Nicholas SCHWIETERMAN, Petitioner-Appellant, v. Kevin SMITH, Warden, Respondent-Appellee, 612 F. App'x 825

MERRITT, Circuit Judge. The state prisoner petitioning for habeas relief in this case was involved in an automobile crash that killed four teenagers. Having no memory of the accident, the petitioner followed his trial counsel’s advice to “throw himself on the mercy of the court” by pleading no contest to charges that included four counts of involuntary manslaughter. After receiving a twenty-four-year sentence, he secured new counsel and commissioned a scientific reconstruction of the accident that flatly contradicted factual conclusions that had caused his plea and influenced his sentence.

Relying on this reconstruction, he pursued a claim of ineffective assistance of counsel at the guilty plea level, first through state postconviction proceedings and then through a federal habeas petition under 28 U.S.C. § 2254. The state failed to include a crucial piece of the postconviction petition when submitting the state record to the federal district court, leaving the district court unable to conduct a complete evaluation of the habeas petition.

Working without this undisclosed reconstruction evidence, the district court misapplied the law governing the Sixth Amendment guarantee of effective assistance of counsel during plea bargaining and incorrectly concluded that the facts contested by the petitioner’s expert reconstruction had no bearing on his habeas petition.

Because these untested facts may control the outcome under clearly established federal law, we reverse the district court’s dismissal of the Sixth Amendment claim and remand for a hearing beginning with the full state-court record, including the reconstruction, to determine whether petitioner’s plea counsel was, in fact, ineffective. I. The petitioner, Nicholas Schwieterman, claims a violation of the Sixth Amendment as a result of his plea counsel’s cursory investigation of a car accident that killed four teenagers.

Schwieterman argues on the basis of an expert affidavit and video reconstruction secured after sentencing that he and the court relied on a misunderstanding of fundamental facts when he entered a no contest plea and the court imposed a twenty-four year sentence. At the time of the accident, Schwieter-man was driving under the influence of cocaine, alcohol, and marijuana, and his passenger was unconscious. They report no memory of the impact, no one in the other car survived, and no bystanders witnessed the collision. The only evidence was blood drawn from the two drivers and the wreckage of the vehicles.

The police documented the resting positions of the two vehicles, but their flawed attempts to recover data from the vehicle sensors destroyed all digital information about the vehicles’ speeds, leaving only the speedometers from the two cars, Schwieterman’s displaying 12 miles per hour (mph) and the other showing 84 mph.

The state advanced a theory primarily based on speculation by firefighters and police investigators that Schwieterman ran a stop sign and collided directly with the side of the teenagers’ car, “T-boning” the car and killing its occupants, including the son of one of the firefighters who responded to the accident. See State v. Schwieterman, No. 10-08-17, 2009 WL 1365087, at *5 (Ohio Ct.App. May 18, 2009) (describing a firefighter who was “apprehended by his colleagues to prevent him from finding his deceased son in the wreckage”); Sentencing Tr. 77.

Schwieterman’s attorneys did not engage in an extensive independent investigation, but instead retained an expert, Doug Heard, to review the state’s reports. On the basis of Heard’s document review, Schwieterman’s attorneys recommended that he “throw himself on the mercy of the court. He pleaded no contest and was sentenced to four consecutive six-year sentences. The first forensic analysis — initially presented in support of Schwieterman’s post-conviction petition — tells a different story.

After sentencing and with new attorneys, Schwieterman hired a different expert, Wilbur Meredith, to reconstruct the accident based on the limited evidence not destroyed by police mishandling. Meredith Aff. ¶¶ 2, 4-10. Meredith reviewed the state’s reports. Id. at ¶ 3.

He also used accident reconstruction software commonly utilized by the Ohio State Police and other law enforcement agencies to reconstruct the collision based on data from the police reports, his “personal inspection and measurements at the crash site and of the two vehicles involved,” photographs taken by himself and the police, and “the approximate crush dimension's on the two vehicles.” Id. at ¶¶ 9,12-13.

Through an affidavit filed in state and federal courts, Meredith. opines, “beyond any reasonable doubt and with engineering certainty, ... this accident was not the result of a ‘T-Bone’ type crash.” Meredith Aff. ¶ 14. Instead, his affidavit indicates that the two cars collided at an angle of 45 degrees. Id. at ¶ 15. It also concludes that Schwieterman’s car was traveling at 12 mph, possibly after stopping at the intersection, and the teenagers’ car was traveling at 84 mph at the time of impact, consistent with the speedometer readings. Id. at ¶¶ 17-18, 20.

Meredith supplemented his affidavit with two sets of video reconstructions of the collision that he created to illustrate his analysis. The videos illustrate Meredith’s reconstruction of both the state’s “T-Bone” theory of the accident and Schwieterman’s alternative theory. One set of videos demonstrates how a collision involving Schwieterman’s car traveling through the intersection at approximately 12 mph (perhaps after stopping) and the teenagers’ car traveling at 84 mph could result in the vehicle placement documented by the state investigators.

The other set shows a “T-Bone” style crash such as the one suggested by the state that results in a dramatically different placement. Although Schwieterman filed these videos on a DVD with the state postconviction court, the state did not include them in the record it provided to the habeas district court. We take no position on the persuasiveness of these videos, which have never been subject to an evidentiary hearing.

The state’s failure to provide them along with the rest of the state postconviction records deprived the district court of important evidence that it needed to properly analyze Schwieterman’s petition. We leave the first full analysis of this evidence to the district court on remand and turn to the legal question still before us. II. After unsuccessfully pursuing direct appeals and postconviction relief in the state courts, Schwieterman filed a habeas petition on June 10, 2011, raising several issues.

The district court rejected each of his claims, and only an ineffective assistance of counsel claim focused on the plea stage was certified for review. The district court first concluded that the ineffective assistance claim was foreclosed by procedural default. After the state postconviction trial court rejected Schwieterman’s ineffective plea assistance claim, State v. Schwieterman, No. 08-CRM-022, slip op. at 6-7 (Ohio Ct.Com.Pl.

July 23, 2009) (relying on an earlier analysis of actual innocence to find an absence of prejudice), the state appeals court affirmed the result through a sua sponte invocation of res judicata. State v. Schwieteman, No. 10-09-12, 2010 WL 169192, at *8-10 (Ohio Ct.App. Jan. 19, 2010). The district court credited the state appellate court’s determination that the claim could and should have been raised on direct appeal and incorrectly concluded that the state court’s anomalous application of that procedural rule triggered procedural default as an independent, adequate state ground for denying further review.

This ruling was in error. An analysis of procedural default begins by determining whether “the petitioner failed to comply with an applicable state procedural rule.” Hanna v. Ishee, 694 F.3d 596, 607 (6th Cir.2012). A state rule must be “firmly established and regularly followed” before its violation can trigger procedural default of a federal habeas claim. Walker v. Martin, 562 U.S. 307, 315-317, 131 S.Ct. 1120, 1127-28, 179 L.Ed.2d 62 (2011) (quoting Beard v. Kindler, 558 U.S. 53, 60, 130 S.Ct. 612, 175 L.Ed.2d 417 (2009)).

Thus, if Schwieterman actually complied with Ohio law by first raising his ineffective assistance claim through a postconviction petition, that claim is not procedurally defaulted. When claims cannot be adjudicated on the trial record alone, Ohio law does not require defendants to raise those issues on direct appeal where they would necessarily fail for lack of evidence. See State v. Cooperrider, 4 Ohio St.3d 226, 448 N.E.2d 452, 454 (1983); State v. Cole, 2 Ohio St.3d 112, 443 N.E.2d 169, 171 (1982).

Instead, Ohio law prescribes the parallel paths Schwieterman took: directly appealing the issues supported by the record and bringing a separate collateral petition to address issues that required further factual development. Cole, 443 N.E.2d at 171; State v. Gibson, 69 Ohio App.2d 91, 430 N.E.2d 954, 959-60 (1980) (Krenzler, C.J., concurring).

The legal distinction between claims appropriate for direct appeal and those reserved for a collateral petition is not whether Schwieterman was aware of the alleged constitutional deprivations but whether those claims “could fairly be determined without examining evidence outside the record.” Cole, 443 N.E.2d at 171. Compare with Schwieterman, 2010 WL 169192, at *9 (“[A]ll of these claims contain issues that Schwieterman was aware of, or should have been aware of, at the time of his appeal, yet he failed_to assert any of these claims in his direct appeal from his conviction and sentence.”).

Schwieterman’s claim of ineffective assistance falls in the second category because, like most ineffective assistance of counsel claims, it cannot be evaluated without examining evidence outside the initial record of the plea and sentencing. Compare Cole, 443 N.E.2d at 171 (“Generally, the introduction ... of evidence dehors the record of ineffective assistance of counsel is sufficient, if not to mandate a hearing, at least to avoid dismissal on the basis of res judicata.”), with Schwieterman, 2010 WL 169192, at *8 (disregarding the petition’s reliance on evidence outside the record).

Ohio Court of Appeals decisions applying Cole’s test reveal that at a minimum, Meredith’s affidavit would likely qualify as evidence “dehors the record” necessary to bring Schwieterman’s claim. In order to evaluate Schwieterman’s claim, it is necessary to know what an expert would have concluded about the accident. See, e.g., In re B.C.S., No. 07CA60, 2008 WL 4823572, at *11 (Ohio Ct.App. Oct. 29, 2008) (“In support of his postconviction claims, ... Sturm presents his trial counsel’s affidavit and supporting documents, including expert reports.

We believe that this constitutes competent, relevant and material evidence outside the record sufficient to avoid operation of the res judicata doctrine. While the record shows that defense counsel did not call these experts, it does not indicate his motivation in failing to call these experts or what testimony the ex-peHs would have provided to the jury.” (emphasis added)); State v. Jenkins, No.2003-CA-1, 2003 WL 21995250 at *7 (Ohio Ct.App.

Aug. 22, 2003) (“Jenkins’ ineffective assistance of counsel claim based on his attorney’s failure to have an expert witness critique police interviews with his accusers ^required the presentation of evidence outside the trial record. Although the trial record revealed that no such expert was called, the trial record did not reveal why defense counsel failed to call an expert or what testimony such an expert would have provided for the jury.... Without record evidence setting forth ivhat such an expert luould have told the jury, he could not demonstrate prejudice.” (emphasis added)).

Sehwieterman complied with Ohio law regarding constitutional challenges that require evidence outside a record of conviction. He complied with the firmly established procedural rules of the state, and the state court’s deviation from those firmly established rules does not constitute default by Sehwieterman.

Because federal law does not permit “novel and unforeseeable” applications of state procedural rules to thwart federal review of habeas claims, Walker, 562 U.S. at 319, 131 S.Ct. at 1130, the state’s irregular invocation of res judi-cata in this case is not an adequate independent state ground barring federal review. We therefore consider the merits of Schwieterman’s claim. Counsel is ineffective under the Sixth Amendment if objectively deficient performance prejudices a criminal defendant. Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984).

A petitioner asserting ineffective assistance following a guilty (or no contest) plea shows prejudice by demonstrating “a reasonable probability that, but for counsel’s errors, he would not have pleaded guilty and would have insisted on going to trial.” Hill v. Lockhart, 474 U.S. 52, 59, 106 S.Ct. 366, 88 L.Ed.2d 203 (1985). Thus, the proper standard for ineffective assistance through failure to investigate is merely the likelihood that investigation would have led to a different recommendation and decision regarding the plea. Id.; Dando v. Yukins, 461 F.3d 791, 802 (6th Cir.2006).

In its limited merits analysis, conducted out of “an abundance of caution” following its conclusion that res judicata barred Schwieterman’s petition, Mem. Op. 16, the district court erroneously concluded that “[Schwietermanj’s burden is to show that, had he gone to trial armed with the report of the expert attached to his post-conviction petition, the ultimate outcome would have been different.” Id. at 20 (emphasis added). But Schwieterman’s burden is less onerous. The relevant “ultimate outcome” in this case is not the result of a speculative trial but the more immediate decision about the recommended plea.

Speculation about trial outcomes is relevant, but only insofar as the likely trial outcome would influence an objective decision to abandon the investigation and enter a plea. Instead of asking whether reasonable counsel and a reasonable defendant would have accepted the state’s facts and entered a no contest plea after reviewing the expert reconstruction, the district court used an actual-innocence standard from an unpublished district court opinion and dismissed the claim because the reconstruction would not “make it impossible for a reasonable jury to convict.” Id. (quoting United States ex rel. Blair v.

Rednour, No. 11-cv-4108, 2012 WL 1280831, at *6 (N.D.Ill. Apr.11, 2012)). The district court found Sehwieterman failed to meet this irrelevant actual-innocence standard of prejudice and rejected his Sixth Amendment claim on the merits without addressing the deficient-performance prong of the ineffective assistance analysis.

Accordingly, the judgment of the district court is reversed as to the ineffective assistance of counsel claim, and the case is remanded to the district court to determine whether the petitioner’s plea counsel was ineffective in light of the postconviction expert evidence concerning how the accident occurred, including the video reconstruction evidence mistakenly withheld from the district court.