.20. “Criminal cases will arise where the only reasonable and available defense strategy requires consultation with experts or introduction of expert evidence, whether pretrial, at trial, or both.” Harrington v. Richter, 562 U.S. 86, 106 (2011); see also ABA Criminal Justice Standards for the Defense Function 4-4.1(d) (2017) (“Defense counsel should determine whether the client’s interests would be served by engaging fact investigators, forensic, accounting or other experts.”). Under AEDPA, a federal court may not grant habeas relief to a state prisoner unless the state court’s adjudication of the claim “(1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States; or (2) resulted in a decision that was based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d). The Montana Supreme Court failed to recognize that Streano’s failure to investigate accident reconstruction constituted a single, serious error rising to the level of ineffective assistance. The Court described the ways in which Streano had performed at a professionally competent level, and it is true that Streano provided effective assistance 40 GARDING V. MONTANA DEP’T OF CORR. in some respects. Garding, 466 P.3d at 507–08. But the Court failed to recognize the critical importance of the accident reconstruction evidence that Streano could have introduced but did not. We know from Garding’s postconviction proceedings that expert accident reconstruction testimony would have been easy to obtain, and that such testimony would have been devastating to the State’s case. Garding’s postconviction counsel put on three experts who unanimously concluded that the accident that killed Parsons could not have occurred in the manner described by the troopers. The State effectively conceded as much. Instead of defending the accident reconstruction theory of Troopers Strauch, Hader and Novak, the State advanced an entirely different theory. With no supporting evidence, Trooper Smart concluded that Garding’s vehicle had been traveling at a speed of less than 20 mph, perhaps even as low as 12– 16 mph, and might have “sneaked up” on Parsons. Garding’s three experts unanimously concluded that the accident could not have happened in the way posited by Smart. The Montana Supreme Court said nothing about any of this. The Court wrote only, “Garding argues, based primarily on an affidavit provided by her trial counsel, that the [Montana] District Court erred by concluding her trial counsel did not render ineffective assistance by failing to hire an accident reconstructionist.” Garding, 466 P.3d at 506 (emphasis added). And it wrote, “Notably, the State did not pursue an accident reconstruction either.” Id. at 508. Neither of the Court’s statements is true. First, while Garding did rely on Streano’s affidavit, that was not the primary basis for her argument in the  GARDING V. MONTANA DEP’T OF CORR. 41 postconviction proceedings. Rather, Garding’s argument relied heavily on the failure of her counsel to present accident reconstruction evidence at trial, and on the utterly convincing accident reconstruction evidence introduced at postconviction. The Montana Supreme Court failed to acknowledge the ease with which accident reconstruction evidence had been obtained from three unanimous experts for postconviction proceedings, and the ease with which that evidence could have been obtained for trial. It also failed to acknowledge the devastating effect that this evidence would have had on the State’s case against Garding. See Avila v. Galaza, 297 F.3d 911, 919 (9th Cir. 2002) (noting that this court has consistently held that a lawyer who fails to investigate evidence that could demonstrate her client’s factual innocence renders deficient performance); Rios v. Rocha, 299 F.3d 796, 805 (9th Cir. 2002) (“The failure to investigate is especially egregious when a defense attorney fails to consider potentially exculpatory evidence.”); Hart v. Gomez, 174 F.3d 1067, 1070 (9th Cir. 1999). Indeed, the Court failed even to acknowledge the existence of the three experts’ accident reconstruction evidence. Second, contrary to the Montana Supreme Court’s statement, the State did “pursue accident reconstruction” at trial. It did so through Troopers Strauch, Hader and Novak. The troopers were not qualified as experts, but they provided accident reconstruction testimony as if they were experts. Streano mounted no defense against the troopers’ testimony. She did not challenge their qualifications and allowed them to provide what was, in effect, expert testimony. Because she had not investigated the possibility of expert accident reconstruction testimony, she could not counter their testimony with expert testimony of her own. See Duncan v. Ornoski, 528 F.3d 1222, 1235 (9th Cir. 2008) (holding that 42 GARDING V. MONTANA DEP’T OF CORR. the duty to consult with an expert is particularly important “when the prosecutor’s expert witness testifies about pivotal evidence” and when counsel “has no knowledge or expertise about the field”). I agree with the district court and conclude under AEDPA that the decision of the Montana Supreme Court majority that Streano provided professionally competent assistance was error. The result was “a decision that . . . involved an unreasonable application of” Strickland, and “a decision . . . based on an unreasonable determination of the facts in light of the evidence presented in the State court proceeding.” 28 U.S.C. § 2254(d). B. Prejudice For counsel’s inadequate performance to constitute a Sixth Amendment violation, the petitioner must also show that counsel’s failures prejudiced her defense. Strickland, 466 U.S. at 692. A petitioner “must show that there is a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” Id. at 694. Because the Montana Supreme Court did not reach the question of prejudice, AEDPA deference is not required. Rompilla v. Beard, 545 U.S. 374, 390 (2005); Barker v. Fleming, 423 F.3d 1085, 1093 (9th Cir. 2005). However, even if AEDPA were to apply, I would reach the same result. The prejudice to Garding’s case is beyond obvious. The three accident reconstruction experts—with over a century of combined experience—uniformly concluded that Garding’s vehicle could not possibly have struck Parsons. The experts disproved the state troopers’ theory at trial and made a laughingstock of Trooper Smart’s theory at postconviction. The weakness of the State’s case at trial  GARDING V. MONTANA DEP’T OF CORR. 43 makes prejudice all the more evident. See Strickland, 466 U.S. at 695–96. The only purported eyewitnesses connecting Garding to the crime were Bordeaux and Cornell. Bordeaux testified in return for a spectacularly good plea deal. He provided a number of different stories before he settled on the story he told on the stand. He could testify accurately as to the location of the crash only because Trooper Hader told him where it occurred. Cornell was so unreliable that the State was unwilling to put him on the stand. The evidence now before us tells one story: Garding is factually innocent. From the moment Garding was first questioned by Trooper Hader until today, she has consistently maintained her innocence. Had the reconstruction evidence been presented at trial, the “likelihood of a different result” is more than “substantial.” Richter, 562 U.S. 86 at 112. It is a virtual certainty. Conclusion Garding has spent many years in prison for a crime she did not commit. In Garding’s own words, “When Bronson was hit that night, not just one innocent life was taken but two.” Garding has suffered a great injustice at the hands of the State of Montana. Today, she suffers another injustice at the hands of this court. I dissent.