Martin Woolley v. Dave Rednour

WILLIAMS, Circuit Judge. Martin M. Woolley was charged in Illinois state court with counts of murder, armed violence, armed robbery, and unlawful possession of a firearm by a felon, arising out the fatal shootings of two victims in 1995. After initially confessing, Martin later recanted, claiming he had falsely implicated himself in order to protect his wife, Marcia Woolley, who committed the murders out of jealousy toward one of the victims. The jury convicted the defendant on all counts. In state post-conviction procedures, Martin produced an expert who pointed out flaws in expert evidence introduced by the State at trial. After obtaining no relief and exhausting review in state court, Martin filed a federal habeas corpus petition pursuant to 28 U.S.C. § 2254, claiming that he was deprived of his constitutional right to effective assistance of counsel. The district court denied Martin relief, but granted a certificate of appealability with respect to his claim that he was prejudiced by ineffective assistance of counsel. Martin appealed. Although Martin’s counsel was ineffective, we affirm because we find that Martin was not prejudiced by the error. *414 I. BACKGROUND Around 10:00 p.m. on February 20, 1995, the bodies of Rane Baldwin and Dianna “Dee” Turley were found in Phylly’s Cue and Brew tavern in Kewanee, Illinois. Nine hours earlier, at 1:00 p.m. on February 20, 1995, Martin Woolley (“Martin”) and his wife, Marcia Woolley (“Marcia”), went to Phylly’s Cue to play pool and drink beer with friends. The Woolleys left the tavern around 5:00 p.m. to go home and prepare dinner for Marcia’s children. They returned to the tavern around 6:00 p.m. and continued to drink beer and play pool. At approximately 7:00 p.m., Baldwin arrived and began her bartending shift. Turley arrived around 8:00 p.m. When Turley arrived, Marcia started to drink shots of hard liquor as well as beer. Soon, Marcia and Turley began arguing with each other while sitting at the bar. By approximately 9:30 p.m., everyone had left the tavern except for Turley, Baldwin, Marcia, and Martin. At 9:45 p.m., Baldwin’s boyfriend, Peter Dolieslager, called the tavern and spoke to Baldwin who told Dolieslager that Marcia was getting drunk. According to Dolieslager, Baldwin gave him the impression that Marcia was being difficult. A few minutes later, Dolieslager drove to the tavern to pick Baldwin up. When he arrived around 10:00 p.m., he discovered the bodies of Turley and Baldwin. An autopsy revealed that Turley had been shot once in the forehead from a distance of three to four feet and that the bullet entered her head at a slightly upward angle. Baldwin had been shot three times in the head. The pathologist concluded that the bullets that struck Baldwin had not been fired at close range. A. Martin’s Initial Confession The next day, Martin and Marcia voluntarily went to the Kewanee police station for questioning. Martin initially denied any involvement in the murders, but eventually confessed. He told the officers that Marcia and he were both drinking beer and playing pool at the tavern. Around 9:30 p.m., Turley, Marcia, and Martin, in that order from east to west, were sitting on some bar stools along the bar. Martin then decided to rob the tavern. He pulled out a 9-millimeter hand gun he had been carrying in the back of his pants and shot Baldwin, who was standing behind the bar, two or three times in the head. He then shot Turley, whom Martin claimed was standing, while Marcia was still sitting on her barstool. Martin then walked around behind the bar, took the cash from the register as well as two money bags that were located underneath the register. While Martin was taking the money, Baldwin made a noise and Martin shot her again in the head from behind the bar. Martin then grabbed Marcia and dragged her out of the bar. When the Woolleys got home, Martin burned the money bags and hid the $300 in cash in a hole in the wall of a closet. He put the gun in a freezer located in a friend’s garage. The next day, Martin asked his friend to throw the gun into the river, but his friend turned it over to the police. Martin told the police that Marcia did not handle the gun and that she had no knowledge of what he was going to do. B. Martin’s Trial Testimony Recanting His Confession During trial, Martin testified in his own defense and recanted his prior confession. He told the jury that around 9:30 p.m., Marcia and Turley were arguing as they had been all evening. Martin decided to leave the women and go to the men’s bathroom to smoke marijuana. He stated that he had previously put the gun in his *415 jacket pocket when he had stopped at home to prepare dinner for Marcia’s children. He left that jacket on the barstool on his way to the bathroom. While he was away from the bar, he heard raised voices followed by gunshots. He ran out of the bathroom and saw Marcia standing up with a foot propped up on her barstool, leaning over the top of the bar. Martin testified that he then saw Marcia shoot down towards the bartender’s side of the bar. Turley was lying next to the bar-stool. Not knowing what else to do, Martin attempted to make the scene look like a robbery. He walked around the bar and took the cash. He then drove home with Marcia. Once home, Martin disposed of the gun. The couple agreed that if the police pursued them, both would say that Martin committed the murders. Marcia had three children from a previous marriage, and Martin testified that he thought that he needed to take the blame for the shootings to protect them because the children would “not miss a step-dad as much as a mother.” C. The State’s Expert Crime Scene Investigator A central component of the prosecution’s case was expert testimony to the effect that the shooter must have fired the shots from the bathroom area near the southwest corner of the bar and that Martin’s account of the shootings was physically impossible. Martin was represented by Eugene Stockton, a part-time public defender and former State’s Attorney. Over two months before trial, Stockton prepared a written statement explaining what Martin had witnessed and his anticipated trial testimony. So the prosecution had advance notice that Martin would testify that he was in the men’s restroom area when Marcia fired the first few shots, that Marcia fired the final shot from her barstool, and that Martin went behind the bar to steal the cash from the register. About a month before trial, the State gave Stockton a crime scene report authored by Michael Ogryzek, the crime scene investigator the State would later call as an expert. The report disclosed Ogryzek’s opinion that the shooter was located in the southwest corner of the tavern by the men’s restroom when at least one of the shots was fired. This was about 18 feet from Marcia’s location in the same area in which Martin would testify that he was located at the time of the shootings. Ogryzek’s conclusions were at odds with earlier investigations of the crime scene following the murders. Initially, police investigators had determined that the location of the shots was consistent with Martin’s confession that they had been fired from the barstools where Martin and Marcia had been sitting. At trial, Ogryzek testified that the physical evidence showed that the person who fired the final shot at Baldwin was standing above her on the bartender’s side of the bar. Stockton did not learn of that opinion until the day trial began, when the State supplemented its answer to the pretrial order. Stockton knew that the State’s disclosure was untimely, and knew the opinion would directly refute Martin’s testimony, but he did not request a continuance to address it or make any motion to bar Ogryzek’s testimony. Ogryzek also testified that it would have been physically impossible for a person where Marcia was located to have delivered the final shot to Baldwin from her position in the third barstool in the manner Martin would testify: [Prosecutor.] Based on your expertise, Officer Ogryzek, could a person with a foot on the bar stool, the third bar stool *416 with their knee on the bar leaning over, have made that shot? [Ogryzek.] No. It is impossible. [Prosecutor.] How long would that person’s arms have to be, to be in that barstool, a foot on the barstool, a knee up on the bar reaching over ... [for a gun shot to] come in at that angle that that gun shot above Rane Baldwin’s ear came in? [Ogryzek.] Oh, twelve feet. Stockton cross-examined Ogryzek but did not get concessions of any significance. D. Other Trial Testimony Several other witnesses testified for the prosecution. One witness testified that while the group was playing pool at the tavern, she saw the imprint of a gun under Martin’s shirt stuck in the back of his pants and that she never saw him move the gun from his pants to his jacket pocket. Another witness testified that on the night of the murders, Martin made a comment saying that he was the type of person who “could walk into McDonald’s and just open up on everybody.” A third witness testified that two nights before Martin commented that it would be easy to commit a robbery if you killed the witnesses. The State also called Donald Tomsha. In March of 1995, Tomsha was in jail and had been charged with burglarizing three businesses. Tomsha later pleaded guilty and received six months in jail and probation. Tomsha testified that Martin confessed to him after the two had begun discussing religion and morality. According to Tomsha, Martin was unhappy that Marcia was not charged for her role in the murders. Martin allegedly told Tomsha that Marcia must have discovered “our pistol” in his jacket as she reached for cigarettes. Tomsha testified that he had questioned Martin’s credibility, at which point Martin confessed that he shot Baldwin as she was turning up the volume of the television at his request. Martin allegedly then handed the gun to his wife to shoot Turley, but Marcia “froze” and Martin took the gun back and shot Turley. Tomsha also testified that Martin told him that he and Marcia planned to rob the tavern that night because it was less busy on Mondays. Tomsha claimed he convinced Martin to write out and sign written confessions to take accountability for his actions. Martin wrote the account down on a legal pad and signed it in his presence. He later supplemented the confession to include a version emphasizing Marcia’s role as a co-conspirator in the shootings. Tomsha then delivered the statements to authorities. At trial, the prosecution produced an FBI handwriting expert who confirmed that the signatures on the documents matched Martin’s and did not belong to Tomsha. The expert could not establish that the information contained inside the documents matched Martin’s handwriting. But the expert explained that Martin refused to provide a natural handwriting exemplar to permit an adequate comparison. For the defense, Stockton called witnesses who testified that Marcia had harbored a grudge against Turley for over a year. Marcia had made many statements to the effect that Turley would “get hers someday.” Additionally, two nights before the murders, Martin and Baldwin had discovered that they had attended school together and had engaged in a lengthy conversation. Witnesses testified for the defense that Marcia was “visibly upset” that Martin and Baldwin were talking. Marcia’s ex-husband also testified that Marcia had become violent with him in the past. Further, it came out at trial that because Marcia had worked at the tavern previously, the Woolleys knew that *417 very little money was kept at the tavern, that the tavern had the most money on Thursdays, and was less crowded on Mondays. During closing arguments, Stockton made an attempt to address Ogryzek’s testimony by drawing a picture on his yellow legal pad and making the following presentation: Then, — I don’t know, I am not a very good artist, — this is her shoulder, the head is straight ... based on their testimony that I believe the bullet would have traveled ... if you tilt the head down to make contact with the floor, that makes the bullet almost straight up and down.... [Ogryzek’s testimony] was inconsistent with what Dr. Jumbelic [the autopsy pathologist] said about the angle of that bullet, and was even inconsistent with what he was saying. Stockton also told the jury that the crime scene did not fit with Ogryzek’s testimony given the angles of the bullets. He argued that it would not have made sense for Martin to go to the bathroom area and shoot the victims from a distance when he was free to fire at them from close range. He emphasized that the bullet that hit Baldwin was going up, and that the prosecution could not explain the upward angle of that bullet if Martin, who was over six feet tall, had been the shooter, but that the angle could have easily been accomplished by Marcia, who was five feet and five inches tall, sitting on the barstool. He also called into question Tomsha’s statement by noting that it would not have made sense for Turley to have stayed seated on the barstool while Martin and Marcia argued about who was going to shoot her. In rebuttal, the prosecution stated: You were told by Mr. Stockton that Martin Woolley couldn’t have made the third shot because he was standing by the cash register.... [B]ut the shot still came ... from the bar area.... Now you heard experts testify as to angles, and I submit to you that you can’t rely on a rough drawing made on a legal pad as to angles here, because I submit to you that this is not the evidence. Martin was convicted and sentenced to death. His sentence was commuted to life in prison when Illinois’s governor commuted the sentences of all inmates on death row. Martin’s conviction was upheld on appeal before he sought post-conviction relief. E. Post-Conviction Proceedings During state post-conviction proceedings, Martin’s current counsel retained Alva Busch, a crime scene reconstruction expert with 20 years of experience who would have been available to testify at Martin’s trial. According to Busch’s analysis, the physical evidence shows that neither the State’s theory nor Martin’s initial confession could be true. Busch opined that the shooter must have been located around the area of the barstools. Using the police's crime scene measurements, Busch built a true-to-scale mockup of the bar and reconstructed the crime scene, reflected in the picture below. Busch conducted an experiment by giving a woman of Marcia’s build and height a model gun. A rod was attached to the gun to show the trajectory of the bullet where the gun was to be fired. Busch found that the wound trajectory, the abrasion ring around the wound, and the angle and damage around the hole in the floorboards underneath Baldwin’s head all confirmed that the shot came from the patrons’ side of the bar. *418 [[Image here]] The shot would have been even easier had the person had a foot on a barstool. Moreover, Busch concluded that Ogryzek’s testimony that the final shot came from someone on the bartender’s side of the bar near Baldwin’s feet was wrong because, had the shooter been at Baldwin’s feet as Ogryzek claimed, the wound trajectory would have pointed in nearly the opposite direction. This testimony would have contradicted and undermined both Ogryzek’s testimony and Martin’s confession to the police that he was behind the bar when he fired the last shot at Baldwin. Busch also concluded that the trajectory of the bullet holes caused by the initial shots to both Baldwin and Turley were consistent with a shooter being located by the barstools and that the shots could not have been made by someone coming out of the men’s restroom. First, the bullet that caused Turley’s wound was found in the tavern’s east door. Had the bullet been fired by someone by the men’s restroom or walking along the south wall (as Ogryzek testified), the bullet would have had to change its course almost 90 degrees after striking Turley to end up in the east door. The diagram below reflects the tavern’s layout and locations of Marcia Woolley, Turley, and Baldwin at the time of the shootings. *419 [[Image here]] In contrast, Marcia had been seen sitting directly to the west side of Turley, meaning that she would have been in an ideal position for a bullet to pass through Turley and hit the east door. Second, Turley’s autopsy report and photographs showed gunpowder residue called “stippling” around Turley’s wound. Stippling is only caused by gunshots fired at a distance of four feet or less. Ogryzek testified that Martin was about 18 feet from Turley when he shot her from the bathroom. In contrast, Marcia’s barstool was less than four feet from Turley. Third, Turley was shot in the forehead and the bullet exited the back of her head at an upward angle. Had the shooter been standing or walking, the gun would have been level or above the wound; but if the gun had been held by someone sitting on Marcia’s barstool, the gun would have been lower than Turley’s forehead and the wound angle would have been upward. This testimony was presented during an evidentiary hearing held before the circuit court of Henry County, Illinois. Stockton also testified that obtaining an expert to rebut Ogryzek’s testimony would have been “ideal,” and although court funds were available for the purpose of procuring experts, Stockton took no steps to find an expert or apply for available funds. He testified that the idea of seeking expert assistance “never crossed [his] mind,” and that he did not know whether such experts were available to the defense in 1995. Busch later testified that there were a number of such experts available in 1995. Based on this new testimony, Martin argued that Stockton’s failure to procure an expert to rebut Ogryzek’s testimony constituted ineffective assistance of counsel. Applying Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), the court concluded that while Martin’s attorney did cross-examine Ogryzek in an attempt to discredit his opinions, counsel was ineffective because “more should have been done.” It then found that “Petitioner’s counsel’s failure to retain an expert witness, ask for a continuance, or move to bar Ogryzek’s testimony, fell below an objective standard of reasonableness.” However, the court also found that Martin was not prejudiced by his counsel’s ineffectiveness. The court concluded that “there was no reasonable probability that *420 ha[d] an expert such as Busch been called and testified, the outcome of the trial would have been different.” The court explained that petitioner “admitted that he had been at [the tavern] at the time of the murder, that he had taken money from the cash register ... that he put the weapon [in the friend’s garage] ... and that he lied to the police.” Thus, the court concluded, “the sole decision for the jury was to determine whether petitioner was lying when he testified, or when he confessed in an oral and written statements [sic] to the police, and to David Tomsha.” The court stated that Martin’s “credibility was illogical, inconsistent, and impeached in numerous ways.” The court also concluded that Busch agreed with Ogryzek on “several matters.” The court further found: [M]uch of the basis for Busch’s disagreement is unconvincing. For instance, Busch discounted Ogryzek’s reliance on the location of the shell casing against the south wall because casings can roll. However the photograph of [Turley’s] body show[s] that the floor is slanted toward the bar, the direction in which the blood flowed, not toward the south wall. The court did not refer to any of the other disagreements between Busch and Ogryzek and did not otherwise explain why “much” of the basis for Busch’s disagreement was unconvincing. The court of appeals affirmed, focusing only on the second prong of Strickland. As to prejudice, the court found that Martin’s guilt or innocence depended on his credibility. It also found that the “key to the case was not where the gunman stood. The jury was called upon to decide if the defendant was honest when he stated he was not the gunman and that he confessed to the details of the crime to protect his wife.” Without opinion, the Supreme Court of Illinois denied Martin’s petition for leave to appeal. Martin then filed a habeas petition, which the district court denied. This appeal followed. II. ANALYSIS A district court’s judgment regarding habeas relief is reviewed de novo. Northern v. Boatwright, 594 F.3d 555, 559 (7th Cir.2010). The Antiterrorism and Effective Death Penalty Act (“AEDPA”) sets the parameters for our review. Under AEDPA, we may grant habeas relief only if a state-court decision was (1) “contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States” or (2) “based on an unreasonable determination of the facts in the light of the evidence presented in the state court proceeding.” 28 U.S.C. § 2254(d); Williams v. Taylor, 529 U.S. 362, 386, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000); Charlton v. Davis, 439 F.3d 369, 374 (7th Cir. 2006). In order for a federal court to find a state court’s application of federal law unreasonable, the court’s application must have been more than incorrect; it must have been objectively unreasonable. Wiggins v. Smith, 539 U.S. 510, 520, 123 S.Ct. 2527, 156 L.Ed.2d 471 (2003). To establish a claim of ineffective assistance of counsel, a petitioner must show that counsel was deficient in his performance and that the deficiency prejudiced his defense. Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984); Pole v. Randolph, 570 F.3d 922, 934 (7th Cir.2009). First, Martin Woolley must demonstrate that his counsel’s performance fell below an objective standard of reasonableness. See Strickland, 466 U.S. at 688, 104 S.Ct. 2052. Second, he must demonstrate that he was *421 prejudiced by the deficient performance. Id. at 694, 104 S.Ct. 2052. On habeas review, a federal court evaluates “the totality of the evidence — both that adduced at trial, and the evidence adduced in the habeas proceeding.” Wiggins, 539 U.S. at 536, 123 S.Ct. 2527 (citation, emphasis, and internal quotations omitted). A. The Performance of Martin’s Counsel Was Ineffective. Before discussing the merits, we will consider the appropriate standard of review for Strickland’s performance prong, which the parties contest. The state trial court first determined that Stockton’s representation of Martin was ineffective, while the state appellate court explicitly declined to reach the ineffectiveness prong of the Strickland test. The State’s highest court then denied leave to appeal the opinion below. Federal courts typically apply de novo review to a Strickland prong left unaddressed by a state court. See Wiggins, 539 U.S. at 534, 123 S.Ct. 2527. Even though no prior decision found Martin’s counsel effective, the State argues that we must nevertheless apply AEDPA deference in its favor on the performance prong because the State ultimately prevailed on Martin’s Strickland claim. Relying on Thompson v. Battaglia, the State contends that we must treat the entire Strickland claim as an indivisible constitutional ground for relief. 458 F.3d 614, 616 (7th Cir.2006). Because the Illinois Appellate Court found no Strickland violation overall, the State maintains that we should apply AEDPA deference to both prongs of the test and presume that the appellate court found defense counsel’s representation adequate even though it remained silent on attorney performance. This novel approach plainly conflicts with Wiggins. But the State argues that the Supreme Court overruled Wiggins in Harrington v. Richter, which held that “§ 2254(d) does not require a state court to give reasons before its decision can be deemed to have been ‘adjudicated on the merits.’ ” — U.S.-, 131 S.Ct. 770, 785, 178 L.Ed.2d 624 (2011). On the other hand, Martin argues that under AEDPA we must defer to the state trial court’s determination that defense counsel was ineffective. Martin contends that since the state trial court considered the ineffectiveness prong and adjudicated it on the merits in his favor, AEDPA operates to preserve this presumption on federal review. He notes that Wiggins applied de novo federal review when “neither of the state courts below reached [the pertinent] prong of the Strickland analysis.” 539 U.S. at 534, 123 S.Ct. 2527 (emphasis added). Since one of the state courts ruled in his favor on the performance prong and the other did not address it, Martin believes we must defer to the state trial court determination on this measure. Both parties’ arguments miss the mark. First, Martin misconstrues the scope of federal review under AEDPA. When a state collateral review system issues multiple decisions, we typically consider “the last reasoned opinion on the claim” — here the opinion of the Illinois Appellate Court. Ylst v. Nunnemaker, 501 U.S. 797, 803, 111 S.Ct. 2590, 115 L.Ed.2d 706 (1991); see also Gonzales v. Mize, 565 F.3d 373, 379 (7th Cir.2009) (“We review the decision of the last state court that substantively adjudicated each claim.”). Unless a state-court opinion adopts or incorporates the reasoning of a prior opinion, “AEDPA generally requires federal courts to review one state decision.” Barker v. Fleming, 423 F.3d 1085, *422 1093 (9th Cir.2005). 1 When Martin appealed the state trial court judgment denying collateral relief, the State reiterated its contention that Stockton’s counsel had been adequate. The state appellate court declined to adopt the trial court’s reasoning and instead remained silent on defense counsel’s performance. This ruling is the “last reasoned opinion” we review under AEDPA. Because the Illinois Appellate Court did not reach Strickland’s ineffectiveness prong, we apply Wiggins to review the issue de novo. Next, the State’s theory that Wiggins has been overruled would stretch Harrington’s holding well beyond the scope of the decision. Harrington addressed a scenario where a conviction was upheld by a summary affirmance of the California Supreme Court. There was no “reasoned opinion” by any lower court on collateral review. By its terms, Harrington applies “[w]here a state court’s decision is unaccompanied by an explanation.... ” 131 S.Ct. at 784. The Harrington Court held that such unexplained determinations may still qualify as adjudications on the merits for purposes of § 2254(d) and should not be presumed to be procedural dismissals absent some opposing indication. But Harrington did not purport to disturb Wiggins or Ylst. To the contrary, the Court cited Ylst’s pass-through rule as a reason why a summary affirmance might not represent a judgment on the merits. See Harrington, 131 S.Ct. at 785 (citing Ylst, 501 U.S. at 803, 111 S.Ct. 2590 (“Where there has been one reasoned state judgment rejecting a federal claim, later unexplained orders upholding that judgment or rejecting the same claim rest upon the same ground.”)). In this case, there is little need for uncertainty regarding the reasoning of the Illinois courts. The state court made the grounds for its ruling abundantly clear. The Illinois Appellate Court explicitly considered Strickland’s prejudice prong in isolation, deeming it “unnecessary to address any other issues raised by the State regarding the [trial] court’s findings on ineffectiveness.” The Supreme Court of Illinois then denied leave to appeal without opinion, presumptively adopting the reasoning of the state appellate court under Ylst. Under such circumstances, Wiggins controls and we review attorney performance de novo. It would be perverse, to say the least, if AEDPA deference required this court to disregard a state court’s expressed rationale for a decision and presume instead that Illinois’s courts affirmatively found defense representation adequate. See Sussman v. Jenkins, 642 F.3d 532, 534 (7th Cir.2011) (Ripple, J., in chambers) (denying motion to stay mandate) (“We certainly cannot assume that the [Harrington] Court overruled sub silentio its holding in Wiggins — a precedent so important to the daily work of the lower federal courts.”). Now to the merits. Under Strickland, our review of defense counsel’s performance is “highly deferential”; Martin must “overcome the presumption that, under the circumstances, the challenged action might be considered sound trial strategy.” Strickland, 466 U.S. at 689, 104 *423 S.Ct. 2052 (internal quotation marks omitted). The choice not to investigate a particular defense does not constitute deficient performance “if a lawyer has made a reasonable decision that makes particular investigations unnecessary.” Adams v. Bertrand, 453 F.3d 428, 436 (7th Cir.2006) (internal quotation marks omitted). We have little difficulty reaching the same conclusion as the state trial court on Strickland’s performance prong: Martin did not receive the effective representation guaranteed by the Sixth Amendment. This case does not present a total failure to provide any meaningful opposition to the State, though defense counsel made more than one mistake in representing his client. Martin’s attorney did engage in substantive cross-examination of government witnesses. But “even an isolated error of counsel” can deny a defendant his right to effective assistance “if that error is sufficiently egregious and prejudicial.” Murray v. Carrier, 477 U.S. 478, 496, 106 S.Ct. 2639, 91 L.Ed.2d 397 (1986); see also Williams v. Lemmon, 557 F.3d 534, 538 (7th Cir.2009) (“a single error may suffice” for a finding of ineffectiveness). Here, Martin’s attorney remained nearly passive in the face of damning, impeachable testimony from Ogryzek that effectively hollowed out the core of his client’s defense in this capital case. Stockton had advance notice that the government planned to introduce testimony from an expert crime scene investigator. Although defense counsel knew Martin’s testimony would turn on the location of the final gunshot, he made no effort to retain a defense expert. More than two months before trial, in accordance with pretrial discovery requirements, the defense disclosed that Martin would testify that he had been in the bathroom when the shootings occurred. The following month, the State indicated that it would reverse its initial conclusion that the gunshots were fired from across the bar; Ogryzek would testify instead that the shots came from the area around Phylly’s bathroom — precisely where Martin now said he had been located during the shootings. Even though he knew that the State would directly inculpate Martin’s trial testimony, Stockton took no steps to secure an expert opinion to rebut the State’s evidence. On the morning of the trial itself, the State made an untimely supplement to its pretrial report: Ogryzek would now testify that it was physically impossible for Marcia to have fired the final shot at Baldwin from across the bar as Martin maintained. Defense counsel made no objection to the state’s request. Defense counsel’s failure to retain an expert witness, ask for a continuance, or move to bar Ogryzek’s testimony due to untimely disclosure fell below the “objective standard of reasonableness” required by Strickland, 466 U.S. at 688, 104 S.Ct. 2052. Even if defense counsel could have initially believed expert testimony unnecessary, the State’s indication that it was shifting its position on the location of the gunshots would have alerted any reasonable attorney to the need to rebut with a defense expert. Though we often defer to an attorney’s calculated decision to forgo a certain trial strategy, it is undisputed that there was no strategic rationale underlying these errors. Stockton testified that the idea of securing an expert witness “never crossed my mind.” Further, it was objectively unreasonable for defense counsel to concede to the late disclosure of Ogryzek’s new theory that the shots could not have been fired by Marcia leaning across the bar. This newly disclosed theory did not simply involve a minor or collateral detail; it went to the heart of whether Martin’s version of the shootings was physically possible. We *424 have no doubt that as a trial approaches, many attorneys feel a strong incentive to proceed with the case, having prepared witnesses, evidence, and argument for the scheduled date. But it was inappropriate here for defense counsel to simply submit to an untimely disclosure of expert opinion when it blew a gaping hole in the defendant’s theory of the case. The Illinois courts have noted that “[t]he goal of discovery, of course, is to eliminate surprise and unfairness and afford opportunity to investigate; and sanctions in aid of that purpose are to compel compliance with discovery orders.” People v. Nelson, 92 Ill.App.3d 35, 44, 47 Ill.Dec. 683, 415 N.E.2d 688, 696 (1980). Though attorneys can sometimes respond to certain late disclosures up to the day of trial, this was not a scenario where Martin’s attorney could simply wing it as Ogryzek’s planned testimony became increasingly damaging. See Stanley v. Bartley, 465 F.3d 810, 812 (7th Cir.2006). This does not mean that defendants enjoy an automatic entitlement to expert rebuttal witnesses whenever the government offers expert testimony in a trial. See United States v. Anderson, 61 F.3d 1290, 1298-99 (7th Cir.1995). Particularly when the State’s theory would be very difficult to controvert with a defense expert, it may be reasonable to rely on cross-examination to cast general doubt on the government’s version of events. See Harrington, 131 S.Ct. 770 at 791. But here, unlike in Harrington, the State made clear from the start that it would present forensic evidence on the shooter’s location at the crime scene. And there were significant holes in Ogryzek’s conclusions that required expert illustration by the defense in order for the jury to weigh the evidence fairly. Cf. Showers v. Beard, 635 F.3d 625, 630 (3d Cir.2011) (distinguishing Harrington). Ogryzek testified that Martin’s account was impossible because the final shot could not have been fired from over the bar. But Busch’s expert analysis later showed that it was demonstrably possible for Marcia to have fired the final shot. At trial, defense counsel vainly sought to extract concessions from Ogryzek through cross-examination that Baldwin’s head may have moved such that the source of the final gunshot would be consistent with Martin’s account. Ogryzek repeatedly denied any alternative explanations. Busch’s post-conviction testimony demonstrates that these alternatives were indeed feasible and, in Busch’s view, more consistent with the evidence at the crime scene. Without a countering defense witness, Ogryzek’s denials in the face of cross-examination only reconfirmed the one-sidedness of the expert opinion before the jury. We take pains not to rely on the “harsh light of hindsight” in judging counsel’s performance in a particular case. Bell v. Cone, 535 U.S. 685, 702, 122 S.Ct. 1843, 152 L.Ed.2d 914 (2002). The “failure to investigate a particular lead may be excused if a lawyer has made a ‘reasonable decision that makes particular investigations unnecessary.’ ” Washington v. Smith, 219 F.3d 620, 631 (7th Cir.2000) (quoting Strickland, 466 U.S. at 691, 104 S.Ct. 2052). But we can perceive no strategic reason why the importance of expert testimony would not have been apparent at the time of trial. See Earls v. McCaughtry, 379 F.3d 489, 494 (7th Cir.2004). Indeed, defense counsel admitted that his failure to obtain an expert was an oversight. Though an inadvertent omission will not always result in constitutionally deficient performance, the failure to conduct a reasonable investigation may. Harris v. Cotton, 365 F.3d 552, 555-56 (7th Cir.2004). Here, defense counsel could not adequately represent his client simply by *425 cross-examining the State’s expert. See Miller v. Anderson, 255 F.3d 455, 457 (7th Cir.2001) (“[C]ross-examination alone could weaken the prosecution’s expert evidence, but not to the point of denying it the essential corroborative value for which the prosecutor was using it.”), judgment modified, 268 F.3d 485 (7th Cir.2001); cf. Stevens v. McBride, 489 F.3d 883, 896 (7th Cir.2007) (finding ineffectiveness due to failure to investigate expert “[w]here an expert witness’s opinion is crucial to the defense theory” (internal quotation marks and citation omitted)). In an effort to rebut Ogryzek’s testimony, defense counsel showed the jury during closing arguments an impromptu diagram he had scrawled on a legal pad. But this was not a case where such ad hoc efforts could adequately discharge counsel’s duty to the defendant under the Sixth Amendment. Counsel’s failures were particularly glaring because this was a capital case. ABA Standard 4-1.2(c) states that “[sjince the death penalty differs from other criminal penalties in its finality, defense counsel in a capital case should respond to this difference by making extraordinary efforts on behalf of the accused.” ABA Standards for Criminal Justice Prosecution Function and Defense Function 120 (3d ed.1993). B. The State Court Reasonably Determined that Martin Suffered No Prejudice From Defense Counsel’s Errors. Though it declined to reach the performance prong, the state court affirmatively held that Martin had not established prejudice under Strickland. Therefore, we must evaluate this prong under AEDPA’s deferential standard. See Harrington, 131 S.Ct. at 787-88. Under AED-PA, we “allow[] the state court’s conclusion to stand if it is one of several equally plausible outcomes.” Hall v. Washington, 106 F.3d 742, 749 (7th Cir.1997). Strickland requires Martin to demonstrate “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different.” 466 U.S. at 694, 104 S.Ct. 2052. “A reasonable probability is a probability sufficient to undermine confidence in the outcome.” Id. As a preliminary matter, we agree with Martin that Ogryzek’s opinion was highly damaging. Of the witnesses introduced during the State’s case-in-chief, prosecutors spent the greatest amount of time examining Ogryzek. Ogryzek testified at trial that Martin’s account was physically “impossible,” a refrain that the State repeated in closing arguments. This unrebutted expert opinion was unjustly inculpatory because it resulted from ineffectiveness. Busch’s analysis suggests possible critical flaws in Ogryzek’s opinion that were left untested at trial. First, it was demonstrably possible for the final shot to have been fired at Baldwin from over the bar. Second, the bullet found in the bar’s east door powerfully refutes Ogryzek’s theory that the shots originated from the bar’s south wall. If the shooter had been to the south of the victims and firing north, one of the bullets would have had to turn sharply to the right in order to end up in the east door. In addition, Turley’s gunshot wound had stippling which is consistent with a close-range shot from the area around the barstools. And the absence of stippling on Baldwin’s body supports a final shot from across the bar rather than from a person standing over the body. 2 *426 Beyond physical evidence, the circumstances of the State’s changed opinion are frankly suspicious. Investigators originally determined that the shots came from the area of the bar stools, consistent with Martin’s initial confession. When defense counsel disclosed Martin’s planned defense — that he was in the bathroom at the time of the murders — the State changed its theory to follow him to this part of the bar. In light of the convenient timing and contradictory physical evidence, Ogryzek’s opinion seems tailored to eliminate Martin’s proposed alibi, rather than deriving from an objective analysis of the crime scene. Due to ineffectiveness of counsel, Martin did not have an adequate opportunity to impeach Ogryzek in this regard. In many cases, this would be sufficient to support a finding of prejudice under Strickland. But we cannot view even a serious error in isolation. The materiality of omitted evidence helpful to a defendant “must be evaluated in the context of the entire record.” United States v. Agurs, 427 U.S. 97, 112, 96 S.Ct. 2392, 49 L.Ed.2d 342 (1976); see also Strickland, 466 U.S. at 694, 104 S.Ct. 2052 (citing Agurs). Looking at the context of the full record here, we observe that there were multiple, independent sources of evidence inculpating Martin that have nothing to do with Ogryzek’s damaging testimony. Martin contends that his counsel’s ineffectiveness prejudiced him because a defense expert would have corroborated his trial testimony that he was not the shooter. This argument misapprehends the character of the expert evidence. A crime scene investigator testifying for the defense could only have helped establish the location of the shooter, not his or her identity. Because Marcia invoked her right not to testify and there were no other witnesses, either spouse could have been in any part of the bar during the murders. 3 And if he were guilty, Martin would have had an obvious motive to put his wife in precisely the same location he was in when the gunshots were fired. The Illinois Appellate Court correctly recognized this fact and concluded, “[t]he key to the case was not where the gunman stood.” Woolley’s initial problem is that the state court found (1) the proposed expert testimony that is the basis for his collateral attack “unconvincing”; and (2) the possibility of a “so called ‘battle of the experts’ ” to have had no reasonable probability of changing the outcome of the trial given that “petitioner’s credibility was illogical, inconsistent, and impeached in numerous ways,” including by “the testimony of at least eight other witnesses.” State court findings, including credibility determinations, are presumed correct on federal *427 habeas review, unless the petitioner rebuts those findings with “clear and convincing evidence.” 28 U.S.C. § 2254(e)(1). Deference is due to “observe the verbal and nonverbal behavior of the witnesses focusing on the subject’s reactions and responses to the interrogatories, their facial expressions, attitudes, tone of voice, eye contact, posture and body movements, as well as confused or nervous speech patterns in contrast with merely looking at the cold pages of an appellate record.” Murrell v. Frank, 332 F.3d 1102, 1112 (7th Cir.2003) (citation, emphasis, and internal quotation marks omitted). Since the Supreme Court has discouraged federal courts from relitigating cases on habeas review, Woolley’s prejudice claim — based on his counsel’s failure to present expert testimony found unconvincing by the state court — appears barred from federal review, especially when it would require that this court second guess the judge’s weighing of evidence and credibility determinations of both side’s experts and the other witnesses, particularly the defendant. Assuming this court were free to consider proposed testimony from a defense expert credible and due equal weight, the question of prejudice may be a closer one, but the outcome is no different. Certainly, Ogryzek’s unrebutted opinion was damaging because it suggested that Martin must be lying when he said the final shot came from over the bar. A defense expert would have assisted Martin by impeaching Ogryzek and potentially neutralizing his damaging testimony. But the omitted evidence is not exculpatory in the manner Martin wishes it were. It would only have leveled the playing field by making Martin’s account of the murders possible rather than “impossible.” Jurors would have had to find Martin credible independent of any expert testimony. But viewing this trial record as a whole, we find numerous reasons to question Martin’s credibility. 4 First, a disinterested witness contradicted Martin’s testimony that the murder weapon was not on his person. It is undisputed that Martin’s gun was used to kill the victims. But Martin said he had moved the gun from the small of his back to his coat pocket when he left the bar between 4:00 and 5:00 p.m. to feed Marcia’s children at home. Later that night, according to Martin, Marcia took the weapon from the pocket after he left his coat on the bar stool to go to the bathroom. The timing of this account conflicts with Deborah Brose’s testimony. Brose said she saw the imprint of the gun in the back of Martin’s pants while the two played pool and Martin leaned over the table. The pool game occurred between 8:00 and 8:30 p.m., a little over an hour before the murders. Perhaps Brose was mistaken, but she mentioned this fact to police unprompted when questioned the following afternoon. And what reason would Brose have had to testify falsely? Potentially Martin could have moved the gun to his coat pocket after Brose left the bar. But then why would Martin have lied about when he did it? Martin offers no explanation for these inconsistencies. Brose was not the only disinterested witness Martin contradicted. The State *428 offered evidence of Martin’s motive from violent, inculpatory statements Martin allegedly made. Peter Dolieslager, Jeff Ince, and David Aldred testified of separate occasions on which Martin told them he could gun down patrons in public restaurants without compunction, that police in Kewanee were an incompetent “circus,” and that it would be easy for a person to commit a crime if he killed all of the witnesses. Additionally, Aldred said that two months before the murders Martin told him of financial problems he was having and money he needed to pay off large tax debts on his tattoo shop. These statements do not prove that Martin committed the murders, though they are obviously damaging. Martin might have argued that he was blowing off very badly timed steam. But he did not take this tack at trial. He testified instead that he never made any of the statements and that all three were lying. Martin has offered no explanation as to why these three apparently disinterested witnesses, including a self-described friend of the defendant, would combine to frame him. Tomsha provided the most directly incriminating evidence. He said Martin made an additional, sincere confession to him while the two were in pretrial detention. Martin contends that this testimony cannot be trusted because it comes from a jailhouse informant. Tomsha was indeed an interested witness, having received a light sentence on pending burglary charges after cooperating with the State. But Martin ignores the key corroborating element in Tomsha’s account. According to Tomsha, Martin said that before he killed the victims, he asked Baldwin to turn the television volume to its highest level. The purpose? To mask the sound of Martin’s impending gunshots. Dolieslager was the first person at the scene of the crime and said he immediately noticed the television playing at “full blast.” This evidence independently supports Tomsha’s testimony even if it might otherwise be suspect. Conceivably, Tomsha could have learned of this crime scene detail from another source and fabricated his story. But Martin has never accounted for the television volume witnessed by Dolieslager. He has never argued, for instance, that Marcia asked for the volume to be raised. Tomsha also provided written copies of confessions (and amendments to the confessions) purportedly made by Martin. When Martin testified, he denied communicating with Tomsha about the murders or composing the documents. He claimed that Tomsha must have forged the confessions from the case files Martin kept in his cell. These confession “drafts” are certainly unusual and potentially suspect coming from a jailhouse informant. But here again there is independent evidence corroborating Martin’s authorship. The State’s handwriting expert testified that Martin’s signature matched each of the documents. 5 Martin has never proposed or offered a potential defense expert to rebut the FBI agent’s conclusions regarding his handwriting. Furthermore, a spelling mistake in the written confessions supports Tomsha’s testimony. Martin prepared a letter for his lawyer to release to the press claiming that Marcia had committed the murders *429 while he was in the bathroom. The press release contained an error: Martin had misspelled Baldwin’s first name as “Renee” rather than “Rane.” The documents Tomsha delivered to police contained detailed information about the crime scene likely obtained only from either personal observation or from police reports. And each of the contested confessions also contained the same spelling error as Martin’s press release, even though the police records used the correct spelling of Baldwin’s name. Finally, Tomsha delivered the documents to authorities before Martin’s press release was dated or published. If Tomsha had forged these confessions using the police reports, how would he have known to mirror the spelling Martin would use in the later press release? The ineffectiveness of Martin’s defense counsel resulted in a significant trial error. But it is difficult to conceive of a defense that would have overcome the State’s remaining evidence. Martin has proposed none. Instead, he relies entirely on the potential effect of Busch’s testimony. As explained above, the omitted defense theory cannot carry the burden Martin desires. If the State had withdrawn Ogryzek’s testimony completely and stipulated to Busch’s account of the crime scene, a jury would still have had to contend with the overwhelming remainder of the State’s evidence. Under such circumstances, we cannot conclude that the state appellate court acted irrationally in finding “no reasonable probability that the omitted evidence would have changed the” outcome. Strideland, 466 U.S. at 700, 104 S.Ct. 2052; see also id. at 696, 104 S.Ct. 2052 (“[A] verdict or conclusion only weakly supported by the record is more likely to have been affected by errors than one with overwhelming record support.”). We observe that even defendants with weak cases deserve vigorous, effective assistance of counsel. The error in this case is troubling. It highlights the difficulty of evaluating inadequate performance when a defendant’s case is tenuous. Cf. Strickland, 466 U.S. at 710-11, 104 S.Ct. 2052 (Marshall, J., dissenting) (“Seemingly impregnable cases can sometimes be dismantled by good defense counsel.... A proceeding in which the defendant does not receive meaningful assistance in meeting the forces of the State does not, in my opinion, constitute due process.”). Nevertheless, a writ of habeas corpus is not a remedy the federal courts have authority to provide in circumstances such as these. See Knowles v. Mirzayance, 556 U.S. 111, 123, 129 S.Ct. 1411, 173 L.Ed.2d 251 (2009) (“The question ‘is not whether a federal court believes the state court’s determination’ under Strickland ‘was incorrect but whether [it] was unreasonable — a substantially higher threshold.’ ”) (quoting Schriro v. Landrigan, 550 U.S. 465, 473, 127 S.Ct. 1933, 167 L.Ed.2d 836 (2007)). III. CONCLUSION The determination that Martin Woolley was not prejudiced by his counsel’s performance at trial was not an unreasonable application of Strickland. We therefore Affirm the district court’s judgment. Footnotes [1] See also Barker, 423 F.3d at 1093 ("[T]he Supreme Court describes AEDPA review as applying to a single state court decision, not to some amalgamation of multiple state court decisions.... In Williams [v. Taylor, 529 U.S. 362, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000) ], the Virginia Supreme Court applied a standard that was contrary to federal law while the trial court applied the correct standard. The Supreme Court did not aggregate the two state court decisions or engage in 'collective review.’ Instead, it reviewed only the Virginia Supreme Court decision and held that it was contrary to federal law.” (citations omitted)). [2] The only evidence supporting Ogryzek’s version is the fact that the gunshot casings were *426 found near the bar's south wall. But this evidence has weak probative value because Ogryzek and Busch both agreed that casings ejected from a firearm do not travel in a predictable manner. [3] Busch testified that Turley’s gunshot wound was at an upward angle, indicating that she was likely shot from a position beneath her head. Martin contends that this demonstrates Marcia was the shooter since, at 5 foot, 3 inches, she was shorter than Turley and would have fired upward at her head. At over 6 feet, Martin was taller than both women. This evidence does indeed assist the defense by undercutting Ogryzek’s theory that Martin fired at Turley while standing from a distance. But the bullet trajectory cannot prove quite as much as Martin desires. First, there is no way to know whether the shooter fired from a seated or standing position. In his initial confession to police, Martin said he shot Turley while he was sitting, which would likely be consistent with the upward angle. Furthermore, Martin’s attorney was conscious of this discrepancy and highlighted it to the jury in closing arguments. [4] The Illinois Appellate Court focused on the fact that Martin had confessed then recanted, which permitted the jury to believe either his trial testimony or the prior admission. Of course it is true that jurors were free to credit the initial confession over the later recantation. But this is not the strongest basis for a finding of no prejudice since a false confession would have been consistent with the defense theory that Martin acted to protect his wife. And contrary to the State's intimations, there is no rule barring a finding of Strickland prejudice where a defendant has recanted a confession. See Soffar v. Dretke, 368 F.3d 441, 478 (5th Cir.2004); Baylor v. Estelle, 94 F.3d 1321, 1324-25 (9th Cir. 1996). [5] The expert could not definitively conclude whether Martin authored the body of the documents. But even this fact does little to help Martin under the circumstances. The expert testified that Martin intentionally refused to provide a natural handwriting exemplar. A jury could reasonably infer from this testimony alone that Martin sought to defeat the expert’s handwriting identification because he knew it would inculpate him. Again, Martin has proposed no defense on this point.