Romano v. Gibson

239 F.3d 1156, 2001 Colo. J. C.A.R. 771, 2001 U.S. App. LEXIS 2127, 2001 WL 120502

• Court: Court of Appeals for the Tenth Circuit
• Decided: February 13, 2001
• Docket: 99-6310, 99-6323
• Status: Published

Opinion

EBEL, Circuit Judge.

These death penalty appeals 1 present a number of issues. The most difficult, which we deal with at some length, include: 1) whether the State violated Brady v. Maryland, 373 U.S. 83, 83 S.Ct. 1194, 10 L.Ed.2d 215 (1963), by failing to disclose a detective’s unrecorded recollection of the temperature of the victim’s apartment; 2) what degree of mental torture or conscious serious physical abuse preceding death is necessary to satisfy Oklahoma’s especially heinous, atrocious or cruel aggravating factor; and 3) the extent to which defense counsel, under Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984), is required to investigate and present psychiatric evidence and evidence of a defendant’s early childhood during a capital sentencing proceeding. During the course of our analysis, we also clarify, with the approval of the en banc court, that the State does not need to appeal separately the district court’s adverse procedural bar determination in order to reassert that defense on appeal.

I. FACTS

Romano and Woodruff were convicted of killing a jeweler, Roger Sarfaty, during the course of a robbery. Sarfaty, Romano’s acquaintance, had been bound hand and foot and strangled, beaten about the head and stabbed five times. Because, at the time of this murder, Romano was serving a prison sentence and was only free on weekends, a critical issue in the case was how long Sarfaty had been dead before a friend first discovered his body, around 11:00 P.M. Tuesday, October 15, 1985. Romano had been out on weekend furlough from Friday evening, October 11 through Sunday evening, October 13. The State’s theory was that the murder occurred on Saturday, October 12, between 2:00 A.M. and 2:00 P.M. Romano and Woodruff do not have alibis for at least part of that time period.

Evidence supporting the State’s theory included the fact that Sarfaty usually went to a bar, the Celebrity Club, every night. The last time he was seen there was when he left the club about 2:00 A.M. Saturday, October 12. And at the time the body was discovered, the October 12 edition of the newspaper was found opened on Sarfaty’s couch; the editions for October 13, 14, and 15 were left delivered, but unopened, on his patio.

There was also, however, evidence contradicting the State’s theory. When Sarfaty’s friend first discovered his body, the volume of the television in Sarfaty’s apartment was turned way up, yet none of the neighbors had ever complained about the noise. In addition, Sarfaty’s neighbor testified that he thought he saw Sarfaty arguing with a blonde woman early Sunday morning, October 13.

Scientific evidence concerning the time of death was not conclusive, although it generally supported the State’s theory. The medical examiner’s investigator believed Sarfaty had died two to three days before the body’s discovery. The investigator based that determination, in part, on the apartment’s sixty-degree temperature, which he noted on the night the body was found. The investigator recorded that temperature at approximately 4:30 A.M. Wednesday, October 16, although authorities had first entered the apartment two hours earlier. A police officer described the apartment, at the time they entered, as warm, but not hot.

Based in part on the sixty-degree temperature, the medical examiner, Dr. Choi, testified that Sarfaty had been dead between two days and one week. The doctor’s best estimate was three or four days. Dr. Choi further testified, however, that, at a warmer temperature, Sarfaty might have been dead only twelve to twenty-four hours.

Apart from the opportunity to commit these crimes, the evidence linking Romano and Woodruff to the crime itself was primarily circumstantial. Sarfaty, a jeweler, often kept a great deal of jewelry with him or in his apartment. He also usually wore rings on each finger. On Thursday, October 10, a friend had seen Sarfaty with a bag of ten to twelve gold necklaces, some older looking. And late Friday, October 11, Sarfaty had shown the Celebrity Club’s manager a “real big” diamond. Yet Sarfaty was not wearing any rings when his body was found and there were only a few items of costume jewelry in the apartment at that time. Detectives did find a diamond in the living room, near where the body was discovered, and a packet of seventeen diamonds in one of Sarfaty’s suits.

On Sunday, October 13, Woodruffs girlfriend observed that Woodruff had a lot of jewelry, including some older looking gold necklaces and five or six rings. Although Woodruff was a trained gemologist, she had never before seen him with that much jewelry, nor did she think he had sufficient money at that time to buy such jewelry. Woodruff mailed this jewelry to an acquaintance in California.

Sarfaty also kept as many as six large containers of quarters in his apartment. There were, however, no such containers in the apartment following the murder. Further, on Saturday afternoon, October 12, Romano and Woodruff, who were then intoxicated, attempted to purchase a television at a mall store, using only quarters. Witnesses estimated the two had between ten and forty dollars’ worth of quarters. A saleswoman testified that, at that time, Woodruff also had what appeared to be spots of blood on his jeans. She also noted a recent cut on Woodruffs hand and that Romano was limping. Others who saw the pair later that day, however, did not notice any injuries. When mall security took them into custody for being drunk and disorderly, Romano had a “lock blade” folding knife. He was also wearing an expensive-looking gold necklace.

Later that evening, Woodruffs girlfriend delivered the two men from police custody to Woodruffs car, still parked at the mall. When she dropped them off, she noticed diamond papers — special papers designed to hold gems securely — on the ground near Woodruffs car. Sarfaty usually carried thirty to forty such papers with him. Woodruff, however, as a gemologist, also used diamond papers in his work.

When police arrested Woodruff, ten months after Sarfaty’s murder, he called his girlfriend and asked her to “clear” the house. In response, she removed a pair of gloves, a watch, and several pieces of rope. The medical examiner testified that these rope pieces could have caused the marks around Sarfaty’s neck, hands and feet. Additionally, one of the rope pieces was fashioned like a garrote, which could be used to strangle a victim. The garrote would have left marks like those found on Sarfaty’s neck — marks on the front and sides of the neck, but with an area at the back of the neck without any ligature marks at all.

In addition to this evidence linking Woodruff and Romano to these crimes, there was also evidence that at least Romano had a motive to rob and kill Sarfaty. A few weeks before the murder, Romano was in need of money and had asked a friend, Tracy Greggs, to help him rob Sar-faty. Romano was particularly interested in stealing the rings Sarfaty usually wore. Romano had told Greggs that, because Sarfaty knew and would recognize him, Romano would have to kill Sarfaty. Greggs had refused to help Romano.

Romano had also previously stolen a former girlfriend’s rings and had sold them to Sarfaty. When she discovered the jewelry was missing, the girlfriend threatened to report the theft to her insurance company and police. Although he had promised to get the rings back, Romano later told his former girlfriend that the man who had her rings had been killed.

The State jointly tried Woodruff and Romano. The jury convicted both of first degree malice murder and robbery with a dangerous weapon, sentencing each to 1000 years in prison on the robbery conviction, which was after former conviction of a felony for both. At the capital sentencing proceeding, the jury found three aggravating factors pertaining to both Woodruff and Romano: they had prior violent felony convictions; the murder was especially heinous, atrocious or cruel; and they were continuing threats to society. The jury also found that Romano had committed the murder to avoid arrest or prosecution for Sarfaty’s robbery. The jury sentenced both Woodruff and Romano to death on the first degree murder convictions.

In a related matter, prior to the Sarfaty trial, the State had jointly tried Woodruff and Romano for murdering another Romano acquaintance, Lloyd Thompson. In that case, a jury also convicted both men of first degree murder and sentenced them to death. The State introduced evidence of those Thompson convictions and death sentences during sentencing in the Sarfaty trial. After the conclusion of the Sarfaty case, however, the Oklahoma Court of Criminal Appeals overturned the Thompson convictions and death sentences, holding the trial court had erred in jointly trying Woodruff and Romano for that murder. See Romano v. State, 827 P.2d 1335 (Okla.Crim.App.1992); Woodruff v. State, 825 P.2d 273 (Okla.Crim.App.1992). The State subsequently retried Woodruff and Romano for Thompson’s murder. Separate juries again convicted both of first degree murder. Romano’s jury sentenced him to death, but Woodruff received a sentence of life imprisonment without parole. See Romano v. State, 909 P.2d 92, 106 & n. 1 (Okla.Crim.App.1995).

The Oklahoma Court of Criminal Appeals affirmed the Sarfaty convictions and sentences on direct appeal. See Woodruff v. State, 846 P.2d 1124 (Okla.Crim.App.), cert. denied, 510 U.S. 934, 114 S.Ct. 349, 126 L.Ed.2d 313 (1993); Romano v. State, 847 P.2d 368 (Okla.Crim.App.1993), aff'd, 512 U.S. 1, 114 S.Ct. 2004, 129 L.Ed.2d 1 (1994). In affirming Romano’s death sentence, however, the state appellate court struck the prior violent felony aggravator, which was based solely on Romano’s conviction for Thompson’s murder, and then reweighed the remaining aggravating circumstances and mitigating evidence. See Romano, 847 P.2d at 389, 393-94. The United States Supreme Court, in a 5-4 decision, affirmed, specifically holding that introduction of evidence of Romano’s earlier death sentence for the Thompson murder did not diminish the Sarfaty jury’s sense of responsibility for sentencing Romano for that murder. See Romano, 512 U.S. at 3, 6, 9-10, 114 S.Ct. 2004. Subsequently, the Oklahoma Court of Criminal Appeals affirmed the denial of state post-conviction relief from the Sarfaty convictions and death sentences. See Romano v. State, 917 P.2d 12 (Okla.Crim.App.1996); Woodruff v. State, 910 P.2d 348 (Okla.Crim.App.1996).

II. STANDARDS OF REVIEW

Because Woodruff and Romano filed their individual habeas petitions after the effective date of the Antiterrorism and Effective Death Penalty Act of 1996 (AED-PA), that Act governs these appeals. See Williams v. Taylor, 529 U.S. 362, 402, 120 S.Ct. 1495, 146 L.Ed.2d 389 (2000). Under AEDPA, petitioners will not be entitled to habeas relief unless they can establish that the state court determination of their claims was contrary to, or an unreasonable application of, clearly established Supreme Court precedent, see 28 U.S.C. § 2254(d)(1), or was an unreasonable determination of the facts in light of the evidence, see id. § 2254(d)(2). We presume the correctness of any state court findings of fact, absent clear and convincing proof to the contrary. See id. § 2254(e)(1).

Where the state court did not address the merits of a habeas claim, this court reviews the district court’s resolution of that ground for relief de novo, reviewing for clear error any district court findings of fact. See, e.g., Thomas v. Gibson, 218 F.3d 1213, 1220 (10th Cir.2000) (citing LaFevers v. Gibson, 182 F.3d 705, 711 (10th Cir.1999)).

III. DISCUSSION

A. Was there sufficient evidence to support petitioners’ convictions for first degree malice murder and robbery with a dangerous weapon?

1. Applicable standard of review

In federal habeas proceedings, the appropriate inquiry into a sufficiency-of-the-evidenee claim is “whether, after viewing the evidence in the light most favorable to the prosecution, any rational trier of fact could have found the essential elements of the crime beyond a reasonable doubt.” Jackson v. Virginia, 443 U.S. 307, 319, 99 S.Ct. 2781, 61 L.Ed.2d 560 (1979). AEDPA additionally directs that, where the state court has already addressed the claim, this court’s review is further limited. See Valdez v. Ward, 219 F.3d 1222, 1237 (10th Cir.2000). Romano asserts that, even though the Oklahoma Court of Criminal Appeals did address the sufficiency of the evidence supporting petitioners’ convictions, it failed to apply Jackson. Therefore, this court’s review, according to Romano, should be de novo.

The state appellate court, because of the circumstantial nature of the evidence against Woodruff and Romano, did apply a state-law standard of review, considering whether the evidence was “inconsistent with any reasonable hypothesis other than the defendant’s guilt.” Romano, 847 P.2d at 378; Woodruff, 846 P.2d at 1133. That standard is actually more onerous than Jackson. Thus, if the evidence was sufficient to meet Oklahoma’s stricter test, it would certainly also meet the Jackson standard.

Applying AEDPA, 2 therefore, we review the reasonableness of the state appellate court’s determination that the evidence was sufficient to support Woodruffs and Romano’s convictions. In doing so, we consider only the evidence presented at trial and do not, at this juncture, address any newly discovered evidence. See Herrera v. Collins, 506 U.S. 390, 402, 113 S.Ct. 853, 122 L.Ed.2d 203 (1993).

2. Sufficiency of the evidence

Under Oklahoma law, see Jackson, 443 U.S. at 324 n. 16, 99 S.Ct. 2781, “ ‘[a] person commits murder in the first degree when that person unlawfully and with malice aforethought causes the death of another human being. Malice is that deliberate intention unlawfully to take away the life of a human being, which is manifested by external circumstances capable of proof.’ ” Bland v. State, 4 P.3d 702, 713 (Okla.Crim.App.2000) (emphasis omitted) (quoting Okla. Stat. tit. 21, § 701.7(A)), cert. denied, — U.S. -, 121 S.Ct. 832, 148 L.Ed.2d 714 (2001).

A design to effect death lie., premeditation] is inferred from the fact of killing, unless the circumstances raise a reasonable doubt whether such design existed. Premeditation sufficient to constitute murder may be formed in an instant[,] or it may be formed instantaneously as the killing is being committed. Malice aforethought may be proved by circumstantial evidence.

Id. (citations, quotation omitted).

In addition, “[t]o sustain a conviction for robbery with a dangerous weapon, the State must prove another’s wrongful taking of one’s personal property against one’s will, by force or fear.” Mitchell v. State, 884 P.2d 1186, 1199-1200 & 1200 n. 89 (Okla.Crim.App.1994) (citing, e.g., Okla. Stat. tit. 21, § 791).

Woodruff and Romano argue that they did not rob or kill Roger Sarfaty. They challenge, in particular, the largely circumstantial case underlying their convictions. The evidence, however, established the following: Romano needed money and also sought the return of his former girlfriend’s rings from Sarfaty. He had asked Tracy Greggs to help him rob Sarfaty, and had told Greggs that in doing so, Romano would have to kill Sarfaty.

Although the time of the murder was critical and disputed, in considering this claim, we must view the evidence in the light most favorable to the State. See Jackson, 443 U.S. at 319, 99 S.Ct. 2781. In that light, the evidence did support the State’s theory that the murder occurred between 2:00 A.M. and 2:00 P.M. Saturday, October 12. The medical examiner testified that Sarfaty could have died between two and seven days before his body was discovered. Additionally, employees of the Celebrity Club last saw Sarfaty alive during the early morning of October 12. And police found the October 12 edition of the newspaper open on Sarfaty’s couch, while the later editions were left delivered, but untouched, on his patio.

Woodruff and Romano lack alibis for at least part of this time frame. Romano was on furlough from State custody that weekend. And, although he may have exchanged a number of telephone calls that Saturday morning with a friend, the last one was before 10:00 A.M. In addition, although Woodruff had told his girlfriend that he would be visiting his parents that Saturday, when she called his parents’ home, he was not there.

Later that day, both Woodruff and Romano possessed a large number of quarters. Sarfaty was known to keep large amounts of quarters in his apartment, yet there were no containers of quarters found there after his death.

Further, although Sarfaty had several gold necklaces just before the murder and he usually wore a number of rings, none of this jewelry was found at his home after the murder. When police took custody of Romano Saturday afternoon, October 12, he was wearing an expensive looking gold necklace. And the next day, Woodruffs girlfriend saw him with a number of gold necklaces and rings.

Sarfaty had been stabbed several times, beaten and strangled. Romano had a knife with him when police took the pair into custody the afternoon of October 12. A salesperson had noticed blood on Wood-ruffs jeans that afternoon and a recent cut on his hand, and Romano was limping. And when Woodruff was arrested months later, he asked his girlfriend to remove from their home, among other things, several pieces of rope, including a garrote, that could have produced the ligature marks found on Sarfaty’s body.

Considered in the light most favorable to the government, see Jackson, 443 U.S. at 319, 99 S.Ct. 2781, this evidence was more than sufficient for a rational trier of fact to find the existence of all of the elements of these charged offenses beyond a reasonable doubt. Therefore, the state appellate court reasonably upheld Romano’s and Woodruffs convictions for first degree malice murder and robbery with a dangerous weapon. See Romano, 847 P.2d at 378-80; Woodruff, 846 P.2d at 1133-35.

B. Did the trial court violate due process or the Sixth or Eighth Amendments when it restricted petitioners’ ability to put on evidence arguably casting suspicion on other, uncharged individuals?

Oklahoma has an evidentiary rule that a criminal defendant cannot put on evidence that someone else might have committed the charged offense, absent proof that person took an overt act toward the commission of the crime. See, e.g., Dennis v. State, 879 P.2d 1227, 1232 (Okla.Crim.App.1994). Proof of another’s motive is not enough. See, e.g., id.

In this case, Woodruff and Romano sought to cast suspicion for Sarfaty’s robbery and murder on T.R. “Tippy” Ballard, Kathy Ford, and Susan Babbitt. The state trial court admitted some evidence connecting these three individuals with Sarfaty, but excluded other such evidence. Applying Oklahoma’s evidentiary rule, the Oklahoma Court of Criminal Appeals upheld the trial court’s exclusion of this evidence. See Romano, 847 P.2d at 380-82; Woodruff, 846 P.2d at 1137-38.

Although, on direct appeal, Woodruff and Romano did challenge the trial court’s exclusion of this evidence on federal constitutional grounds, the Oklahoma Court of Criminal Appeals addressed these claims only under state law. See Romano, 847 P.2d at 380-82; Woodruff, 846 P.2d at 1137-38. We, therefore, review de novo the federal district court’s denial of habeas relief on these due process and Sixth and Eighth Amendment claims. See Thomas, 218 F.3d at 1220.

Of course, sitting as a federal habeas court applying 28 U.S.C. § 2254, it is not for us to review a state court’s evidentiary rulings. See, e.g., Estelle v. McGuire, 502 U.S. 62, 67-68, 112 S.Ct. 475, 116 L.Ed.2d 385 (1991). Rather, a federal habeas court reviews only for violation of “the Constitution, laws, or treaties of the United States.” Id. at 68, 112 S.Ct. 475 (citing, e.g., 28 U.S.C. § 2241). Further, state evidentiary determinations ordinarily do not present federal constitutional issues. See Crane v. Kentucky, 476 U.S. 683, 689, 106 S.Ct. 2142, 90 L.Ed.2d 636 (1986) (noting Court’s “traditional reluctance to impose constitutional constraints on ordinary evidentiary rulings by state trial courts”). However, the Supreme Court, in, e.g., Chambers v. Mississippi, 410 U.S. 284, 302, 93 S.Ct. 1038, 35 L.Ed.2d 297 (1973), and Green v. Georgia, 442 U.S. 95, 97, 99 S.Ct. 2150, 60 L.Ed.2d 738 (1979) (capital sentencing proceeding), has provided an exception, under some circumstances, if a state court applies the State’s evidentiary rules unfairly to prevent a defendant from presenting evidence that is critical to his defense. See also, e.g., Washington v. Texas, 388 U.S. 14, 16, 23, 87 S.Ct. 1920, 18 L.Ed.2d 1019 (1967). This, however, is not such a case. The state trial court did not exclude any evidence critical to Woodruffs or Romano’s defense. 3

Here, Woodruff and Romano were in fact able to put on a significant amount of evidence concerning these three individuals. See Boyd, 179 F.3d at 921; see also Richmond v. Embry, 122 F.3d 866, 873-74 (10th Cir.1997). They first presented evidence that it was generally well known that Sarfaty frequently carried lots of jewelry and money with him. Additionally, Sarfaty had been a crime victim on several previous occasions, having had things taken from his home and jewelry stolen from him. Sarfaty told a friend he suspected “[s]ome woman” had perpetrated these crimes. Within two weeks prior to the murder, Sarfaty had reported separate in cidents of assault and burglary and, immediately prior to the murder, Sarfaty told a friend that he anticipated another robbery attempt against him. His friend urged Sarfaty to change the way he did business, late at night in bars and clubs.

Woodruff and Romano also presented evidence at trial specifically pertaining to Ballard, Babbitt and Ford. T.R. Ballard was Sarfaty’s acquaintance. They occasionally went to the same pool hall, and had, at times, also discussed business together at the Celebrity Club. Prior to Sar-faty’s death, Ballard had indicated that he was broke. Yet, immediately after the murder, Ballard appeared to have a large amount of money and jewelry. No one had seen Ballard in the pool hall or Celebrity Club after Sarfaty was killed. However, a detective investigating Sarfaty’s death testified that, although he had come across Ballard’s name early in the investigation, the detective never considered him a suspect. Nor could the detective find a link between Ballard and either Kathy Ford or Susan Babbitt.

The detective did come across Kathy Ford’s name during his investigation, but he was never able to find her or even verify that she existed. He indicated she may have been a prostitute. The owner of a pawn shop and his employee testified that they knew Ford and had sold her a knife just before Sarfaty’s murder. The medical examiner could not rule out that knife as the murder weapon.

The investigating detective had also come across the name of Susan Babbitt, purportedly another prostitute, but the detective was unable to link her to either Ballard or Ford. The only connection he could make between Babbitt and Sarfaty was that they were acquaintances from the Celebrity Club.

Based upon this evidence, defense counsel could have argued to the jury that someone other than Woodruff and Romano killed Sarfaty. Nonetheless, Woodruff and Romano assert that the trial court erred in excluding additional evidence concerning Ballard, Ford and Babbitt. The record, however, does not support their contention that the trial court excluded evidence that Ballard had threatened Sarfaty. The trial court did sustain an objection to defense counsel’s cross-examination of Sarfaty’s friend as to whether Sarfaty ever had any problems with Ballard. In response, defense counsel asserted to the court that one of the witnesses had indicated that, on one occasion, Ballard had slapped Sarfaty, but defense counsel was not sure which witness it was. The trial court then instructed defense counsel to find out which witness it was before counsel posed the question, and indicated counsel could recall this witness if it was necessary to do so. Defense counsel, however, never recalled that witness, nor did she attempt to elicit this information from any other witness. The trial court, therefore, did not preclude the defense from introducing this evidence. Cf. United States v. Ramone, 218 F.3d 1229, 1237 (10th Cir.) (noting defense counsel’s failure to seize opportunity to cross-examine did not show that trial court unfairly limited right to confront witnesses), cer t. denied, — U.S. -, 121 S.Ct. 598, 148 L.Ed.2d 512 (2000).

Nor could we find any instance in the record that the trial court prevented Woodruff or Romano from asserting any evidence specifically concerning Susan Babbitt. The trial court, therefore, did not deprive petitioners of any opportunity to present evidence concerning these two uncharged individuals.

The trial court did exclude some defense evidence concerning Kathy Ford. At trial, defense counsel made an offer of proof that Sarfaty’s friends had mentioned that Kathy Ford had possibly set up and robbed Sarfaty on previous occasions. Sarfaty himself had told friends that she had robbed him and, on one occasion, Sar-faty had asked a pawn shop owner to help him find Ford so that he could get back the property she had stolen from him. Nonetheless, this evidence was not critical to Woodruffs and Romano’s defense. Rather, the state court’s application of Oklahoma’s evidentiary rule pertaining to uncharged individuals only prevented Woodruff and Romano from presenting incremental evidence concerning another individual who might have been involved in this crime. It did not deprive them of significant and fundamentally exculpatory evidence. See Boyd, 179 F.3d at 921 (holding evidence must be material; that is, that it might have affected trial’s outcome); see also Richmond, 122 F.3d at 872, 874-75. See generally United States v. Valenzuela-Bernal, 458 U.S. 858, 867, 102 S.Ct. 3440, 73 L.Ed.2d 1193 (1982).

Woodruff and Romano argue that it is somehow unfair to apply the state-law overt act requirement concerning uncharged individuals who may have committed the crime because no similar requirement applies in order for the State to convict the individual charged. This particular argument is like comparing apples and oranges. At trial, the jury unavoidably has to focus on whether the charged individual did or did not commit the crime alleged. Crimes can be proven by circumstantial evidence; there is no requirement that substantive crimes must be proven only through direct evidence. See generally Lucero v. Kerby, 133 F.3d 1299, 1312 (10th Cir.1998) (reviewing sufficiency of both direct and circumstantial evidence to support conviction). Moreover, the jury did have to find that Woodruff and Romano had committed overt acts toward commission of robbery and murder. The jury could do so, however, based on circumstantial evidence.

The ability to cast aspersions on uncharged individuals, however, is a very different story. See Irvin v. State, 11 Okla.Crim. 301, 146 P. 453, 464 (1915) (“While it is competent for the defendant to show, by any legal evidence, that some other person committed the crime with which he is charged, ... evidence which could have no further effect than to cast a bare suspicion upon another is incompetent and inadmissible.”). The main purpose of Oklahoma’s evidentiary rule requiring proof of an overt act by an uncharged individual is to prevent juries from embarking on wild goose chases. See, e.g., id. at 464-66 (citing authority). Here, there was no proof, either direct or circumstantial, that Ballard, Ford or Babbitt ever took any overt act toward commission of these offenses. The jury, then, had to remain focused on the trial’s central issue — whether the charged individuals committed the crimes alleged. See id. at 466 (citing Horn v. State, 12 Wyo. 80, 73 P. 705 (1903)).

In any event, here, in light of Green and Chambers, we need ask no more than whether the trial court’s application of this state evidentiary rule excluded critical exculpatory evidence. See Green, 442 U.S. at 97, 99 S.Ct. 2150; Chambers, 410 U.S. at 302, 93 S.Ct. 1038. It did not.

C. Did the State violate Brady by failing to disclose a detective’s unrecorded opinion about the temperature of Sarfaty’s apartment at the time the body was discovered?

1. Procedural default

As an initial matter, respondent argues Romano procedurally defaulted this claim because he failed to raise it on direct appeal. Respondent, however, did not appeal from the district court’s rejection of this procedural bar defense. Nonetheless, a federal habeas court can always raise procedural bar sua sponte. See Duvall v. Reynolds, 139 F.3d 768, 796 n. 11 (10th Cir.1998), and cases cited therein; see also, e.g., Hale, 227 F.3d at 1334 n. 16. Further, this court can affirm the district court’s decision on any legal ground the record supports. See, e.g., Hernandez v. Starbuck, 69 F.3d 1089, 1093 (10th Cir.1995). Therefore, this court can address procedural default even though respondent did not separately appeal the district court’s determination. Cf. Jones v. United States, 527 U.S. 373, 396, 119 S.Ct. 2090, 144 L.Ed.2d 370 (1999) (holding, in direct criminal appeal where circuit court held aggravating factors were invalid, but error was harmless, that government could argue on appeal validity of aggravators without filing cross appeal).

Admittedly, previous Tenth Circuit law has not been consistent on whether the State must specifically appeal an adverse procedural bar determination. In Robison v. Maynard, 829 F.2d 1501, 1502 (10th Cir.1987), this court did hold that the State’s failure to appeal from a district court’s adverse procedural default determination precluded the State from raising that issue to this court on appeal. In several subsequent cases, however, this court has held that the State’s separate appeal was unnecessary. See, e.g., Walker v. Gibson, 228 F.3d 1217, 1226 n. 2 (10th Cir.2000); Jones v. Gibson, 206 F.3d 946, 955 n. 4 (10th Cir.), cert. denied, — U.S. -, 121 S.Ct. 496, 148 L.Ed.2d 467 (2000); see also Smith v. Massey, 235 F.3d 1259, 1273 n. 9 (10th Cir.2000).

When confronted, as we are here, with inconsistent Tenth Circuit holdings, we are generally bound by the first decision. See Calderon v. Kan. Dep’t of Soc. & Rehab. Servs., 181 F.3d 1180, 1187 (10th Cir.1999); Haynes v. Williams, 88 F.3d 898, 900 n. 4 (10th Cir.1996). That would require the State here to appeal from the district court’s procedural bar determination in order to raise that issue now on appeal. We believe, however, that this court’s earlier authority addressing this issue was incorrect. We have submitted this determination to the entire en banc court, which agrees. See, e.g., United States v. Meyers, 200 F.3d 715, 721 & n. 3 (10th Cir.2000); Murphy v. Klein Tools, Inc., 935 F.2d 1127, 1128 n. 2 (10th Cir.1991). Therefore, following Walker and Jones, we overrule Robison in this regard. Other circuits have reached this same conclusion. See Hull v. Kyler, 190 F.3d 88, 98 n. 2 (3d Cir.1999); Moore v. Ponte, 186 F.3d 26, 31 & n. 4 (1st Cir.), cert. denied, 528 U.S. 1053, 120 S.Ct. 596, 145 L.Ed.2d 495 (1999); Nichols v. McCormick, 929 F.2d 507, 509 n. 2 (9th Cir.1991); Washington v. Lane, 840 F.2d 443, 445-46 (7th Cir.1988); see also Young v. Catoe, 205 F.3d 750, 762 n. 12 (4th Cir.) (noting, in dicta, government’s cross-appeal unnecessary), ce rt. denied, — U.S. -, 121 S.Ct. 164, 148 L.Ed.2d 111 (2000). See generally El Paso Natural Gas Co. v. Neztsosie, 526 U.S. 473, 479, 119 S.Ct. 1430, 143 L.Ed.2d 635 (1999) (noting even without cross-appeal appellee may assert any ground for affirmance that is apparent on record, even if that ground challenges lower court’s reasoning so long as it does not enlarge the relief previously given appellee below).

We turn, then, to respondent’s procedural bar argument. In these cases, the Oklahoma Court of Criminal Appeals ruled inconsistently on this procedural default issue when it considered the merits of this claim in Woodruffs post-conviction proceeding, but declined to do so in Romano’s post-conviction case. The Oklahoma Court of Criminal Appeals held Romano had waived this claim because, with due diligence, he could have discovered its factual basis and asserted it on direct appeal. See Romano, 917 P.2d at 15-16. Conversely, the same court concluded Wood-ruff could not have raised this claim on direct appeal. 4 See Woodruff, 910 P.2d at 350-51.

One of the requirements of an adequate and independent state procedural bar, see, e.g., Lambrix v. Singletary, 520 U.S. 518, 522-23, 117 S.Ct. 1517, 137 L.Ed.2d 771 (1997), is that the state courts must have applied it consistently, see, e.g., Johnson v. Mississippi, 486 U.S. 578, 587, 108 S.Ct. 1981, 100 L.Ed.2d 575 (1988), and cases cited therein. “[T]he procedural rules must be applied evenhandedly to all similar claims.” Maes v. Thomas, 46 F.3d 979, 986 (10th Cir.1995) (further quotation omitted). Here the Oklahoma Court of Criminal Appeals applied the same procedural rule inconsistently in the cases of two co-defendants charged with the very same offenses and tried together. In light of that inconsistency, this procedural bar will not preclude our consideration of the merits of these habeas claims. See, e.g., Gutierrez v. Moriarty, 922 F.2d 1464, 1469-71 (10th Cir.1991) (holding state procedural bar, which state courts had not strictly or regularly applied, did not provide adequate state law grounds precluding federal habeas review).

2. Merits

Because both Woodruff and Romano had access to Sarfaty and did not have an alibi for the morning and early afternoon of October 12, it was important to determine Sarfaty’s date of death. Sarfaty’s friend discovered the body late Tuesday night, October 15. At trial, the medical examiner, Dr. Choi, testified Sarfaty had been dead no less than two and no more than seven days. This would have included the time Woodruff and especially Romano had access to Sarfaty. Woodruff and Romano, on the other hand, argued that Sarfaty may have been dead less than two days, which would have precluded at least Romano’s opportunity to commit these crimes.

Further, it is undisputed that the apartment’s temperature is relevant to this time-of-death determination because it would affect how quickly the body decomposed. And at trial, the medical experts relied at least in part on the decomposition rate to estimate the time of Sarfaty’s death.

Of course, date of death was not strictly based on the decomposition rate. There was other evidence suggesting Sarfaty died on October 12: The last time Celebrity Club employees saw Sarfaty, a nightly customer, was early morning October 12. And police found the October 12 edition of the newspaper open on Sarfaty’s couch, while the October 13, 14, and 15 editions were found delivered, but unopened, on his patio. On the other hand, there was also evidence suggesting a later time of death — Sarfaty’s neighbor testified that he saw Sarfaty arguing with a woman early Sunday morning, October 13.

In any event, the evidence at trial concerning the temperature of Sarfaty’s apartment during the time in question included the following: Sarfaty’s friend discovered the body at approximately 11:00 P.M. Tuesday, October 15, and called police. Officer Tom Dale reported finding the body at approximately 1:00 A.M. Wednesday morning. Authorities, however, did not enter the apartment until 2:30 A.M. At that time, Officer Dale described the apartment as “warm but not hot.” Two hours later, at approximately 4:30 AM., the medical examiner’s investigator noted the apartment’s thermostat registered sixty degrees. The investigator believed Sarfaty had been dead two or three days. Relying on the investigator’s sixty-degree temperature, the medical examiner testified Sarfaty had been dead no less than two and no more than seven days, and most probably three to four days. Both the medical examiner and the investigator noted the body’s “moderate decomposition.”

Almost nine years after Sarfaty’s murder, during the sentencing stage of Wood-ruffs retrial for the Thompson murder, Detective Jerry Martin testified that when he entered Sarfaty’s apartment that morning, the apartment was “quite warm ... probably about 80 degrees.” Detective Martin, however, indicated that he was just estimating that temperature. He had not determined the temperature scientifically by, for example, checking the thermostat. Nor was it Martin’s job specifically to note the apartment’s temperature; rather, his responsibility at the crime scene was to investigate evidence outside the apartment, including interviewing neighbors. And Detective Martin never recorded his temperature observation in any written document or report.

In light of these facts, the State challenges whether Detective Martin’s recollection of the apartment’s temperature, unrecorded and not recalled until nine years after the investigation, was in fact extant Brady material at the time of the Sarfaty trial. For purposes of these appeals, we will accept, without adopting, petitioners’ characterization of the detective’s temperature recollection as Brady material. Under Brady, therefore, Wood-ruff and Romano must establish that Detective Martin’s recollection was “favorable to the accused, either because it is exculpatory, or because it is impeaching; [it] must have been suppressed by the State, either willfully or inadvertently; and prejudice must have ensued.” Strickler v. Greene, 527 U.S. 263, 281-82, 119 S.Ct. 1936, 144 L.Ed.2d 286 (1999).

Although the state appellate court, in Woodruffs state post-conviction proceeding, did address the newly discovered evidence of the apartment’s temperature, that court did not specifically address Wood-ruffs Brady claim. See Woodruff, 910 P.2d at 350-51. Nor did that court address the merits of Romano’s Brady claim, because it deemed Romano to have waived that issue. See Romano, 917 P.2d at 15 n. 2, 16. We, therefore, review these Brady claims de novo. See Thomas, 218 F.3d at 1220. We review for clear error the federal district court’s factual findings, made in this case after an evidentiary hearing. See id.

Detective Martin’s temperature estimate was not necessarily exculpatory. It was not inconsistent with, but rather was merely cumulative of, Officer Dale’s testimony at the Sarfaty trial that the apartment was “warm but not hot.” See Foster v. Ward, 182 F.3d 1177, 1192 (10th Cir.1999), cert. denied, 529 U.S. 1027, 120 S.Ct. 1438, 146 L.Ed.2d 326 (2000) (holding cumulative evidence, which added only marginally to defense, was not exculpatory).

Using Detective Martin’s eighty-degree temperature estimate as their foundation, however, Woodruff and Romano presented, during their federal habeas proceedings, the testimony of three experts whose opinions support petitioners’ contention that Sarfaty had been dead only twelve to thirty hours before the body’s discovery. A thermodynamics expert, Dr. Sutton, conducted a study indicating the temperature in Sarfaty’s apartment during the days preceding the body’s discovery was approximately ninety degrees. An anthropologist, Dr. Marks, testified that Sarfaty had probably been dead only thirty to forty-eight hours. And a Texas medical examiner thought Sarfaty had died no more than one and one-half days before his body was discovered, and “quite probably sooner than that.” Woodruff and Romano claim that, had they known about Detective Martin’s eighty-degree temperature estimate earlier, they could have garnered and presented this expert testimony at trial in support of their defense.

Petitioners’ experts, however, predicated their time-of-death opinions on a more rapid decomposition rate, which in turn was based upon Dr. Sutton’s thermodynamics study indicating the temperature in the apartment was actually around ninety degrees or more. If anything, then, Detective Martin’s eighty-degree estimate undercuts, rather than supports, Dr. Sutton’s temperature estimate which, in turn, undercuts the premise of the decomposition experts’ testimony.

Nonetheless, we will assume, for purposes of our analysis here, that Detective Martin’s temperature estimate was favorable to Woodruff and Romano. We also assume that the State suppressed this evidence. 5 See, e.g., Strickler, 527 U.S. at 280-81, 119 S.Ct. 1936 {Brady includes information known only to police investigators and not to prosecutors). We, therefore, need address here only whether the State’s suppression of Detective Martin’s eighty-degree temperature estimate prejudiced Woodruff and Romano.

Brady evidence will be material, and thus prejudicial, “if there is a reasonable probability that, had the evidence been disclosed to the defense, the result of the proceeding would have been different,” Strickler, 527 U.S. at 280, 119 S.Ct. 1936 (further quotation omitted). The appropriate inquiry, then, is whether, absent knowledge of Detective Martin’s eighty-degree temperature estimate, Woodruff and Romano received “a fair trial, understood as a trial resulting in a verdict worthy of confidence.” Id. at 289-90, 119 S.Ct. 1936 (further quotation omitted).

Brady’s, prejudice inquiry is equivalent to the prejudice analysis that applies to an ineffective assistance of counsel claim under Strickland, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674. See United States v. Bagley, 473 U.S. 667, 681-83, 105 S.Ct. 3375, 87 L.Ed.2d 481 (1985) (separate opinion of Justice Blackmun); see also Kyles v. Whitley, 514 U.S. 419, 436, 115 S.Ct. 1555, 131 L.Ed.2d 490 (1995). The state appellate court did not address the merits of petitioners’ Brady claims, deeming Romano’s claim procedurally barred and addressing Woodruffs claim only in terms of newly discovered evidence. We will, therefore, review Brady’s, materiality determination de novo, see Moore v. Gibson, 195 F.3d 1152, 1165 (10th Cir.1999) (noting pre-AEDPA review of Brady materiality is de novo), cert. denied, 530 U.S. 1208, 120 S.Ct. 2206, 147 L.Ed.2d 239 (2000), just as we would Strickland’s prejudice inquiry, see Smith v. Gibson, 197 F.3d 454, 461 (10th Cir.1999), cert. denied, — U.S. -, 121 S.Ct. 102, 148 L.Ed.2d 60 (2000).

We also review for prejudice against the backdrop of the other objective evidence pertaining to Sarfaty’s date of death. See Banks v. Reynolds, 54 F.3d 1508, 1518 (10th Cir.1995) (determining materiality in light of entire record). The rate of decomposition, therefore, was not the sole, or even the most compelling, evidence of the time of death. See Foster, 182 F.3d at 1192. Accordingly, the importance of the decomposition rate diminishes. In addition, Detective Martin’s recollection that the apartment’s temperature was approximately eighty degrees was again not inconsistent with Officer Dale’s testimony at the Sarfaty trial that the apartment was warm, but not hot. Thus, the allegedly new material pertaining to the apartment’s temperature is largely cumulative of other evidence presented at trial.

Additionally, Detective Martin’s temperature estimate is itself not entitled to much weight because it was based on the detective’s unrecorded and subjective recall nine years after the investigation. The detective acknowledged his duties at the crime scene did not include recording the apartment’s temperature. Nor did Detective Martin objectively make that determination. He did not check his observation with the prevailing ambient temperatures during the period. Rather, his testimony concerning the apartment’s temperature represented merely an after-the-fact effort to recall an ancient subjective impression of a detail upon which he did not focus even at the time. It is, therefore, of little relevance.

Lastly, petitioners’ decomposition experts’ based their time-of-death opinions on Dr. Sutton’s conclusion that the apartment’s temperature remained at approximately ninety degrees. The district court, however, afforded Dr. Sutton’s testimony little weight, determining cross-examination had exposed several weaknesses in his study and faulting his demeanor and manner of responding to questions posed at the evidentiary hearing. We must defer to the district court’s credibility determinations. See, e.g., Anderson v. City of Bessemer City, 470 U.S. 564, 575, 105 S.Ct. 1504, 84 L.Ed.2d 518 (1985).

For all of these reasons, we conclude the State’s failure to disclose Detective Martin’s temperature observation, even assuming it was extant Brady material, did not prejudice Woodruffs and Romano’s defense.

D. Did the State violate Brady and Giglio 6 by failing to disclose its deal exchanging the favorable disposition of pending forged check charges against Greg Myers for Myers’ testimony in this case?

Greg Myers, Romano’s former cellmate, testified during sentencing that Romano admitted to Myers that Romano and his partner had killed Lloyd Thompson, and Romano had then solicited Myers, every day for a week, to kill several witnesses in the Thompson case. After the Sarfaty trial, however, Myers recanted this testimony, asserting instead that Romano had never admitted any involvement in Thompson’s murder; neither had Romano solicited Myers to kill any witnesses. According to Myers, he had testified falsely against Romano in exchange for the State’s favorable disposition of then pending felony bogus check charges against Myers. In this habeas proceeding, Romano challenges both the State’s failure to disclose its deal with Myers and Myers’ false testimony providing part of the basis for his death sentence. 7

1. Did the State fail to disclose a deal it had with Greg Myers?

Romano shared a jail cell with Greg Myers during October and November 1985. Myers was at that time facing felony bogus check charges filed by the same district attorney’s office prosecuting Woodruff and Romano. Because he had a prior felony conviction for committing lewd acts with a child, Myers was concerned that he might be charged as a former felon. If so, Myers faced a minimum ten-year prison sentence. See Okla. Stat. tit. 21, § 51(A)(1) (repealed effective July 1, 1999).

When Myers bonded out of county jail, on approximately November 6, 1986, authorities transferred him to the Oklahoma City jail. There, Myers gave a detective a sworn statement against Romano and later met with those prosecuting the Thompson murder trial, District Attorney Bob Macy and Lou Keel. Myers subsequently testified against Romano during the Thompson trial. Just a few weeks after that, Myers pled guilty to the felony check charges and received a one-year deferred sentence, notwithstanding the fact that, under Oklahoma law, he was not eligible to receive such a sentence because of his previous felony conviction. See Okla. Stat. tit. 22, § 991c (subsequently amended). Several months later, Myers similarly testified against Romano at the Sarfaty trial.

During Romano’s Sarfaty direct appeal, however, Myers recanted his testimony, instead asserting that Romano had never admitted to Myers any involvement in Thompson’s murder; neither had Romano ever solicited Myers to kill any witnesses. According to Myers, he had testified falsely against Romano in exchange for the State’s favorable disposition of the then pending felony check charges.

Romano challenged Myers’ testimony on direct appeal and in his state post-conviction proceedings. The state appellate court did not specifically address the issue on direct appeal. In state post-conviction proceedings, the Oklahoma Court of Criminal Appeals denied relief without benefit of an evidentiary hearing, noting, in part, that “[t]he issue of an alleged deal between the State and Greg Myers was raised at trial. Myers’ attorney testified there was no deal. The prosecutor testified there was no deal. The trial judge determined there was no deal.” Romano, 917 P.2d at 17.

The trial court, during the Sarfaty sentencing proceeding, did conduct an in camera hearing concerning the existence of a possible deal. Contrary to the state appellate court’s determination, however, none of the prosecutors testified during that in camera hearing. Only one of Myers’ two defense attorneys testified, and he stated only that he was not aware of any deal Myers had with prosecutors to testify against Romano, but that he had entered the case late and did not know what had occurred prior to that time. Romano’s attorney asserted to the trial court that she had spoken with Myers’ first defense attorney, but had not asked that attorney if she had helped arrange any deal between Myers and the State. And, although the trial court, at the conclusion of the in camera hearing, did indicate that it could not find, on that evidence, that Myers had a deal, the trial court did inform Romano’s attorney that she was free to recall Myers and ask him, in the jury’s presence, whether he had a deal. 8 This evidence does not support the state appellate court’s determination. See Jones, 206 F.3d at 958 (determining evidence did not support Oklahoma Court of Criminal Appeals’ finding that victim pleaded for his life). We, therefore, focus our habeas review on the federal district court’s denial of relief.

The government must disclose any understanding or agreement it has with its witnesses. See Giglio, 405 U.S. at 150-51, 154-55, 92 S.Ct. 763. Nonetheless, the district court, after conducting an evidentiary hearing, 9 found that there was no deal between Myers and the State. We review such factual findings only for clear error. See Thomas, 218 F.3d at 1220. Although, on this record, we remain suspicious about whether Myers in fact had a deal with the State, particularly in light of the timing of these events and the significant benefit Myers derived from avoiding a ten-year prison sentence when the State did not charge him as a former felon, we are unable to conclude the district court’s factual finding is clearly erroneous. The district court heard testimony directly from Myers, as well as the prosecutor and Myers’ first defense attorney. We must defer to the district court’s implicit credibility determination underlying its factual finding. See Smith, 197 F.3d at 459 (deferring to state court’s implicit credibility determination); see also Smith v. Secretary of N.M. Dep’t of Corr., 50 F.3d 801, 831 (10th Cir.1995). On that basis, we affirm the district court’s denial of habeas relief. 10

2. Is Romano entitled to habeas relief because Myers recanted his testimony?

After trial, Myers recanted his testimony, asserting instead that Romano had never admitted any involvement in Thompson’s murder; neither had he solicited Myers to kill any witnesses. A prosecutor’s knowing use of false testimony deprives a criminal defendant of due process, warranting a new trial if there is a reasonable likelihood the false testimony affected the judgment. See Giglio, 405 U.S. at 153-54, 92 S.Ct. 763, citing cases; see also Kyles, 514 U.S. at 433 & n. 7, 115 S.Ct. 1555; Bagley, 473 U.S. at 679 & 680 nn. 8 & 9, 105 S.Ct. 3375, citing cases. Even assuming the truth of Myers’ recantation, however, Romano has failed to assert any evidence indicating prosecutors knew Myers’ testimony was false. See, e.g., Van Woudenberg ex rel. Foor v. Gibson, 211 F.3d 560, 569 (10th Cir.2000), cert. denied, — U.S. -, 121 S.Ct. 1117, — L.Ed.2d - (2000); Smith, 197 F.3d at 459-60.

E. Was there sufficient evidence to support the especially heinous, atrocious or cruel aggravating factor?

Woodruff and Romano challenge the sufficiency of the evidence supporting the jury’s finding that Sarfaty’s murder was especially heinous, atrocious or cruel. The question presented is whether the evidence of what Woodruff and Romano did was sufficient to satisfy a constitutional aggravating circumstance. A constitutional aggravating factor channels and limits the capital “sentencer’s discretion in imposing the death penalty” in order to minimize sufficiently the risk of “wholly arbitrary and capricious action.” Maynard v. Cartwright, 486 U.S. 356, 362, 108 S.Ct. 1853, 100 L.Ed.2d 372 (1988). It must provide a principled means by which a sentencer can distinguish between those murders warranting a death sentence and those that do not. See id. at 363, 108 S.Ct. 1858. In Cartwright, the Supreme Court held that Oklahoma’s especially heinous, atrocious or cruel aggravating factor was unconstitutionally vague and overbroad because it failed to give a capital sentencer sufficient guidance — “an ordinary person could honestly believe that every unjustified, intentional taking of human life is ‘especially heinous.’ ” Id. at 363-64, 108 S.Ct. 1853; see also, e.g., Medlock v. Ward, 200 F.3d 1314, 1321 (10th Cir.), cert. denied, - U.S. -, 121 S.Ct. 197, 148 L.Ed.2d 137 (2000).

Oklahoma subsequently narrowed the application of this aggravator to only cases involving torture or serious physical abuse. See Stouffer v. State, 742 P.2d 562, 563 (Okla.Crim.App.1987). The Supreme Court held that was one means by which Oklahoma could constitutionally narrow this aggravating factor. See Cartwright, 486 U.S. at 364-65, 108 S.Ct. 1853.

Oklahoma has since further refined that narrowing. A murder is especially heinous, atrocious or cruel under Oklahoma law if it is “preceded by torture or serious physical abuse. Torture includes the infliction of either great physical anguish or extreme mental cruelty, while physical abuse requires evidence of conscious physical suffering.” Hale, 227 F.3d at 1335 (further quotation, citations omitted); see also, e.g., Thomas, 218 F.3d at 1226-27, citing cases. We have upheld the facial validity of this aggravating factor, as thus narrowed. See, e.g., Medlock, 200 F.3d at 1321.

Recent Oklahoma cases, however, have begun to blur the common understanding of the requisite torture and conscious serious physical suffering, more and more often finding the existence of these elements in almost every murder. See, e.g., Fluke v. State, 14 P.3d 565, 568 & n. 9 (Okla.Crim. App.2000) (noting evidence that victim was aware of attack is sufficient to show torture, citing cases); Washington v. State, 989 P.2d 960, 974-75 (Okla.Crim.App.1999) (holding sufficient evidence supported this aggravator where victim may have consciously suffered for less than one minute after defendant shot her eight times during brief encounter). There is certainly a concern that Oklahoma’s interpretation of its narrowing language could again render this aggravating factor unconstitutional. See Thomas, 218 F.3d at 1228 & n. 17; see also Medlock, 200 F.3d at 1324 (Lucero, J., concurring) (noting that if Oklahoma permitted capital sentencers to find the especially heinous, atrocious or cruel aggravator, “based merely on the brief period of conscious suffering necessarily present in virtually all murders [it] would fail to narrow the sentencer’s discretion” as constitutionally required, citing Godfrey v. Georgia, 446 U.S. 420, 100 S.Ct. 1759, 64 L.Ed.2d 398 (1980)). This court, for example, noted in Thomas that Oklahoma's application of the especially heinous, atrocious or cruel aggravating factor in cases where there is evidence of multiple injuries, but no other evidence of conscious suffering, raised “serious constitutional questions about whether Oklahoma’s heinous, atrocious, or cruel aggravator legitimately narrows the class of those eligible for death.” Id. at 1228 n. 17.

In this case, however, petitioners’ challenge to this aggravating circumstance is an evidentiary one. Jackson’s rational factfinder standard, therefore, governs our review. See, e.g., Hale, 227 F.3d at 1335. The issue thus presented is whether there was sufficient evidence to meet the especially heinous, atrocious or cruel aggravating factor, as Oklahoma has constitutionally narrowed that aggravator. Although the trial evidence does not compel a jury finding of torture or conscious serious physical abuse, we, nevertheless, conclude that the evidence is constitutionally sufficient.

Here, there is evidence Sarfaty suffered abrasions and scrapes, prior to his death, on his left knee and shin, right elbow and back, indicating a struggle did take place. Further, Sarfaty’s wrists and ankles were bound; he was injured due to the force with which his limbs were tied. The fact that Sarfaty’s killers bound his arms and legs is evidence in this case that he was conscious during at least part of the attack; there would be no need to bind a dead person, although there could be a motive to bind an unconscious person to guard against the possibility that the person would regain consciousness.

Further, Sarfaty suffered a number of nonlethal wounds, which the jury could have concluded preceded his loss of consciousness and death. But see Thomas, 218 F.3d at 1227-29 & 1229 n. 17 (rejecting inference that, because killer inflicted multiple blows, victim had to be conscious during part of attack, where there was no evidence victim consciously suffered and two wounds were in fact postmortem). The medical examiner, Dr. Choi, testified that Sarfaty died of strangulation by ligature. It would have taken a minimum of three minutes for Sarfaty to be strangled to death, and two to three minutes before he would lose consciousness from the strangulation.

Sarfaty also suffered five stab wounds, two of which were to the heart and potentially fatal, and five or six blows to the head, lacerating his scalp. While the wounds to the head could have caused immediate loss of consciousness and the injuries to the heart could have been immediately fatal, the evidence supported an inference that Sarfaty was strangled prior to the infliction of these wounds. According to Dr. Choi, she would have expected a great loss of blood from Sarfaty’s stab wounds and the wounds to his head if he were alive at the time those wounds were administered, yet in this case there was very little bleeding from those wounds. Dr. Choi testified that, although this absence of bleeding could result from the stab wounds to the heart immediately stopping its beating, it was more likely because Sarfaty’s heart had been stopped or slowed, and the body sent into shock, due to the earlier strangulation. This evidence would support a finding that Sarfaty consciously suffered serious physical abuse.

Further, this evidence could also support the inference that Sarfaty’s killers bound and tortured him, perhaps in an effort to get him to reveal the location of his money and valuable jewelry. Although this evidence does not compel these inferences, the evidence permits them. In light of this record, therefore, the state appellate court’s decisions upholding the jury’s finding this aggravating factor, see Romano, 847, P.2d at 386-87; Woodruff, 846 P.2d at 1146-47, were not unreasonable. See 28 U.S.C. § 2254(d); 11 cf. Thomas, 218 F.3d at 1226-29 (holding no reasonable factfinder could have found murder was especially heinous, atrocious or cruel where, although victim suffered severe beating, there was no evidence of struggle or defensive wounds, and no evidence indicating order in, which killer inflicted wounds).

F. Was there sufficient evidence supporting the continuing threat aggravating factor?

Romano asserts there was insufficient evidence to support the jury’s finding that there was a probability that he would commit future violent criminal acts such that he should be considered a continuing threat to society. The question presented is whether, viewing the evidence in the light most favorable to the prosecution, a reasonable trier of fact could have found this aggravating factor beyond a reasonable doubt. See, e.g., LaFevers, 182 F.3d at 723 (citing, e.g., Jackson, 443 U.S. 307, 99 S.Ct. 2781, 61 L.Ed.2d 560).

Although Romano presents this as an evidentiary issue, there is also a legal question subsumed within that evidentiary challenge. Romano’s argument suggests that the continuing threat aggravator should be evaluated only within the context of a prison population. Oklahoma courts, however, have held that this aggravating factor is not limited to the threat Romano poses within a prison population. See Salazar v. State, 978 P.2d 315, 326 (Okla.Crim.App.1998) (noting continuing threat aggravator is not limited to any particular segment of society), cert. denied, 528 U.S. 895, 120 S.Ct. 226, 145 L.Ed.2d 190 (1999); see also, e.g., McCarty v. State, 977 P.2d 1116, 1137 (Okla.Crim.App.1998). That is especially true here, where one of the sentencing options before the jury was life imprisonment, which admits the possibility Romano might someday obtain his release on parole. Myers’ testimony that Romano solicited him to kill witnesses indicates Romano may present a continuing threat, within, as well as outside, a prison population.

As so construed, the evidence was more than sufficient to support the jury’s factual détermination. In addition to Myers’ testimony, the evidence indicated that Romano, with Woodruff, robbed and murdered another of Romano’s acquaintances, Lloyd Thompson. In light of this evidence of other dangerous criminal conduct, as well as the facts of Sarfaty’s murder, the state appellate court’s determination that there was sufficient evidence to support this aggravating factor, see Romano, 847 P.2d at 389, 394, was not unreasonable under either 28 U.S.C. § 2254(d)(1) or (2), see, e.g., Hale, 227 F.3d at 1335 n. 17. See also, e.g., Van Woudenberg, 211 F.3d at 574.

G. Did the State’s charging Romano with the continuing threat aggravating factor violate double jeopardy?

In the first Thompson trial, the capital sentencing jury rejected the aggravating factor that Woodruff and Romano presented continuing threats to society. During the Sarfaty trial, the State again charged, and the jury this time found, that both petitioners were continuing threats to society. Romano asserts the State’s charging him with the continuing threat aggravator in the Sarfaty murder prosecution violated double jeopardy because the first Thompson jury, having heard essentially the same evidence presented in the Sarfaty trial, previously rejected that aggravating factor.

The Supreme Court rejected a similar argument in Poland v. Arizona, 476 U.S. 147, 106 S.Ct. 1749, 90 L.Ed.2d 123 (1986). In Poland, the Court determined the relevant inquiry, in applying the Double Jeopardy Clause to a capital sentencing proceeding, is “whether the sentencfer] or the reviewing court has ‘deeid[ed] that the prosecution has not proved its case’ for the death penalty and hence has ‘acquitted’ petitioner[].” Id. at 154, 106 S.Ct. 1749 (quoting Bullington v. Missouri, 451 U.S. 430, 443, 101 S.Ct. 1852, 68 L.Ed.2d 270 (1981)); see also id. at 155, 106 S.Ct. 1749; Osborn v. Shillinger, 997 F.2d 1324, 1327-28 (10th Cir.1993). The Court rejected the premise that “a capital sentencer’s failure to find a particular aggravating circumstance alleged by the prosecution always constitutes an ‘acquittal’ of that circumstance for double jeopardy purposes.” Poland, 476 U.S. at 155, 106 S.Ct. 1749. “Aggravating circumstances are not separate penalties or offenses, but are ‘standards to guide the making of [the] choice’ between the alternative verdicts of death and life imprisonment.” Id. at 156, 106 S.Ct. 1749 (quoting Bullington, 451 U.S. at 438, 101 S.Ct. 1852); see also Walton v. Arizona, 497 U.S. 639, 648, 110 S.Ct. 3047, 111 L.Ed.2d 511 (1990). Like the Arizona capital sentencing scheme at issue in Poland, under Oklahoma law, the capital sentencer’s “finding of any particular aggravating circumstance does not of itself ‘convict’ a defendant (i.e., require the death penalty), and the failure to find any particular aggravating circumstance does not ‘acquit’ a defendant (ie., preclude the death penalty).” Poland, 476 U.S. at 156, 106 S.Ct. 1749.

In further support of this habeas claim, Romano relies in part on Ashe v. Swenson, 397 U.S. 436, 444-46 & 445 n. 10, 90 S.Ct. 1189, 25 L.Ed.2d 469 (1970), holding double jeopardy includes the concept of collateral estoppel. See also Schiro v. Farley, 510 U.S. 222, 232, 114 S.Ct. 783, 127 L.Ed.2d 47 (1994). “[W]hen an issue of ultimate fact has once been determined by a valid and final judgment,” collateral es-toppel precludes the same parties from litigating that same issue in any future lawsuit. Ashe, 397 U.S. at 443, 90 S.Ct. 1189. We need not explore this argument in detail, however, because the first Thompson trial, upon which Romano relies, was not a “valid and final” judgment; rather, the state appellate court overturned the jury’s decision in that case because the state court erred in jointly trying Woodruff and Romano. See United States v. Lacey, 982 F.2d 410, 412 (10th Cir.1992) (judgment has no preclusive effect, for res judicata and collateral estop-pel purposes, if it has been vacated, reversed or set aside); see also, e.g., United States v. Sackett, 114 F.3d 1050, 1052 (10th Cir.1997) (state judgment that had been set aside had no preclusive effect in federal civil action); In re Hedged-Investments Assocs., Inc. (Sender v. Nancy Elizabeth R. Heggland Family Trust), 48 F.3d 470, 472-73 (10th Cir.1995) (no final judgment existed, as required by collateral estoppel, where earlier decision had been reversed on appeal). See generally Schiro, 510 U.S. at 232, 114 S.Ct. 783 (holding preclusive effect of the state court’s judgment, in a habeas action, is matter of federal law). Collateral estoppel principles, therefore, do not support Romano’s claim for relief. See Goff v. United States, 446 F.2d 623, 627 (10th Cir.1971).

H. Was there sufficient evidence that Romano killed Sarfaty to avoid arrest or prosecution?

Under Oklahoma law, the focus of this aggravator is on the defendant’s intent, which can be proved either by his own statements or through circumstantial evidence. See, e.g., Hale, 227 F.3d at 1334. Additionally, Oklahoma requires proof of a predicate crime, other than the murder, for which the defendant seeks to avoid arrest or prosecution. See id. The state appellate court’s determination that there was sufficient evidence to support this aggravating factor, see Romano, 847 P.2d at 387, was reasonable under either 28 U.S.C. § 2254(d)(1) or (2). See Hale, 227 F.3d at 1335 n. 17.

Tracy Greggs testified that Romano asked him to help rob Sarfaty. Romano further told Greggs that Romano would have to kill Sarfaty because he knew and could identify Romano. Although Romano challenged Greggs’ credibility, the jury appropriately resolved that dispute. See, e.g., Valdez, 219 F.3d at 1238 & n. 4. Furthermore, Sarfat/s robbery provides the requisite predicate crime for which Romano sought to avoid arrest or prosecution. There is, therefore, ample evidence to support this aggravating factor. See, e.g, Hale, 227 F.3d at 1334-35; LaFevers, 182 F.3d at 723; Moore v. Reynolds, 153 F.3d 1086, 1115 (10th Cir.1998).

I. Was the failure of Woodruffs attorney to investigate and present mitigating evidence at the sentencing stage such that Woodruff received ineffective representation?

Woodruff argues there was mitigating evidence, not presented during sentencing, that his natural mother abandoned him at age eighteen months and his natural father’s girlfriend then gave him away to a customer at the coffee shop where she worked. Further, at the time the Wood-ruffs adopted petitioner, there was evidence of cigarette burns on his legs and diaper area. His natural mother had abused alcohol during her pregnancy with him and, as a toddler, he had almost drowned while left in the care of his seven-year-old brother. Woodruff also contends trial counsel should have had Woodruffs mental health evaluated because such an exam would have shown that he had brain damage and a fourteen year old’s reasoning ability, was a follower rather than a leader, and would not pose a threat while in a structured prison environment. Woodruff now argues his attorney’s failure to investigate and present this evidence amounted to ineffective assistance of counsel.

The Oklahoma Court of Criminal Appeals held Woodruff had waived this claim because he failed to raise it on direct appeal. See Woodruff, 910 P.2d at 351-52. Contrary to respondent’s assertion, however, the trial record before the state appellate court on direct appeal did not contain sufficient evidence concerning either the extent of defense counsel’s sentencing preparation or the additional evidence Woodruff asserts counsel should have discovered and presented. The State’s procedural bar, therefore, is inadequate to preclude federal habeas review. See English v. Cody, 146 F.3d 1257, 1264 (10th Cir.1998); see also McGregor v. Gibson, 219 F.3d 1245, 1252-53 (10th Cir.2000), reh’g granted on other grounds. Because the state appellate court did not address the merits of Woodruffs ineffective assistance claim, our review is de novo. See Thomas, 218 F.3d at 1220; Smith, 197 F.3d at 461.

The sentencing stage is the most critical phase of a death penalty case. Any competent counsel knows the importance of thoroughly investigating and presenting mitigating evidence. “As a practical matter, the defendant probably has little or no chance of avoiding the death sentence unless the defense counsel gives the jury'-, something to counter both the horror of the crime and the limited information the prosecution has introduced about the defendant.” Jonathon P. Tomes, Damned If You Do, Damned If You Don’t: The Use of Mitigation Experts in Death Penalty Litigation, 24 Am. J. Crim. L. 359, 364 (1997). Mitigating evidence plays an overwhelmingly important role in the “just imposition of the death penalty.” Mayes, 210 F.3d at 1288. It “affords an opportunity to humanize and explain — to individualize a defendant outside the constraints of the normal rules of evidence.” Id. In light of its importance, investigation and preparation of a case in mitigation should begin prior to trial, well before any determination of guilt at the first stage. See Williams, 529 U.S. at 395, 120 S.Ct. 1495 (noting counsel’s deficient preparation for sentencing did not begin until one week prior to trial).

To prevail here on this ineffective assistance claim, Strickland requires Woodruff to establish that counsel’s deficient performance prejudiced Woodruffs defense. See Strickland, 466 U.S. at 687, 104 S.Ct. 2052. The “ultimate focus of inquiry must be on the fundamental fairness” of the challenged proceeding. Id. at 696, 104 S.Ct. 2052.

Counsel’s representation will be constitutionally deficient if it “fell below an objective standard of reasonableness,” measured “under prevailing professional norms,” and considered in light of all of the circumstances. Id. at 688, 104 S.Ct. 2052. “Judicial scrutiny of counsel’s performance must be highly deferential[,]” making every effort “to eliminate the distorting effects of hindsight, to reconstruct the circumstances of counsel’s challenged conduct, and to evaluate the conduct from counsel’s perspective at the time.” Id. at 689, 104 S.Ct. 2052. We “indulge a strong presumption that counsel’s conduct falls within the wide range of reasonable professional assistance; that is, the defendant must overcome the presumption that, under the circumstances, the challenged action might be considered sound trial strategy.” Id. (further quotation omitted).

To establish prejudice, Wood-ruff must show there is a “reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different. A reasonable probability is a probability sufficient to undermine confidence in the outcome.” Id. at 694, 104 S.Ct. 2052. Woodruff need not, however, go so far as to establish that counsel’s deficient performance “more likely than not altered the outcome.” Nix v. Whiteside, 475 U.S. 157, 175, 106 S.Ct. 988, 89 L.Ed.2d 123 (1986); see also Strickland, 466 U.S. at 693, 104 S.Ct. 2052. In the context of a capital sentencing proceeding, the relevant inquiry is “whether there is a reasonable probability that, absent the errors, the sentencer ... would have concluded that the balance of aggravating and mitigating circumstances did not warrant death.” Strickland, 466 U.S. at 695, 104 S.Ct. 2052.

This court can affirm the denial of habeas relief on whichever Strickland prong is the easier to resolve. See Smith v. Robbins, 528 U.S. 259, 286 n. 14, 120 S.Ct. 746, 145 L.Ed.2d 756 (2000) (citing Strickland, 466 U.S. at 697, 104 S.Ct. 2052). Here, we determine trial counsel’s performance was not deficient. 12

“Because [the adversarial] testing process generally will not function properly unless defense counsel has done some investigation into the prosecution’s case and into various defense strategies, ... ‘counsel has a duty to make reasonable investigations or to make a reasonable decision that makes particular investigations unnecessary.’ ” Kimmelman v. Morrison, 477 U.S. 365, 384, 106 S.Ct. 2574, 91 L.Ed.2d 305 (1986) (quoting Strickland, 466 U.S. at 691, 104 S.Ct. 2052). “[Strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable.” Strickland, 466 U.S. at 690, 104 S.Ct. 2052. “[A]nd strategic choices made after less than complete investigation are reasonable precisely to the extent that reasonable professional judgments support the limitations on investigation.” Id. at 690-91, 104 S.Ct. 2052.

Nonetheless, “[t]he reasonableness of counsel’s actions may be determined or substantially influenced by the defendant’s own statements or actions. Counsel’s actions are usually based, quite properly, on informed strategic choices made by the defendant and on information supplied by the defendant.” Id. at 691, 104 S.Ct. 2052; see also, e.g., James v. Gibson, 211 F.3d 543, 557 (10th Cir.2000), cert. denied, — U.S. -, 121 S.Ct. 886, 148 L.Ed.2d 794 (2001); Wallace v. Ward, 191 F.3d 1235, 1247 (10th Cir.1999), cert. denied, — U.S. -, 120 S.Ct. 2222, 147 L.Ed.2d 253 (2000). In this case, Wood-ruff instructed his attorney not to have his parents and friends testify. Although trial counsel has an independent duty to investigate and make a case in mitigation, counsel also has to be responsive to the wishes of his client, see Wallace, 191 F.3d at 1247-48 (holding counsel’s performance during capital sentencing proceeding was not deficient, where counsel acquiesced in petitioner’s wishes not to present any mitigating evidence or challenge State’s evidence).

In any event, notwithstanding Wood-ruffs wishes, trial counsel did present a significant amount of mitigating evidence on Woodruffs behalf, including the testimony of a family friend, Woodruffs mother and his girlfriend. The family friend testified about Woodruffs Little League activities as a youth, his love and concern for the witness’s children, and Woodruffs carpentry work with his adoptive father. Woodruffs girlfriend also testified about Woodruffs love for children and for people in general, as well as his work history, his training as a gemologist and the fact that he had never harmed her, nor had she seen him harm anyone else. Woodruffs mother testified that she and her husband adopted Woodruff at age eighteen months and that she had never met his natural parents. She asserted that he had a “very normal upbringing.” She also described Woodruffs anguish upon discovering, at age eighteen, that he was adopted.

Woodruff himself related details concerning his childhood, schooling, sports activities, work history, military service, gemology training and his love of children. He described his childhood as “outstanding.” He further explained to the jury his difficulty coping with the discovery that he was adopted. He informed the jury that he did not like having his friends and family testifying on his behalf and that his defense attorney had presented what mitigating testimony he had against Wood-ruffs wishes.

The childhood evidence Woodruff now wishes his attorney had presented was potentially in tension with his attorney’s logical strategy to portray Wood-ruffs childhood as normal and happy. Showing a dysfunctional and difficult childhood basically asks a jury not to hold a defendant responsible for his own dysfunctional or antisocial conduct. On the other hand, showing a defendant had a normal and socially acceptable background goes toward showing the defendant does not present a continuing threat to society because his baseline personality is not antisocial. Counsel’s strategic choice here to portray Woodruffs childhood as normal and happy was “well within the range of professionally reasonable judgments.” Strickland, 466 U.S. at 699, 104 S.Ct. 2052. That reasonable strategic choice justifies counsel’s failure to search further for any information of a possibly dysfunctional background. See id.

Moreover, nothing in the record would have reasonably required a prudent attorney to do the additional investigation Woodruff now suggests. See Mayes, 210 F.3d at 1289 n. 3 (holding counsel’s failure to obtain psychiatric evaluation did not amount to deficient performance where there was nothing in record that would have caused reasonable attorney to believe petitioner’s mental condition was potentially mitigating factor). Through his pretrial preparation, counsel did discover that Woodruff had been adopted, but both Woodruff and his mother asserted he had a happy and normal childhood. Nothing the attorney knew suggested an abusive childhood prior to the adoption. See Strickland, 466 U.S. at 691, 104 S.Ct. 2052 (noting where petitioner has given counsel facts concerning potential defense or reason to believe pursuing further investigation would be fruitless, petitioner may not later challenge reasonableness of counsel’s failure to investigate those matters further).

Nor did counsel’s failure to pursue further psychiatric evaluation amount to deficient performance. See Mayes, 210 F.3d at 1289 n. 3. There was no indication, at the time of trial, that a mental status exam would produce mitigating evidence. A report compiled during a competency evaluation prior to his first trial indicated he did not suffer from any psychiatric disorders and his test results at that time appeared normal.

Considering all of the above, we do not believe that the lawyer’s failure to investigate Woodruffs adoption and to obtain a psychiatric evaluation was “ ‘outside the wide range of professionally competent assistance.’ He ‘has made no showing that the justice of his sentence was rendered unreliable by a breakdown in the adversary process caused by deficiencies in counsel’s assistance.’ ” Burger v. Kemp, 483 U.S. 776, 795-96, 107 S.Ct. 3114, 97 L.Ed.2d 638 (1987) (quoting Strickland, 466 U.S. at 690, 700, 104 S.Ct. 2052).

IV. CONCLUSION

For the reasons stated above, we, therefore, AFFIRM the district court’s denial of habeas relief on Woodruffs and Romano’s first degree murder and armed robbery convictions and their resulting death sentences.

1 . Although these related appeals have not been consolidated, we address them in one opinion because of the similarity of the issues raised by the parties.

2 . This court's authority is divided as to whether, under AEDPA, we review a sufficiency-of-the-evidence issue as a legal determination under 28 U.S.C. § 2254(d)(1) or a factual finding under § 2254(d)(2) and (e)(1). See, e.g., Mayes v. Gibson, 210 F.3d 1284, 1293 (10th Cir.), cert. denied, - U.S. -, 121 S.Ct. 586, 148 L.Ed.2d 501 (2000); see also Hale v. Gibson, 227 F.3d 1298, 1335 n. 17 (10th Cir.2000) (reviewing sufficiency of evidence to support capital jury’s finding aggravating factor). We need not resolve this issue here because the result would be the same under either AEDPA section. See, e.g., Hale, 227 F.3d at 1335 n. 17; Mayes, 210 F.3d at 1293.

3 . The Chambers line of cases does not create a constitutional entitlement to present any evidence critical to a criminal defendant's defense. See Montana v. Egelhoff, 518 U.S. 37, 51-53, 116 S.Ct. 2013, 135 L.Ed.2d 361 (1996); see also United States v. Scheffer, 523 U.S. 303, 308, 316, 118 S.Ct. 1261, 140 L.Ed.2d 413 (1998). Rather, "the introduction of relevant evidence can be limited by the State for a 'valid' reason.” Egelhoff, 518 U.S. at 53, 116 S.Ct. 2013. In this case, however, because the trial court did not exclude any evidence critical to the defense, we need not weigh the State's justification for this eviden-tiary rule with petitioners’ right to present a defense. See Scheffer, 523 U.S. at 308-09, 118 S.Ct. 1261; see also Boyd v. Ward, 179 F.3d 904, 921 (10th Cir.1999) (holding evidence excluded from capital sentencing stage would have had no effect on trial's outcome, without further addressing State's proffered interests in having evidence excluded), cert. denied, 528 U.S. 1167, 120 S.Ct. 1188, 145 L.Ed.2d 1093 (2000).

4 . At oral argument, the State argued the Oklahoma appellate court did not treat petitioners' post-conviction claims inconsistently because the precise issue in Romano’s state post-conviction proceeding was different than the issue in Woodruffs post-conviction action. The State asserts Romano’s state post-conviction application presented the issue now before this court, whereas Woodruff’s state post-conviction proceeding addressed only an ancillary issue-whether the state courts should grant him an evidentiary hearing on this claim. We disagree. The Oklahoma Court of Criminal Appeals noted Woodruff “claims newly discovered evidence regarding the temperature in the victim's apartment after the murder warranted an evidentiary hearing ..., and presumably post-conviction relief.” Woodruff, 910 P.2d at 350. That court denied Woodruff an evidentiary hearing because “there was no reasonable probability that if the evidence had been introduced, different results would have been reached.” Id. at 351. Thus, in denying Woodruff an evidentia-ry hearing, the Oklahoma Court of Criminal Appeals ultimately reached and rejected the merits of this claim. That was inconsistent with its refusal to address the merits of Roma no’s post-conviction claim because he failed to raise it on direct appeal.

5 . While the State had an open file policy, that policy would not have revealed Detective Martin's undocumented and as yet unarticu-lated opinion. Nevertheless, "defense counsel may reasonably rely on that file to contain all materials the State is constitutionally obligated to disclose under Brady." Strickler, 527 U.S. at 283 n. 23, 119 S.Ct. 1936.

6 . Giglio v. United States, 405 U.S. 150, 92 S.Ct. 763, 31 L.Ed.2d 104 (1972).

7 . Although Woodruff also raises a habeas challenge to Myers' testimony, that testimony was directed primarily against Romano. In any event, because we determine Romano is not entitled to habeas relief on this claim, Woodruff would not be, either.

Woodruff further asserts that his trial attorney was ineffective for failing to investigate adequately whether Myers had a deal with prosecutors. As we have said, however, Myers' testimony at sentencing had only very limite,d impact upon Woodruff. Myers’ testimony only indirectly linked Woodruff to Thompson's murder. And it implicated only Romano in the scheme to have witnesses to that murder killed. Moreover, we affirm here the district court's factual finding that there was no deal between Myers and the State. For these reasons, Woodruff cannot establish the requisite prejudice to succeed on this ineffective-assistance claim. See Strickland, 466 U.S. at 687, 104 S.Ct. 2052.

8 . The transcript actually indicates that, in light of the evidence, the trial court could find that there was a deal. Nonetheless, the parties have consistently treated this as a typographical error, interpreting that passage to indicate instead that the trial court could not find that there had been a deal. This interpretation is consistent with the rest of the trial court’s remarks.

9 . Because the district court already conducted a hearing, we need not now address whether Romano was entitled to an evidentia-ry hearing under 28 U.S.C. § 2254(e)(2). See Williams v. Taylor, 529 U.S. 420, 424, 437, 120 S.Ct. 1479, 146 L.Ed.2d 435 (2000) (holding, under § 2254(e)(2), habeas petitioner who failed diligently to pursue development of factual basis of habeas claim in state court must meet § 2254(e)(2)'s stringent requirements before being entitled to federal evidentiary hearing). Romano did have an opportunity, at trial, to cross-examine Myers concerning whether he was testifying pursuant to a deal, but Romano failed to take advantage of that opportunity. And, while Romano did raise the possibility that Myers had a deal during the trial's in camera hearing, he did not diligently pursue that inquiry. See Smith, 235 F.3d at 1275 (noting, in dicta, habeas petitioner had failed to develop, in state court, evidence supporting her claim that State did not disclose complete transcript of its interview with key government witness, where petitioner received state evi-dentiary hearing, but failed to present any evidence there supporting her transcript claim); see also Smallwood v. Gibson, 191 F.3d 1257, 1266 (10th Cir.1999) (holding petitioner failed diligently to seek to develop factual basis of habeas claim where, although Oklahoma appellate court invited petitioner to assert ineffective assistance claim, petitioner failed to do so), cert. denied, - U.S. -, 121 S.Ct. 88, 148 L.Ed.2d 49 (2000). Nonetheless, since the district court held an evi-dentiary hearing and we have benefitted from that court’s fact finding, there is no utility at this time in addressing whether Romano was entitled to that hearing.

Moreover, respondent, on appeal to this court, does not challenge the district court’s decision to hold an evidentiary hearing. See Weaver v. Thompson, 197 F.3d 359, 362 n. 3 (9th Cir.1999).

10 . The district court, alternatively, held that even if the State had violated Giglio and Brady by failing to disclose a deal it had with Myers, there was no reasonable probability that, had such a deal been brought to the sentencing jury’s attention, the jury would have returned a different sentencing determination. See, e.g., Foster, 182 F.3d at 1183, 1192 (applying same standard in pre-AEDPA case). Myers’ testimony strongly supported the State’s allegation that Romano presented a continuing threat to society. Although the State did present other evidence that Romano was involved in Thompson's murder, Myers’ testimony that Romano, from his prison cell, had sought to have witnesses in the Thompson case murdered would have had to impact strongly the jury’s continuing threat determination. See, e.g., Smith, 50 F.3d at 830 (holding police report was material where its nondisclosure clearly impacted defense preparation and presentation and could reasonably have affected trial’s outcome).

11 . We reach the same result under either § 2254(d)(1) or (2). See, e.g., Hale, 227 F.3d at 1335 n. 17.

12 . Although the district court did not address whether counsel's performance was deficient, this court is free to affirm on any basis supported by the record. See United States v. Sandoval, 29 F.3d 537, 542 n. 6 (10th Cir.1994).