Collinsv.Sec'y of the Pa. Dep't of Corr.

742 F.3d 528 (3d Cir. 2014)

Cases citing this document

How cited

  • Woodson v. Delbalso

    Preiser v. Rodriguez, 411 U.S. 475, 499 (1973). Recently, the Third Circuit in Collins v. Sect. of PA Dept.…

  • Xavier v. Harlow

    When faced with a petition seeking habeas relief from a conviction after trial, the Third Circuit has stated…

124 Citing cases

Summaries written by judges


  • holding that deference under AEDPA "applies to any claim that has been adjudicated on the merits in any state court proceeding"

    Summary of this case from Nelson v. Overmyer

  • holding that Collins procedurally defaulted his claim that cumulative errors deprived him of due process where the "cumulative error claim was not presented to the Pennsylvania Supreme Court as an individual claim for relief"

    Summary of this case from Payne v. Bickell

  • holding cumulative error claim is a stand-alone constitutional claim subject to procedural default and exhaustion

    Summary of this case from Brown v. Bickell

No. 12–3472.


Rodney COLLINS, Appellant v. SECRETARY OF the PENNSYLVANIA DEPARTMENT OF CORRECTIONS; Michael Wenerowicz; District Attorney of the County of Philadelphia; Attorney General of the State of Pennsylvania.

Kimberly M. Dolan, [argued], Regional Housing Legal Services, Glenside, PA, Counsel for Appellant. Molly S. Lorber, [argued], Philadelphia County Office of District Attorney, Philadelphia, PA, Counsel for Appellees.

Kimberly M. Dolan, [argued], Regional Housing Legal Services, Glenside, PA, Counsel for Appellant. Molly S. Lorber, [argued], Philadelphia County Office of District Attorney, Philadelphia, PA, Counsel for Appellees.
Before: RENDELL, JORDAN and LIPEZ , Circuit Judges.

Honorable Kermit V. Lipez, United States Court of Appeals Senior Judge for the First Circuit, sitting by designation.


JORDAN, Circuit Judge.

Rodney Collins, a Pennsylvania prisoner convicted in 1993 of first-degree murder, appeals from an order of the United States District Court for the Eastern District of Pennsylvania denying his petition for a writ of habeas corpus. Although the District Court denied his petition, it certified two questions for appeal: whether Collins was deprived of his Sixth Amendment right to effective assistance of counsel because his trial counsel “inadequately prepared for trial and completely failed to conduct any investigation, including into the ballistics evidence” (J.A. at A0004), and whether trial counsel's alleged ineffective assistance, combined with alleged errors of the trial court, cumulatively caused him prejudice. Despite serious doubt that trial counsel conducted an adequate investigation, we conclude that, given the uncontroverted evidence presented against Collins at trial, the state court determination that Collins failed to show he suffered prejudice was not an unreasonable application of the Supreme Court's decision in Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), which sets forth the standard for ineffective-assistance-of-counsel claims. See28 U.S.C. § 2254(d)(1). Collins also has not exhausted his claim of cumulative error, which is therefore procedurally defaulted and not properly before us. Consequently, we will affirm the District Court's ruling denying his habeas corpus petition.

I. Factual Background and Procedural History

A. Factual Background

In the summer of 1992, a feud developed between, on one side, Collins, Andre Graves, and Kevin Cofer and, on the other, a West Philadelphia gang known as the “Boys from the Bottom.” On the night of July 12th of that year, Collins went to his girlfriend's house and reported that the Boys from the Bottom were going to kill Graves, whom the gang members had recently beaten. Collins told Graves of the threat, and Graves and Cofer then joined Collins in driving around Philadelphia in a station wagon, looking for the Boys from the Bottom. Cofer drove, Graves sat in the front passenger seat, and Collins sat in the backseat. After searching for several hours, they eventually returned to the neighborhood where they started, about a block from Collins's girlfriend's house. At that point, according to Cofer, while the three were still in the car, Collins suddenly drew a gun and shot Graves.

Early the next morning, police found Graves's body in the front seat of the station wagon with gunshot wounds as the apparent cause of death. Among the wounds were two bullet holes in Graves's head, with exit wounds under his right eye and near his right ear. A bullet had also grazed his skull. Police found bullet casings in the rear passenger compartment, under the driver's seat, and in the street behind the car. They also found two bullets in the passenger-side dashboard and in the passenger door, and two other bullets in the porches of nearby houses, as well as a fragment of a bullet in the street next to the car. Forensic testing established that all of the bullets were fired from the same .45 caliber gun.

Homicide detectives interviewed Cofer, who told them that he saw Collins shoot Graves at point blank range from the backseat of the station wagon. Cofer said that, after the shooting, he followed as Collins ran to Collins's girlfriend's house. Once there, Collins told her that someone had driven by the station wagon and shot Graves. Cofer also said that he later returned to the station wagon to retrieve the car keys and a shotgun.

Collins was eventually arrested for Graves's murder, and, in May 1993, he was tried and convicted of that crime. His trial counsel was Louis Savino. At the trial, the Commonwealth presented eye-witness testimony from Cofer; ballistics testimony from Police Officer John Finor and a chemist named Ronald McCoy; testimony from a medical examiner, Dr. Gregory McDonald, regarding the physical evidence from Graves's body; and other testimony bearing on the events surrounding the murder. Collins testified in his own defense. He told the jury that, on the day in question, after he, Graves, and Cofer had searched the neighborhood, Cofer dropped him off and, while walking to his girlfriend's house, he heard gunshots. Collins stated that, after he heard the shots, Cofer ran up behind him and told him that some people “just got finished dumping on us,” i.e., shooting at the car. (J.A. at A0828.) Based on Collins's testimony and proposed inferences from the evidence, Savino argued to the jury that the shots came from outside the car. He later described his trial strategy as attempting to create reasonable doubt by casting suspicion alternatively on the Boys from the Bottom and on Cofer as the possible shooters.

During post-conviction proceedings, Savino testified he had argued that “Kevin Cofer ... is the one who killed [Graves] by shooting from outside of the car” (J.A. at A1231), and that he also “stated ... with a lot of emphasis that the [B]oys from the [B]ottom might have been involved” ( id. at A1245). The Pennsylvania Supreme Court characterized Savino's trial strategy as “choosing to cast suspicion simultaneously on both Cofer and the supposed Boys from the Bottom.” Commonwealth v. Collins, 598 Pa. 397, 957 A.2d 237, 249 (2008). That court described Savino's cross-examination of the Commonwealth's experts as “elicit[ing] testimony ... that the shots could have been fired by Kevin Cofer from the driver's seat,” id.; however, we find no evidence in the record that Savino varied from his insistence that the shots were fired from outside the car.

1. Ballistics Testing

As do the parties, we use the term “ballistics testing” broadly to include not only testing and conclusions about bullet trajectories and identification but also about chemical testing and related evidence.

Much of the argument in this and earlier iterations of Collins's battle for post-conviction relief has centered on the trial court's admission of “last-minute” testing on the front passenger-seat headrest. (Appellant's Opening Br. at 6.) Two days before trial, the Commonwealth told Savino that McCoy had conducted additional tests on the headrest, that the results were positive for lead residue from gun powder, and that McCoy would identify the residue and its implications, while Officer Finor would opine on ballistics conclusions that could be drawn from the testing. During jury selection, Savino learned that the tests showed a particular pattern of lead residue on the headrest, but he did not see the actual test results until the trial had begun.

On the first day of trial, Savino informed the court that he had recently been notified of the testing but had not seen a report. Savino said that the testing was a “complete surprise” and that the report “could be crucial in light of the case the shooter might have been outside of the car as compared to being inside of the car.” (J.A. at A0265.) Regarding the new evidence about the headrest, Savino also told the court “I am sure it will require great investigation on my part and possibly some work with experts to see if the tests are accurate.” ( Id. at A0266.) The court noted that “[Savino] might as well get rolling on an expert now.” ( Id.) Savino never consulted an expert.

When the Commonwealth provided the testing report to Savino, it showed a lead residue pattern on the passenger-seat headrest “traversing from the left side and front to the middle of the headrest.” ( Id. at A0508–09.) Savino moved to exclude the report, but the court denied the motion. He alternatively requested “a reasonable period to conduct whatever testing that we can do to try to refute” the evidence, but that was likewise denied. ( Id. at A0509–11.) The court entered into a discussion with Savino about ballistics testing for the defense, at one point asking “[w]hat kind of test do you want to make?” (J.A. at A510.) Savino responded, “I don't know what kind of tests. The Commonwealth puts the defense in a very difficult position.” ( Id.) Savino went on to argue that “[t]he question remains whether someone was in the car doing the shooting or outside the car,” to which the court confusingly responded “[i]t would be lead whether it came from the outside or the inside. The bullets were in the inside and that is where the lead was.” ( Id. at A510–11.)

At that point, the Commonwealth also told Savino that McCoy and Finor would testify about the lead residue and corresponding ballistics conclusions and that Savino could talk to Finor before the testimony. Savino, however, did not interview Finor. In fact, he did not interview any of the Commonwealth's witnesses. He later said, “I interviewed no Commonwealth witnesses in Mr. Collins'[s] case, nor do I, with few exceptions, ever interview Commonwealth witnesses in homicides.” (J.A. at A1236.)

As promised by the Commonwealth, Finor took the stand. Relying on the headrest testing, as well as ballistics tests he performed that very morning, he testified that the murder weapon was fired no more than eighteen inches from the headrest. He also testified that the presence and shape of the bullet holes in nearby porches and in the dashboard, as well as the hole in the passenger door, were consistent with shots fired from behind Graves. McCoy testified in conformity with his report regarding the lead residue on the headrest.

2. Cofer's Testimony

Although Cofer was the Commonwealth's key witness, at trial he recanted his prior statements and instead claimed that, when initially interviewed by the police and at the pretrial hearing, he had lied about being present when Graves was shot because the police threatened to charge him with the homicide and told him that Collins had implicated him in the shooting. The prosecution thus decided to read into the record Cofer's prior statements to police and his preliminary hearing testimony. The prosecutor went through every line of the prior statements, asking Cofer if he gave a specific answer, to which Cofer answered yes, and then asking if the answer given was true, to which Cofer answered no. By this process, the prosecution highlighted details of Cofer's original testimony that would have been difficult to fabricate.

In addition, the prosecutor was able to point out that Cofer had earlier told detectives of Collins's motive for the murder. According to Cofer, Collins had explained to him that he killed Graves because “[h]e could have got us rocked,” i.e., killed, in the feud with the Boys from the Bottom. (J.A. at A0307 (internal quotation marks omitted).)

On the second day of his testimony, Cofer attempted to invoke his Fifth Amendment right against self-incrimination. Retaking the witness stand, after a weekend recess, Cofer stated, “[a]t this point, I will invoke my Fifth Amendment Right and plead the Fifth.” (J.A. at A0390.) The prosecutor asked whether he had been threatened over the weekend and, after saying that he had not, Cofer continued to answer questions. Savino objected, stating his concern that Cofer had invoked the Fifth Amendment. To which the court replied, “[h]ow can you invoke a Fifth Amendment privilege when you have been spilling it out for the last two days?” ( Id. at A0393.) After more testimony, Savino again raised his concern with the court, stating, “if this witness does invoke his Fifth Amendment privilege now,” there would be “no opportunity to cross examine him.” ( Id. at A0412.) As it turned out, however, Savino was able to cross-examine Cofer and to elicit the admission that Cofer had originally told detectives he was not present at the shooting but later changed his story. In addition, the cross-examination established that Cofer had known Graves longer than Cofer had known Collins, and that Cofer was closer to Graves than to Collins. Savino also questioned Cofer regarding his possible implication in the homicide. Although Cofer was clearly a crucial witness, Savino did not try to interview him before the cross-examination.

To discredit Cofer's changed testimony, the Commonwealth introduced two letters, which were given to the prosecutor during a lunch break on the second day of the trial by a woman identifying herself as Cofer's mother. The letters were purportedly from Collins and asked Cofer not to testify against him. On the stand, Cofer refused to read the letters, so the prosecutor read them into the record over Savino's objection. The first letter began:

What's up, Kev. (Kabir). In an effort to straighten out this (lie) that's about to cost an innocent man his life, the necessary steps must be taken to clear us both. I prefer to do it this way cause (Louis Savino) is trying to turn the tables from one innocent man (me) to another.... [S]o to clear us both, you must sign and fill out this form, and return it to me cause I didn't no [sic] anything to tell the police so there I refused to work with this lawyer to try and put you on the spot.
( Id. at A0537–38.)

The letter went on to ask that Cofer swear to a statement that read, in part, “I, (Kevin Cofer) have made a grave mistake and worked along with the police ... in an effort to commit and convict an [i]nnocent man of murder.” ( Id. at A0539.) The proposed statement continued, “[a]ll these (lies) forced me into signing a false statement for fear that I would be charged with (murder).” ( Id.) When the prosecutor asked Cofer whether “everything you told this jury about your not being involved ... [and] the police setting you up and threatening you was a result of this letter,” Cofer refused to answer. ( Id. at A0539–40.) Savino objected and moved for a mistrial, neither of which succeeded. The prosecutor then read into the record the second letter, which stated in part:

Look Blood, you know from first hand experience about jail, so why is you putting me through this sucker.... I mean if you going to straighten this out, and free an innocent black man, look, I know how them devils put they thing down when it comes to these bodies but that's past and you hold the key to my release.
( Id. at A0541.)

After another objection from Savino, the prosecutor finished reading the letter, which ended with: “Look, man, I know you ain't that type of dude that want to be label a snitch and I know for real you ain't built like that, so straighten this out as soon as possible.” ( Id. at A0542.)

Savino once more objected to questioning a witness who had invoked his Fifth Amendment rights and moved for a mistrial. The court again overruled the objection.

The prosecution also requested and received a handwriting exemplar from Collins. Collins asserts that the Commonwealth failed to prove that the letters were written by him and that “[a]n analysis at trial confirmed that they were not in [his] handwriting.” (Appellant's Reply Br. at 18 n.6.) Collins, however, only directs our attention to sections of his own testimony where he denied writing the letters.

3. The Medical Examiner

Dr. Gregory McDonald, the medical examiner who performed Graves's autopsy, also testified at trial, specifically discussing the three gunshot wounds to Graves's head. McDonald found an “entrance wound in the left side of the head slightly above the left ear and slightly behind the left ear and that went through the skull ... and exit[ed] out of the right side of the head, slightly behind and slightly above the right ear.” (J.A. at A0674.) He summarized the trajectory of the shot that caused the wound as “rightward, slightly downward and forward.” ( Id. at A0676.) He testified that another bullet went from “slightly above and slightly behind the left ear ... out the right side of the face just below the right eye.” ( Id. at A0674–75.) McDonald said the trajectory of that shot was “forward, to the right, and slightly downward.” (Id. at A0678.) He also found a “graze wound of the left upper portion of the head going forward and to the right.” (Id. at A0674.) The Commonwealth argued that Graves's wounds and the fact that the bullet trajectories were primarily forward and to the right were consistent with shots fired from the backseat. On cross-examination, Savino elicited the admission that McDonald could not state exactly how far away from Graves's head the shots were fired.

4. Corroborating Witnesses

Other witnesses testified at trial that Collins had been trying to find Graves on the day of the murder; that Collins owned a black .45 caliber handgun that he was carrying when he met with Cofer and Graves; that he had been in the backseat of the car, with Cofer and Graves in front; and that he later ran into his girlfriend's apartment with blood on his knee and behaving nervously.

5. Savino's Theory of the Case

Notwithstanding the physical evidence and the expert testimony about it, Savino's theory of the case was that the shooter was outside the car because there were bullet casings found outside and behind the car and because the Boys from the Bottom had a motive to kill Graves. Savino also wanted the jury to consider that Cofer had something to hide because he immediately fled the scene of the crime when his lifelong friend, Graves, was shot. The defense closing, therefore, pursued two different themes: that the Boys from the Bottom had a motive for murder and that Cofer was an unreliable witness and perhaps was the actual shooter. Whoever the shooter was, Savino maintained, it was not Collins because the shots came from outside the car. Savino's closing argument was almost evenly split between casting suspicion on Cofer and on the Boys from the Bottom.

The jury evidently believed neither of those alternatives. It found Collins guilty, and he was sentenced to death.

B. Procedural History

1. Direct Appeal

The Pennsylvania Supreme Court affirmed Collins's conviction and sentence on direct appeal. See Commonwealth v. Collins, 549 Pa. 593, 702 A.2d 540, 541 (1997). It held that the evidence was sufficient to sustain a conviction for first-degree murder, and it rejected Collins's three claims of error, namely: (1) that the Commonwealth's tactics, specifically eliciting testimony from Cofer regarding his failure to testify in accordance with his prior statements, were inflammatory; (2) that trial counsel was ineffective for failing to object to those tactics; and (3) that trial counsel was ineffective for not requesting an alibi instruction related to Collins's testimony that he left the car before the shots. Having concluded that Collins's conviction for first-degree murder was sound, the court affirmed his death sentence. Id. at 546. The United States Supreme Court denied certiorari. Collins v. Pennsylvania, 525 U.S. 835, 119 S.Ct. 92, 142 L.Ed.2d 73 (1998).

2. Post–Conviction Relief Proceedings

Collins then collaterally attacked his conviction and sentence. See Commonwealth v. Collins, 2005 WL 6347804 (Phila.Ct.Com.Pl., Feb. 15, 2005); Commonwealth v. Collins, 598 Pa. 397, 957 A.2d 237 (2008). He filed a petition under Pennsylvania's Post–Conviction Relief Act (“PCRA”), 42 Pa. Cons.Stat. Ann. § 9541 et seq. , making 25 claims of error. The PCRA court granted a hearing as to six claims, which included the assertion that trial counsel was ineffective for failing to hire a ballistics expert. At the PCRA hearing, Savino testified that he had considered retaining a ballistics expert but decided not to, for strategic reasons. Specifically, he testified regarding his belief that, if he called an expert, the resulting report could have been discoverable by the Commonwealth and might have undermined Collins's defense. In addition, Savino testified that his strategy was to create a reasonable doubt in the jurors' minds as to where the shots came from. He said he “argued to the jury that there was no proof that Mr. Collins fired any shots from the back seat of that car. And in fact, Mr. Cofer, Kevin Cofer, the Commonwealth witness, is the one who killed the gentleman by shooting from outside of the car.” (J.A. at A1231.) Savino confirmed that he interviewed no witnesses and did not consult any experts in preparation for the guilt phase of Collins's trial.

.Pennsylvania Rule of Criminal Procedure 305 applied to pre-trial discovery at the time of Collins's trial. Under Rule 305(C)(2)(a), a court could order the production of “results or reports ... of scientific tests ... that the defendant intends to introduce as evidence in chief, or were prepared by a witness whom the defendant intends to call at the trial....” Pa. R.Crim. P. 305(C)(2)(a). Collins argues that if the expert report was not favorable for the defense, Savino never would have had to call the expert, and the report would not have been discoverable.

William Welch, a ballistics expert, testified at the PCRA hearing on behalf of Collins. His testimony focused on the direction from which the shots were fired. Welch said that Finor's ballistics testing was unreliable because the test shots had been fired directly into the test material. He also disputed the results because only one shot was fired in testing despite multiple shots having been fired in the actual murder, and because the testing was not done with the murder weapon. Welch testified that if the shooter had been sitting in the backseat of the car, it would have been awkward to reach around the headrest to fire in a way that was consistent with the physical evidence. He also stated that he would have been available to review the evidence at Collins's trial, had he been called.

On cross-examination, however, Welch conceded that the physical evidence was consistent with a shooter inside the car—either in the driver's seat or the backseat. As to physical evidence—particularly lead residue and bullet holes—counsel for the Commonwealth asked Welch if it was “consistent with a gun being fired in that car,” and Welch responded, “It is.” ( Id. at A1194.) The questioning went on:

Q: Would you agree that the physical evidence in that car is, in fact, consistent with the testimony of Kevin Cofer [implicating Collins]?

A: Physical evidence, meaning the bullet holes?

Q: Yes.

A: It is consistent, but also consistent with that of the driver firing the shots.
( Id. at A1195.)

Q: Mr. Welch, your opinion that it would come from the driver's side, you're giving us your opinion it would be awkward physically to shoot someone from the back seat, but in order to explain a direct back-to-front shot in the dashboard, you would have to opine that there was a struggle between the driver and the victim of some sort and his hand was pushed in order to fire directly into the dash board; is that correct?

A: Yes.

Q: That's not supported by any of the evidence that you've reviewed in this case; is that correct?

A: There doesn't seem to be any physical evidence to that effect. No.
( Id. at A1199–1200.)

The PCRA court dismissed all of Collins's claims except those regarding ineffective assistance of counsel at the sentencing phase. Regarding Savino's failure to obtain a ballistics expert, the PCRA court held that Savino had a “reasonable strategy for not retaining a ballistics expert and no prejudice has been shown by his failure to do so.” ( Id. at A0118.) Looking to the sentencing phase of the trial, however, the court held that Collins was entitled to relief because Savino was “ineffective for failing to investigate and present [mitigation] evidence,” and appellate counsel was ineffective for failing to raise Savino's ineffectiveness with respect to sentencing. ( Id. at A0135.) The court therefore vacated the death sentence and remanded for a new sentencing hearing. The Commonwealth decided not to seek the death penalty again, and Collins was resentenced to life in prison.

The PCRA court stated: “If the jury was provided accurate information about defendant's juvenile adjudications and available mitigation evidence, there is a reasonable probability that at least one juror would have struck a different balance and voted not to impose the death penalty.... Accordingly, trial counsel was ineffective for failing to investigate and present such evidence and appellate counsel was ineffective for failing to raise trial counsel's ineffectiveness in this regard.” (J.A. at A0135 (citations omitted).)

Collins then appealed 15 claims to the Pennsylvania Supreme Court, which affirmed the PCRA court's denial of a new trial. Collins, 957 A.2d at 243, 272. Relevant to the claims before us, the Pennsylvania Supreme Court noted that the issue of Savino's failure to present expert ballistics testimony was properly layered with his claim that appellate counsel was ineffective for failing to raise Savino's ineffectiveness. It also held that “[Savino] did not act in a constitutionally unreasonable fashion in choosing to cast suspicion simultaneously on both Cofer and the supposed Boys from the Bottom gang rather than attempt to find a competing defense ballistician to paint Cofer as the only possible perpetrator.” Id. at 249. The court determined that trial counsel made a strategic decision to not hire a ballistics expert but, instead, to create a reasonable doubt by casting suspicion on others, and, “because appellant failed to show that trial counsel lacked a reasonable strategic basis for choosing not to call a ballistics expert, his underlying claim fails.” Id. at 250. In his briefing before that court, Collins supported his claim that Savino had pursued a deficient trial strategy with, among other things, references to Savino's general lack of preparation and inadequate investigation. The Pennsylvania Supreme Court, however, did not address the lack of preparation or investigation that went into Savino's trial strategy, focusing instead on Savino's decision not to hire a ballistics expert.

Collins did not make a cumulative error claim during post-conviction proceedings. The Pennsylvania Supreme Court nevertheless reviewed his claims regarding each individual error that he now alleges cumulatively prejudiced him, and it rejected them all. See id. at 253–54, 259–60.

Collins subsequently filed in the District Court a petition for a writ of habeas corpus, pursuant to 28 U.S.C. § 2254, arguing that he is entitled to relief based on violations of his right to the effective assistance of counsel, his right to confront Cofer, and his right to due process of law. The District Court referred the case to a Magistrate Judge who, in a Report and Recommendation, recommended that the petition be denied and no certificate of appealability be issued. The Magistrate Judge determined that the Pennsylvania Supreme Court's denial of Collins's ineffectiveness claim was not an unreasonable application of federal law under 28 U.S.C. § 2254 because it was based on deference owed to Savino's strategic decisions and that Savino did not “automatically lose[ ] the benefit of a presumption of reasonable strategic decisionmaking simply because he did not consult with a ballistics expert.” (J.A. at A0021.) Further, the Magistrate Judge noted that, given Welch's concession on cross-examination that the shots could have been fired from the driver's side or the backseat of the car, there was no “reasonable probability of an acquittal” even if Savino had consulted with an expert. ( Id.) The Magistrate Judge also rejected Collins's claim of cumulative error because, he said, there were no underlying errors.

The District Court approved and adopted the Magistrate Judge's Report and Recommendation as to each claim for relief, but it nevertheless stated that trial counsel was “clearly inadequate.” ( Id. at A0003.) It thus certified two claims for appeal: (1) Collins's claim that he was deprived of his Sixth Amendment right to effective assistance of counsel “because his attorney inadequately prepared for trial and completely failed to conduct any investigation, including into the ballistics evidence”; and (2) his claim of cumulative prejudice from various errors allegedly committed at trial. ( Id. at A0003–04.) Those are the issues Collins brings to us now.

II. Procedural Default

As a preliminary matter, we must consider the Commonwealth's contention that we should affirm the District Court's denial of Collins's habeas petition on the basis of procedural default. “The doctrine of procedural default prohibits federal courts from reviewing a state court decision involving a federal question if the state court decision is based on a rule of state law that is independent of the federal question and adequate to support the judgment.” Fahy v. Horn, 516 F.3d 169, 187 (3d Cir.2008) (citing Nara v. Frank, 488 F.3d 187, 199 (3d Cir.2007)). Procedural default occurs when a state court determines that “the prisoner ... failed to meet a state procedural requirement.” Coleman v. Thompson, 501 U.S. 722, 730, 111 S.Ct. 2546, 115 L.Ed.2d 640 (1991).

The Commonwealth argues that the claims certified for appeal are not properly before us. It asserts that Collins raised only the issue of appellate counsel's ineffectiveness before the Pennsylvania Supreme Court and, therefore, his claims against trial counsel were waived and thus procedurally defaulted. We reject that argument.

Collins did raise his ineffective assistance claim as to trial counsel, and it was reviewed on the merits. The Pennsylvania Supreme Court found that the claim was properly layered with a claim of ineffective assistance from appellate counsel, consistent with Commonwealth v. Grant, 572 Pa. 48, 813 A.2d 726, 733 (2002). The court stated that, “[b]ecause [Collins] was represented by new counsel on direct appeal, and his appeal was pending on collateral review prior to our decision in [Grant ], these [ineffective assistance of trial counsel] claims are cognizable only as ‘layered claims.’ ” Collins, 957 A.2d at 244. Under a layered-claim analysis, Collins had to plead and prove that trial counsel was ineffective. We have held that “[u]nder Pennsylvania law, where ineffectiveness claims are properly layered, there is no waiver and no procedural default.” Showers v. Beard, 635 F.3d 625, 629 (3d Cir.2011) (rejecting Commonwealth's procedural default claims and finding them waived; proceeding to the merits where the state court addressed both effectiveness claims on the merits). Here, “the PCRA court and the [Supreme Court] addressed the ineffectiveness claims against both trial and appellate counsel on the merits.” Id. at 629 n. 4. Likewise, the Magistrate Judge viewed the ineffectiveness of trial counsel as the main issue. Therefore, the Commonwealth's procedural default arguments regarding ineffective assistance of counsel are unpersuasive.

The Commonwealth also contends that Collins only raised claims of ineffective assistance of appellate counsel before the District Court and raises claims of ineffective assistance of trial counsel for the first time on appeal. This is inaccurate, as Collins's habeas petition clearly indicates. See Petition for Habeas Corpus at 9, Collins v. Beard, No. 10–05950, 2012 WL 3136768 (E.D.Pa. Aug. 1, 2012) (“A. Trial Counsel Adopted a Facially Implausible Defense.”). The Commonwealth may actually have waived its procedural default argument on this point, given that, in the District Court, it did not assert that Collins's claims were waived as to everything except appellate counsel. See Showers v. Beard, 635 F.3d 625, 629 (3d Cir.2011) (citing Trest v. Cain, 522 U.S. 87, 89, 118 S.Ct. 478, 139 L.Ed.2d 444 (1997) (in habeas context, procedural default is normally a defense that the State is obligated to raise)).

As for Collins's claim of cumulative error, the Commonwealth argues that it too is defaulted because it was not raised in state court and is therefore unexhausted. Collins responds that “the cumulative error doctrine is a required method of conducting prejudice analysis,” not a standalone constitutional claim, and that the underlying errors were raised in state court and are therefore exhausted. (Appellant's Reply Br. at 21.) We have not had occasion before to hold that a cumulative error argument constitutes a standalone constitutional claim subject to exhaustion and procedural default, but, with the issue squarely presented now, we so rule. Collins's cumulative error claim was not raised before the Pennsylvania Supreme Court and is therefore not properly before us.

The Magistrate Judge concluded that Collins had presented the underlying claims to the PCRA court, but did not know if the cumulative error claim had been presented to the Pennsylvania Supreme Court. Accordingly, the Magistrate Judge recommended that the claim be dismissed on the merits.

“Before seeking a federal writ of habeas corpus, a state prisoner must exhaust available state remedies, 28 U.S.C. § 2254(b)(1), thereby giving the State the ‘opportunity to pass upon and correct alleged violations of its prisoners' federal rights.’ ” Baldwin v. Reese, 541 U.S. 27, 29, 124 S.Ct. 1347, 158 L.Ed.2d 64 (2004) (quoting Duncan v. Henry, 513 U.S. 364, 365, 115 S.Ct. 887, 130 L.Ed.2d 865 (1995) (per curiam)). The Supreme Court has instructed that a claim is not “fairly presented” if the state court “must read beyond a petition or a brief ... in order to find material” that indicates the presence of a federal claim. Id. at 32, 124 S.Ct. 1347. A claim is procedurally defaulted if the petitioner failed to exhaust that claim in state court and if state procedures prohibit the petitioner from later presenting the claim in state court. See Jimenez v. Walker, 458 F.3d 130, 149 (2d Cir.2006) (“Under the procedural-default doctrine, when a prisoner has exhausted his state remedies but has not given the state courts a fair opportunity to pass on his federal claims, the prisoner has procedurally defaulted his claims....”); Bridges v. Beard, 941 F.Supp.2d 584, 621 (E.D.Pa.2013) (discussing exhaustion and procedural default requirements).

Collins does not attempt to show either cause and prejudice for the default or a fundamental miscarriage of justice and so we do not address those exceptions to the procedural default rule. See Jimenez v. Walker, 458 F.3d 130, 149 (2d Cir.2006) (citing Murray v. Carrier, 477 U.S. 478, 495–96, 106 S.Ct. 2639, 91 L.Ed.2d 397(1986)) (holding petitioner ineligible for habeas relief on cumulative error claim when it was not exhausted and petitioner did not attempt to show cause and prejudice or a fundamental miscarriage of justice).

We thus do not agree with Collins's assertion that cumulative error is only a method of conducting prejudice review. The cumulative error doctrine allows a petitioner to present a standalone claim asserting the cumulative effect of errors at trial that so undermined the verdict as to constitute a denial of his constitutional right to due process. See Albrecht v. Horn, 485 F.3d 103, 139 (3d Cir.2007) (holding that petitioner could not show that the cumulative prejudice of trial errors “undermined the reliability of the verdict”). Specifically, we have said that

Individual errors that do not entitle a petitioner to relief may do so when combined, if cumulatively the prejudice resulting from them undermined the fundamental fairness of his trial and denied him his constitutional right to due process. Cumulative errors are not harmless if they had a substantial and injurious effect or influence in determining the jury's verdict, which means that a habeas petitioner is not entitled to relief based on cumulative errors unless he can establish actual prejudice.
Fahy v. Horn, 516 F.3d 169, 205 (3d Cir.2008) (internal quotation marks omitted) (citations omitted). Here, the Pennsylvania Supreme Court reviewed each alleged underlying error and rejected each on its merits, but it was not presented with a separate claim of cumulative error. See Collins, 957 A.2d at 243–44 (listing the fifteen claims raised).

Although the Fifth Circuit Court of Appeals held in Derden v. McNeel, 978 F.2d 1453, 1454 (5th Cir.1992) (en banc), that cumulative error relief is available so long as the individual errors were themselves not procedurally defaulted, several other circuits disagree and treat cumulative error claims as distinct claims subject to exhaustion and procedural default. See, e.g., Wooten v. Kirkland, 540 F.3d 1019, 1026 (9th Cir.2008) (“[A] cumulative error claim must be clearly identified in a petitioner's brief before a state court to be exhausted.”); Jimenez, 458 F.3d at 149 (holding that cumulative error claim must be fairly presented to state court to later be considered by federalcourts on habeas review); Keith v. Mitchell, 455 F.3d 662, 679 (6th Cir.2006) (“Because [the petitioner] did not raise his claim of cumulative error in the state courts, it is procedurally defaulted.”); Gonzales v. McKune, 279 F.3d 922, 925 (10th Cir.2002) (finding default on habeas review because state court was not asked to consider cumulative error and therefore never had opportunity to consider it); Bridges, 941 F.Supp.2d at 621 (listing cases supporting the holding that cumulative error claim, which was not presented to state court, was unexhausted and defaulted as an independent basis for habeas relief). Those decisions comport with the long-held view that in order to satisfy exhaustion, a state habeas petitioner must present the “substantial equivalent” of his federal claim to the state courts in order to give the state courts “an opportunity to apply controlling legal principles to the facts bearing upon his constitutional claim.” Picard v. Connor, 404 U.S. 270, 277–78, 92 S.Ct. 509, 30 L.Ed.2d 438 (1971). Because “[b]riefing a number of isolated errors that turn out to be insufficient to warrant reversal does not automatically require the court to consider whether the cumulative effect of the alleged errors prejudiced the petitioner,” Wooten, 540 F.3d at 1025, we now join those courts that hold that a claim of cumulative error must be presented to the state courts before it may provide a basis for habeas relief. See28 U.S.C. § 2254(b)(1)(A) (exhaustion requirement). Collins's cumulative error claim was not presented to the Pennsylvania Supreme Court as an individual claim for relief and, hence, was not exhausted in state court. It is now too late for him to return to the state courts to exhaust that claim, and it is therefore procedurally defaulted and not properly before us.

III. Standard of Review

As is generally true in habeas corpus cases, this appeal is heavily influenced by a standard of review that dictates how much deference we must give to state court rulings. That standard, a function of the Anti–Terrorism and Effective Death Penalty Act of 1996 (“AEDPA”), 28 U.S.C. § 2254(d), does not “permit federal judges to ... casually second-guess the decisions of their state-court colleagues or defense attorneys.” Burt v. Titlow, ––– U.S. ––––, 134 S.Ct. 10, 13, 187 L.Ed.2d 348 (2013). Under AEDPA, habeas relief is not available for any “claim that was adjudicated on the merits” in state court unless that adjudication either “resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established federal law, as determined by the Supreme Court,” or was founded on an “unreasonable determination of the facts.” 28 U.S.C. § 2254(d)( l )-(2). Collins contends that “[t]he state court's assertions are unreasonable applications” of federal law “and unreasonable in light of the state court record.” (Appellant's Opening Br. at 28.) We focus on the “unreasonable application” clause of § 2254(d). A state court decision is an “unreasonable application” of federal law “if the state court identifies the correct governing legal principle from [the Supreme] Court's decisions but unreasonably applies that principle to the facts of the prisoner's case.” Williams v. Taylor, 529 U.S. 362, 413, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000).

Under the strictures of the “unreasonable application” clause, a federal court may not issue a writ of habeas corpus simply because it “concludes in its independent judgment that the state-court decision applied a Supreme Court case incorrectly.” Blystone v. Horn, 664 F.3d 397, 417 (3d Cir.2011) (internal quotation marks omitted). Instead, the petitioner must “show that the state court applied that case to the facts of his case in an objectively unreasonable manner.” Id. (internal quotation marks omitted). “It bears repeating that even a strong case for relief does not mean the state court's contrary conclusion was unreasonable.” Harrington v. Richter, ––– U.S. ––––, 131 S.Ct. 770, 786, 178 L.Ed.2d 624 (2011). Section 2254(d) authorizes us to issue a writ of habeas corpus only “in cases where there is no possibility fairminded jurists could disagree that the state court's decision conflicts with th[e Supreme] Court's precedents.” Id. Under this highly deferential standard, “we will not surmise whether the state court reached the best or even the correct result in [a] case; rather, we will determine only ‘whether the state court's application of [federal law] was unreasonable.’ ” Rountree v. Balicki, 640 F.3d 530, 538 (3d Cir.2011) (quoting Richter, 131 S.Ct. at 785).

The clearly established law chiefly at issue in this case is the Supreme Court's landmark decision in Strickland v. Washington, which provides a two-pronged test for reviewing ineffective-assistance-of-counsel claims. 466 U.S. at 687–88, 104 S.Ct. 2052. A petitioner must prove both (1) that “counsel's representation fell below an objective standard of reasonableness” and (2) that petitioner was prejudiced by that subpar performance. Id. at 688, 694, 104 S.Ct. 2052. But, before examining the state courts' application of Strickland, we first have to determine whether Collins's claims were adjudicated on the merits, since the distinction between claims that have been so adjudicated and claims that have not been means the difference between highly deferential review and de novo review. Cf. Porter v. McCollum, 558 U.S. 30, 130 S.Ct. 447, 452, 175 L.Ed.2d 398 (2009) (“Because the state court did not decide whether [the prisoner's] counsel was deficient, we review this element of [his] Strickland claim de novo.”); Appel v. Horn, 250 F.3d 203, 210 (3d Cir.2001) (applying de novo review when Pennsylvania Supreme Court never considered petitioner's constructive denial of counsel claim, and instead treated claim as one of ineffective assistance of counsel).

If there has been no adjudication on the merits of a claim, “the federal habeas court must conduct a de novo review over pure legal questions and mixed questions of law and fact.” Appel v. Horn, 250 F.3d 203, 210 (3d Cir.2001). State court factual determinations are “presumed to be correct,” absent clear and convincing evidence of error. Id. (citing 28 U.S.C. § 2254(e)(1)).

Here, the Pennsylvania Supreme Court focused its decision on Savino's failure to present ballistics evidence. See Collins, 957 A.2d at 250. Collins, however, presents his Strickland claim as the broader assertion that Savino utterly failed to investigate and prepare for the case, which included, in part, the failure to present ballistics evidence. That broader claim is indeed the one that the District Court certified for appeal to us. We are therefore faced with the question of whether Collins “fairly presented” that broader claim to the state courts and if the state courts adjudicated it on the merits. We conclude that the broader claim was presented to and adjudicated on the merits by the state courts, and that the narrower focus of the Pennsylvania Supreme Court on ballistics evidence does not affect our standard of review.

Often, a question of this type is stated in terms of “exhaustion,” that is, whether the claim at issue was pursued in a manner that took advantage of all state court remedies. In looking at what claims were presented to state courts for the purposes of determining the applicable standard of review, we ask whether a court was given the opportunity to address a specific claim, regardless of whether the court actually addressed that claim. See Smith v. Digmon, 434 U.S. 332, 333–34, 98 S.Ct. 597, 54 L.Ed.2d 582 (1978) (per curiam) (holding that habeas exhaustion requirement is not dependent on whether the state court's opinion references a claim, which was raised in petitioner's brief before that court, but rather if it was properly raised).

The record reveals that Collins presented to the PCRA court, and that the court reached, both the narrow claim that Savino failed to refute the ballistics evidence and the broader claim that he did not properly investigate and prepare for trial and retain expert testimony. The PCRA court denied both claims on the merits, explicitly ruling that Collins had not established that Savino's performance was constitutionally deficient. As to the narrower claim, the PCRA court ruled that “no prejudice has been shown by [Savino's] failure to [retain a ballistics expert].” (J.A. at A0118.) The PCRA court likewise decided that Collins had not demonstrated prejudice regarding his broader claim of lack of preparation and presentation of exculpatory evidence. ( See JA. at A0106–09.) On PCRA appeal, Collins challenged the PCRA court's denial of both the narrower and the broader claims, and argued that the PCRA's findings as to counsel's representation and lack of prejudice were erroneous. The Pennsylvania Supreme Court acknowledged the PCRA court's rulings on these claims, as well as Collins's arguments, but affirmed on the ground that Savino's performance was not constitutionally deficient without ruling on prejudice. Collins, 957 A.2d at 248–50.

We are persuaded that these state court decisions resulted in an adjudication on the merits of both Collins's narrower and broader claims of ineffective assistance of counsel to which we must defer under AEDPA. Section 2254(d) deference applies to any claim that has been adjudicated on the merits in any state court proceeding, which “can occur at any level of state court” as long as the state court's resolution has preclusive effect. Thomas v. Horn, 570 F.3d 105, 117 (3d Cir.2009) (holding that under § 2254(d) a claim has been adjudicated on the merits “when a state court has made a decision that finally resolves the claim based on its substance, not on a procedural, or other, ground”). In Collins's case, there is no question that the PCRA court denied both the narrower and broader claims of ineffective assistance on the merits and expressly ruled on each of the two prongs of the Strickland test. Although in affirming the PCRA court's decision, the Pennsylvania Supreme Court discussed counsel's performance in the context of the narrower claim only, nothing in its opinion questionedor undermined the PCRA court's more specific rulings. The lack of an express ruling from the Pennsylvania Supreme Court on the question of prejudice does not negate the PCRA court's decision that Collins was not prejudiced. Collins's claim of ineffective assistance of counsel, framed both narrowly and broadly and tested under each of Strickland 's prongs, was adjudicated on the merits in state court, even if only at the PCRA court level. Therefore, under the circumstances presented in this case, the PCRA court's determination as to prejudice is owed deference under AEDPA. Cf. Bond v. Beard, 539 F.3d 256, 289 (3d Cir.2008) (reviewing the Pennsylvania Supreme Court's decision for the first prong of the Strickland analysis, but the PCRA court's ruling for the prejudice prong).

An argument can be made that the Pennsylvania Supreme Court answered the question as it was put to it. In his briefing before that court, Collins presented what may be characterized as a single ineffective-assistance-of-counsel claim with two issues as supporting arguments: that Savino was ineffective for following a deficient “outside shooter” defense theory because of his inadequate preparation and investigation; and, more pointedly, that he was ineffective for failing to consult a ballistics expert. Even though the Pennsylvania Supreme Court focused on the ballistics evidence, it ultimately held that Savino's trial strategy was not, in fact, limited to the outside shooter theory, but was more generally aimed at creating a reasonable doubt, and that he “cho[se] to cast suspicion simultaneously on both Cofer and the supposed Boys from the Bottom gang rather than attempt to find a competing defense ballistician to paint Cofer as the only possible perpetrator.” Collins, 957 A.2d at 249. Therefore, the Pennsylvania Supreme Court arguably adjudicated the claim as presented to it, determining that Savino, in fact, pursued a defense strategy to cast doubt on both Cofer and the Boys from the Bottom and rejecting Collins's initial assertion about Savino's strategy.

We recognize that our approach in this case may seem at odds with the approach approved by the Seventh Circuit in Woolley v. Rednour, 702 F.3d 411, 421–22 (7th Cir.2012), cert. denied, Woolley v. Harrington, ––– U.S. ––––, 134 S.Ct. 95, 187 L.Ed.2d 33 (2013). In Woolley, the Seventh Circuit applied de novo review where the opinion of the state's appellate court was “silent on defense counsel's performance,” even though the state PCRA court had expressly ruled on both prongs of the Strickland test. Id. at 422. In the Seventh Circuit's view, the opinion of the state's appellate court was the “last reasoned opinion on the claim,” and thus the only decision entitled to AEDPA deference. Id. at 421–22. We think our approach is the better course in reviewing Collins's claims of ineffective assistance of counsel. Our approach is especially appropriate in light of the United States Supreme Court's repeated admonitions that AEDPA mandates broad deference to the decisions of the state courts. See, e.g., Titlow, 134 S.Ct. at 16 (“We will not lightly conclude that a State's criminal justice system has experienced the extreme malfunction for which federal habeas relief is the remedy.” (internal quotation marks omitted)); Richter, 131 S.Ct. at 786–88. In any event, even if we were to apply de novo review to the PCRA court's prejudice rulings in Collins's case, the result would be the same for the reasons discussed infra Part IV. See Richter, 131 S.Ct. at 787–88 (stating that even under de novo review, prejudice is established only where counsel's errors are “so serious as to deprive the defendant of a fair trial, a trial whose result is reliable.”).

IV. Discussion

The District Court had jurisdiction over Collins's petition pursuant to 28 U.S.C. §§ 2241 and 2254. We have appellate jurisdiction pursuant to 28 U.S.C. §§ 1291 and 2253. The District Court did not conduct an evidentiary hearing, so our review of that Court's decision is plenary. Blystone v. Horn, 664 F.3d 397, 416 (3d Cir.2011).

To summarize, we must decide whether, viewed with the deference required by AEDPA, the state courts' denial of Collins's ineffective-assistance-of-counsel claim was based on an “unreasonable application” of the Supreme Court's decision in Strickland.

Turning to the merits, we begin with the center of gravity in this case: the claim of ineffective assistance of counsel. As previously noted, Strickland provides a two-pronged test for reviewing such claims. 466 U.S. at 687–88, 104 S.Ct. 2052. First, the petitioner must show that “counsel's representation fell below an objective standard of reasonableness” under “prevailing professional norms.” Id. at 688, 104 S.Ct. 2052. Second, the petitioner must show prejudice such 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, 104 S.Ct. 2052. On habeas review, we are “doubly deferential” in considering counsel's performance: the state court was obligated on post-conviction review to view that performance deferentially, and, under AEDPA, we must give wide deference to the state court's conclusions, disturbing them only if the state court unreasonably applied either of the prongs of Strickland. Titlow, 134 S.Ct. at 13;Knowles v. Mirzayance, 556 U.S. 111, 123, 129 S.Ct. 1411, 173 L.Ed.2d 251 (2009).

While we believe that the appropriate route to resolving this appeal is to consider whether the PCRA court correctly applied the prejudice prong—which we think it did in concluding that Collins was not prejudiced by Savino's representation—we would be remiss if we did not comment briefly on trial counsel's dismal preparation for a high stakes case. Under Strickland 's objective-reasonableness prong, the question of whether counsel had a reasonable strategy necessarily includes the question of whether that strategy was reasonably arrived at. See Strickland, 466 U.S. at 690–91, 104 S.Ct. 2052 (“[S]trategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.”); Rolan v. Vaughn, 445 F.3d 671, 682 (3d Cir.2006) (“Only choices made after a reasonable investigation of the factual scenario are entitled to a presumption of validity.”). Savino's approach to this capital case was in many ways deeply troubling. It is not that he did everything wrong. He obtained useful information on cross-examination and he quite rightly, and vigorously, objected to the trial court's determination to permit the last-minute ballistics evidence and other expert testimony. But it seems he conducted no investigation whatsoever. He interviewed no witnesses, including the one eye-witness, and instead cross-examined them cold at trial; he failed to interview the Commonwealth's firearms expert or obtain his own; and he did not engage any other forensic expert. Cf. Siehl v. Grace, 561 F.3d 189, 196 (3d Cir.2009) (holding that a state-court ruling was not an objectively reasonable application of Strickland when trial counsel sought no forensic experts or evidence). His representation of Collins was so deficient at the sentencing phase that it was declared ineffective by the Pennsylvania courts. Collins's assertion is not, as the Magistrate Judge concluded, that “Savino automatically loses the benefit of a presumption of reasonable strategic decisionmaking simply because he did not consult with a ballistics expert” (J.A. at A0021) but, rather, that Savino loses the presumption because he conducted no investigation and consulted with no one. During the PCRA hearing, Savino responded to the question “that's the strategic decision that you made, to rely on yourself?” with the answer, “[y]es, sir.” ( Id. at A1281.) It is hard to imagine why effective counsel in a capital case involving ballistics, chemical, and medical expert testimony would decline to seek the assistance of competent experts to advise the defense. So, while we do not rule on whether Savino's performance fell below the constitutional minimum of effectiveness, it is not surprising that Judge Brody declared the performance to be “clearly inadequate.” ( Id. at A0003.)

Our focus, though, is on the prejudice prong of Strickland. Again, to show prejudice under Strickland, a petitioner must demonstrate that there is a “reasonable probability that, but for counsel's unprofessional errors, the result of the proceedings would have been different.” 466 U.S. at 694, 104 S.Ct. 2052. A reasonable probability is one “sufficient to undermine confidence in the outcome.” Id. Prejudice is reviewed in light of the totality of the evidence at trial and the testimony at the collateral review hearing. Rolan, 445 F.3d at 682. The PCRA court held that “no prejudice has been shown by [Savino's] failure to [retain a ballistics expert]” (J.A. at A0118); that the lack of preparation and inadequate pretrial consultationwith Collins would not “change the verdict in this case” ( id. at A0107); that Collins “suffered no prejudice” ( id. at A0108) from Savino's failure to interview additional witnesses; and that Collins did not “demonstrate how he was prejudiced” by Savino's lack of investigation of witnesses ( id. at A0109). Although these are admittedly cursory statements, AEDPA requires that we “determine what arguments or theories supported or ... could have supported, the state court's decision.” Richter, 131 S.Ct. at 786.

The primary evidence before the jury consisted of Cofer's eye-witness account and the damning physical evidence, especially Graves's wounds, the bullet trajectories, and lead residue. Therefore, Collins argues, he was prejudiced because Savino's failure to investigate the forensic evidence affected his entire defense theory, and the jury was deprived of competing evidence on the main factual issue: where the shots originated. Collins claims that “[a] jury that was aware both of Cofer's credibility problems and that the physical evidence favored Cofer as the shooter would likely have acquitted Collins.” (Appellant's Opening Br. at 39.)

The problem for Collins is that Cofer's credibility was aggressively attacked and, more importantly, the physical evidence does not favor the defense. Especially when viewed through AEDPA's deferential lens, the PCRA court's determination on prejudice—a determination that has preclusive effect, see supra Part III—was not an unreasonable application of federal law under § 2254(d)(1).

Reviewing the arguments that “could have supported” that court's holding, Richter, 131 S.Ct. at 786, we conclude that the decision on prejudice reflects a reasonable application of Strickland, based on the totality of the available evidence—both presented at trial and at the PCRA hearing. See Williams, 529 U.S. at 398, 120 S.Ct. 1495 (describing state prejudice determination as unreasonable for failing to evaluate the totality of the evidence). Given the medical examiner's testimony, the damaging letters from Collins to Cofer, Welch's admission at the PCRA hearing that the shots could have come from a passenger in the backseat, the blood on Collins's pants, and his ownership of a gun matching the caliber of the murder weapon, the state court was not objectively unreasonable in determining that there was no reasonable probability that the result of the trial would have been different if Savino had prepared differently. “Although ‘fairminded jurists could disagree’ with the way the state court weighed the evidence in this case,” Rountree, 640 F.3d. at 544 (quoting Yarborough v. Alvarado, 541 U.S. 652, 664, 124 S.Ct. 2140, 158 L.Ed.2d 938 (2004)), it was not an unreasonable application of federal law for the state court to say that Savino's failures—whether described as a lack of preparation or confined to the ballistics evidence—did not raise a reasonable probability of a changed outcome, for reasons we more fully consider next.

First, the Commonwealth correctly asserts that the “uncontradicted and unequivocal testimony of the medical examiner,” Dr. McDonald, which would not have been refuted by an expert ballistician's testimony, was of great importance in showing that the shooter was behind the victim. (Appellee's Br. at 33.) At trial, McDonald testified that one of the gunshot wounds had a pattern that was “forward, to the right, and slightly downward.” (J.A. at A0678.) This was no doubt highly convincing evidence to the jury that a shot was fired from behind, especially given the exit wound under the eye. McDonald also testified that the other bullet went “rightward, slightly downward and forward” ( id. at A0676) and that there was a “graze wound of the left upper portion of the head going forward and to the right” ( id. at A0674). Those trajectories indicate a shot traveling forward rather than a shot directly to the right, and no alternative forensic explanation of that evidence was introduced. It is reasonable under Strickland to conclude that no amount of additional investigation on Savino's part would have changed the physical evidence relied on by the medical examiner, which alone went far in establishing that the shots came from behind Graves.

Second, although Welch testified at the PCRA hearing that he would have refuted Finor's ballistics testing and proposed that the shots could not have come from the backseat, that testimony did not withstand cross-examination. As the Magistrate Judge noted, Welch “was forced to concede ... that the physical evidence was consistent with the shooter” inside the car—either in the driver's seat or the backseat. ( Id. at A0021 n. 7.) The claim that the shooter was in the driver's seat rather than the backseat was substantially undermined by the bullets that went straight into the dashboard, as Welch effectively acknowledged. Welch also had to concede that “his principle reason for opining that the shooter did not sit in the rear of the car had nothing to do with specialized ballistics ... but rather his belief that it would be awkward” to shoot around the headrest. ( Id.) The Magistrate Judge was therefore on firm ground in saying that Welch's proposed expert testimony would not “have created a reasonable probability of an acquittal” (J.A. at A0021) because, at most, it only showed another possible shooter, and Savino had already pointed to both Cofer and the Boys from the Bottom to fill that role. Welch's testimony on the whole adds only marginally to the defense Savino mounted, as it raises some suspicion about Cofer but gives no supporting physical evidence. This is not a case where, as Collins argues, “rebuttal testimony from a credible, objective expert witness ... would have cast serious doubt on the prosecution's case.” (Appellant's Opening Br. at 40 (quoting Showers, 635 F.3d at 634).) On the contrary, Welch's cross-examination testimony at the PCRA hearing shows that he would have largely confirmed the Commonwealth's case—namely that the shots were fired from inside the car and possibly from the rear seat. Had Savino more explicitly argued that Cofer shot Graves from inside the car, as Collins claims Savino should have argued, it would still be sheer speculation to conclude that Welch's testimony would have changed the jury's decision to believe that Cofer, not Collins, was the shooter.

The Commonwealth contends that Collins told the police that he was sitting in the middle of the backseat and not directly behind the passenger. He thus would not have had to reach around the seat. However, at trial, Collins testified that there was a tire behind the driver's seat and a radio in the middle of the backseat.

Third, the Commonwealth introduced letters from Collins to Cofer that hurt Collins's credibility and supported Cofer's initial statements to police and his pretrial hearing testimony that Collins shot Graves. For example, one letter ended with the plea, “[l]ook, man, I know you ain't that type of dude that want to be label [sic] a snitch and I know for real you ain't built like that, so straighten this out as soon as possible.” ( Id. at A0542.) It is true that the jury could have decided that those words from Collins were a sincere effort to persuade Cofer to recant lies that Cofer had earlier told when implicating him. But it is also true that the jurors could have viewed the letters as witness tampering by a guilty man, and it appears that that is indeed how they viewed them. Moreover, taken together with the letters, Cofer's dodgy testimony at trial made his initial statements to the police all the more credible. What is important here is that no pretrial investigation or ballistics expert was going to make the letters disappear or explain them away. That conclusion is buttressed by the fact that the letters were given to the Commonwealth in the middle of trial, making any rebuttal preparation improbable. Although Savino's choice not to interview Cofer before cross-examination was highly questionable, we cannot say that it was objectively unreasonable for the state court to conclude that additional preparation for cross-examination would not have created a reasonable probability of an acquittal.

Finally, and though there is a disconcerting irony in it, the fact that the ballistics evidence and expert testimony were introduced at such a late stage in the trial undercuts Collins's prejudice claim. Because the headrest testing was a surprise, and despite Savino's failure to promptly prepare when the prosecution first told him of it, the effect of anything he could have done to counter the evidence would have been limited by the short time that he had. It was not Savino's fault that the Commonwealth introduced the testing so late. There was no forewarning and little, if anything, he could have done before the trial to prepare for it. Thus, the timing of the introduction of the ballistics evidence also weighs against a finding of prejudice.

This is not a case where the Commonwealth is alleged to have withheld evidence or delayed in bad faith. The Pennsylvania Supreme Court noted that Collins “does not dispute the PCRA court's finding that the Commonwealth disclosed the lead residue report on the very same day it was obtained, nor does he allege that the prosecutor was aware of the test results before that time.” Collins, 957 A.2d at 254.

Our decision about the lack of prejudice is not made lightly. Counterfactuals necessarily involve some speculation, and we cannot say with certainty that the result of Collins's trial would have been the same even if Savino had been better prepared for trial and hired appropriate experts. But that is not the standard we must apply. We look only far enough to determine if the state court reasonably applied federal law. See28 U.S.C. § 2254(d)(1). We do not ask whether we “believe[ ] the state court's determination under the Strickland standard was incorrect but whether that determination was unreasonable—a substantially higher threshold.” Knowles, 556 U.S. at 123, 129 S.Ct. 1411 (citations omitted) (internal quotation marks omitted). On the whole, it was not an unreasonable application of federal law for the state court to say that Savino's failures—whether they be broadly described as a lack of preparation or confined to the ballistics evidence—did not raise a reasonable probability of a different outcome.

Therefore, the PCRA court, the highest state court to expressly address the question of prejudice, did not unreasonably apply Strickland in holding that Collins could not establish prejudice, and we are bound to uphold that conclusion.

V. Conclusion

Although the performance of Collins's trial counsel may well have been constitutionally deficient, fairminded jurists could agree with the PCRA court's decision that Collins failed to show prejudice and so cannot establish ineffective assistance of counsel under Strickland. In addition, his cumulative error claim is procedurally defaulted. We will therefore affirm the decision of the District Court to deny Collins's petition for a writ of habeas corpus.