According to the prosecution, Kenneth Phoenix "almost committed the perfect crime . . . except for one mistake." Based on that mistake -- a blood-soaked fingerprint left near the scene -- Phoenix was convicted of the first degree murder of Raymond Green. Although Phoenix's defense counsel cross-examined the forensic serologist and fingerprint experts presented by the Commonwealth of Massachusetts, he did not call defense experts to further contradict their testimony. Phoenix now claims that his attorney's decision not to call such experts denied him his constitutional right to effective assistance of counsel. Having had his petition for a writ of habeas corpus denied by the district court, he appeals to this Court. For the reasons stated herein, we affirm the decision of the district court.

As our previous decision, Phoenix v. Matesanz, 189 F.3d 20, 22-24 (1st Cir. 1999), summarizes this case's procedural history prior to the district court's denial of habeas, we need provide only a brief summary here.

Raymond Green was a plant manager at the Belchertown State School. On August 4, 1986, he was found dead in his office, shot five times. Police found pieces of a green scouring pad on his face and on the floor of his office, as well as in a metal drum on the first floor of the plant. In that drum police also found a crumpled brown paper bag. Fingerprints and blood were found on the paper bag. Two experts presented by the Commonwealth testified at trial that the identifiable fingerprints on the bag belonged to Phoenix, who was one of 63 employees supervised by Green. Expert serologist Dr. Moses Schanfield testified that the blood from the only successfully tested blood stain was consistent with Green's blood and inconsistent with Phoenix's blood. Based largely on these two pieces of incriminating evidence, Phoenix was convicted.

Although Phoenix's counsel, William Bennett, had retained a forensic serologist and a fingerprint analyst, he called neither to testify. The serologist, Dr. Brian Wraxall, later filed an affidavit stating that he would have testified that the allotype blood test performed by Schanfield yielded scientifically meaningless results. Wraxall would have further testified that no scientific basis existed to conclude that the tested stain was either consistent with Green's blood or inconsistent with Phoenix's blood. The fingerprint expert, Herbert MacDonnell, filed an affidavit stating that the fingerprint lacked sufficient detail to be identified either as Phoenix's or not Phoenix's.

After Phoenix's direct appeals were denied, ultimately by the Massachusetts Supreme Judicial Court, Commonwealth v. Phoenix, 567 N.E.2d 193 (Mass. 1991), he filed a second motion for a new trial claiming that he had been denied effective assistance of counsel based on Bennett's failure to call Wraxall and MacDonnell. The Superior Court denied his motion, as did a single gatekeeper justice of the SJC. The federal district court then found that Phoenix was not procedurally barred from filing a habeas petition based on ineffective assistance of counsel, a decision we affirmed. See Phoenix v. Matesanz, 189 F.3d 20 (1st Cir. 1999). After remand, the district court ruled on the merits of Phoenix's petition, ultimately concluding that the state court decisions did not involve an unreasonable application of the Strickland v. Washington, 466 U.S. 668 (1984), standard for determining ineffective assistance of counsel.

In enacting the Antiterrorism and Effective Death Penalty Act (AEDPA), Congress placed new restrictions on the power of federal courts to grant writs of habeas corpus to state prisoners. As relevant here, 28 U.S.C. 2254(d)(1) was revised to provide:

(d) An application for a writ of habeas corpus on behalf of a person in custody pursuant to the judgment of a State court shall not be granted with respect to any claim that was adjudicated on the merits in State court proceedings unless the adjudication of the claim (1) resulted in a decision that was contrary to, or involved an unreasonable application of, clearly established Federal law, as determined by the Supreme Court of the United States.

Until last Term, the Supreme Court had not been presented with the opportunity to elucidate the meaning of the revised provision, and the various courts of appeals, including this Court, had been left to take our best shots. See, e.g., Green v. French, 153 F.3d 865 (4th Cir. 1998); O'Brien v. DuBois, 145 F.3d 16 (1st Cir. 1998); Drinkard v. Johnson, 97 F.3d 751 (5th Cir. 1996); Lindh v. Murphy, 96 F.3d 856 (7th Cir. 1996). However, in Williams v. Taylor, 120 S. Ct. 1495 (2000), the Supreme Court provided the first explicit guidance on the construction of this section of the AEDPA, fortuitously in the very context of a claim of ineffective assistance of counsel.1 It is to this decision that we must now turn.

Addressing a case out of the Fourth Circuit, Williams v. Taylor, 163 F.3d 860 (4th Cir. 1998), the Supreme Court generally upheld Green's interpretation of 2254(d)(1), albeit with several significant differences. The Court began by sustaining the Green conception of the "contrary to" clause, holding that the clause applied in two types of situations. First, "a state-court decision will certainly be contrary to our clearly established precedent if the state court applies a rule that contradicts the governing law set forth in [Supreme Court] cases." Williams, 120 S. Ct. at 1519. For example, requiring a petitioner to meet a higher burden than that provided for in Strickland would be contrary to clearly established Supreme Court precedent. See id. Second, "a state-court decision will also be contrary to this Court's clearly established precedent if [it] confronts a set of facts that are materially indistinguishable from a [Supreme Court decision] and nevertheless arrives at a [different result]." Id. at 1519-20. However, Justice O'Connor noted that the run-of-the-mill state-court case, applying the correct legal rule to a new set of facts, would not fit comfortably within the "contrary to" clause of 2254(d)(1). To place such a case within that clause would sap the "unreasonable application" clause of any meaning. See id.

The Court then found that the Fourth Circuit was again generally correct as to its interpretation of the "unreasonable application" clause. Green held that an "unreasonable application" of Supreme Court precedent occurs when (i) "the state court identifies the correct governing legal rule . . . but unreasonably applies it to the facts of the particular [case]," and (ii) when "the state court either unreasonably extends a legal principle from [Supreme Court] precedent to a new context where it should not apply or unreasonably refuses to extend that principle to a new context where it should apply." Id. at 1520 (citing Green, 153 F.3d at 869-70). The Court endorsed the first approach, see id. at 1520-21, while withholding judgment on the second, see id. at ____ __ ____, 120 S.Ct. at 1521.

However, the Court refused to endorse the Fourth Circuit's determination of "what exactly qualifies" as an unreasonable application of law under 2254(d)(1). Green limited unreasonable applications to those cases where all reasonable jurists would agree that the state court application of law is unreasonable. See Green, 153 F.3d at 870. The Supreme Court concluded that such an approach provides little assistance to federal courts, and is in fact misleading in its subjectivity. See Williams, 120 S. Ct. at 1521 (noting that under such a standard, any conflicting authority, including a 2-1 split, would foreclose review). Instead, the federal habeas court "should ask whether the state court's application of clearly established federal law was objectively unreasonable." Id. Although "unreasonable" may be difficult to define, Justice O'Connor noted that it is a term familiar to the legal world and to federal judges. At the very least, an unreasonable application of federal law differs from an incorrect application of federal law. See id. at 1522 (citing Wright v. West, 505 U.S. 277, 287 (1992), for this distinction). In sum, "[u]nder 2254(d)(1)'s 'unreasonable application' clause, a federal habeas court may not issue the writ simply because that court concludes in its independent judgment that the relevant state-court decision applied clearly established federal law erroneously or incorrectly. Rather, that application must also be unreasonable." Id.

Finally, Justice O'Connor noted that so-called "old rules" under Teague v. Lane, 489 U.S. 288 (1989), would qualify as "clearly established Federal law," the only caveat being that post-AEDPA, such "old rules" could only stem from the Supreme Court's jurisprudence. See Williams, 120 S. Ct. at 1523.

The Supreme Court then applied its Williams v. Taylor analysis to the denial of effective assistance of counsel, concluding that "[i]t is past question that the rule set forth in [Strickland] qualifies as 'clearly established law, as determined by the Supreme Court of the United States.'" Williams, 120 S. Ct. at 1512. The Court noted that the Strickland test, "of necessity," requires a case-by-case examination of the evidence, but held that such case-specific concerns "obviate neither the clarity of the rule nor the extent to which the rule must be seen as 'established' by this Court." Id.

The Court in Williams nicely summarized the relevant aspects of the Strickland test:

[A] violation of the right [to effective assistance of counsel] has two components:

"First, the defendant must show that counsel's performance was deficient. This requires showing that counsel made errors so serious that counsel was not functioning as the 'counsel' guaranteed the defendant by the Sixth Amendment. Second, the defendant must show that the deficient performance prejudiced the defense. This requires showing that counsel's errors were so serious as to deprive the defendant of a fair trial, a trial whose result is reliable." Strickland, 466 U.S. at 687.

To establish ineffectiveness, a "defendant must show that counsel's representation fell below an objective standard of reasonableness." Id. at 688. To establish prejudice, he "must show that 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.

Williams, 120 S. Ct. at 1511-12.

In Strickland itself the Supreme Court spoke in more detail about the deferential level of scrutiny involved in this review, particularly with respect to potentially strategic decisions made by counsel. The Court cautioned as to the use of hindsight: "It is all too tempting for a defendant to second-guess counsel's assistance after conviction . . . , and it is all too easy for a court, examining counsel's defense after it has proved unsuccessful, to conclude that a particular act or omission of counsel was unreasonable." Strickland, 466 U.S at 689; see also United States v. Natanel, 938 F.2d 302, 310 (1st Cir. 1991) (cautioning that "[the fact that] counsel's selection of a stratagem may, in retrospect, have proved unsuccessful, or even unwise, is not the issue"). The defendant, as a result, must "overcome the presumption that, under the circumstances, the challenged action 'might be considered sound trial strategy.'" Id. (quoting Michel v. Louisiana, 350 U.S. 91, 101 (1955)). Specifically, a court must judge the reasonableness of counsel's challenged conduct on the facts of the case at the time of that conduct. Moreover, "counsel is strongly presumed to have rendered adequate assistance and made all significant decisions in the exercise of reasonable professional judgment." Id. at 690. In short, "strategic choices made after thorough investigation of law and facts relevant to plausible options are virtually unchallengeable." Id. (emphasis added).2

Under Strickland, we must ask whether defense counsel's decision not to call either defense expert was an act that falls below "an objective standard of reasonableness." However, if the state court applied the correct test, we cannot answer this question de novo. See Williams, 120 S. Ct. at 1519-20 (holding that application of the Strickland test is under the unreasonable application prong of 2254(d)(1)). Although Justice Fried, as single gatekeeper justice of the SJC, did not cite Strickland in his analysis, it is clear from his language that his decision denying Phoenix's motion was based on the judgment that the failure to call Wraxall was a strategic choice:

It is this strategy which defendant now claims amounted to ineffective assistance of counsel. The defendant claims that if only the expert witness, Dr. Wraxall, had been called to testify, it would have had a damaging effect on the Commonwealth's own expert testimony regarding the crucial blood testimony. But the subject of the reliability of the Commonwealth's testimony on this score had been so thoroughly canvassed both at trial, by cross-examination by defendant's counsel, and then examined by the full court on plenary review, that I think the claim of ineffectiveness of counsel is so far-fetched as to be insubstantial.

Phoenix v. Commonwealth, No. SJ-96-0571, at 4-5 (SJC Memorandum, Fried, J., May 8, 1997). The strategy to which Justice Fried referred was described in the prior decision of the full SJC, see Commonwealth v. Phoenix, 409 Mass. 408, 421 n.8 (1991). In denying leave to admit into the record on direct appeal the Wraxall affidavit, the SJC referred to defendant's arguements that he did not have Wraxall testify because of the expense and because of his belief that Schandfield's testimony would expose the unreliability of the critical test. The SJC concluded that "[t]hese considerations were purely ones of trial strategy, and, therefore, there is no reason to permit the filing of the affidavit."Id.

As for the fingerprint evidence, Judge Moriarty plainly found that the failure to call MacDonnell was a strategic one. See Phoenix, March 18 Memorandum, at 8 ("[Bennett] made a tactical decision - and probably a wise one.").

Hence, we must assess whether the respective state court applications of Strickland to these ineffective assistance claims was in fact objectively unreasonable. In other words,were the state court findings that Bennett made strategic choices in not calling either Wraxall or MacDonell reasonable determinations? Cf. Tucker v. Catoe, 221 F.3d 600 (4th Cir.2000)(applying Williams/Strickland analysis in a similar manner); Barnabeli v. Angelone, 214 F.3d 463, 469 (4th Cir.2000)(recognizing "the legal effect of the prior state court adjudication while independently reviewing the issused raised").

We make this assessment under the guidance of our decision in Lema v. United States, 987 F.2d 48, 53 (1st Cir. 1993).3 In Lema, we denied an ineffective assistance claim based on counsel's failure to call three witnesses proposed by the defendant. We noted that "[t]he decision whether to call a particular witness is almost always strategic, requiring a balancing of the benefits and risks of the anticipated testimony." Id. at 54. "Where the prosecution's case is less than compelling . . . the risk of 'rocking the boat' may warrant a decision by trial counsel to forego the presentation of further defense testimony, even favorable testimony." Id. (citing Johnson v. Lockhart, 921 F.2d 796, 800 (8th Cir. 1990)). Moreover, choices in emphasis during cross-examination are prototypical examples of unchallengeable strategy. See Matthews v. Rakiey, 54 F.3d 908, 916-18 (1st Cir. 1995).

We now examine each claim in turn.

In response to the prosecution experts' identification of his client's prints and exclusion of another suspect's prints, Bennett cross-examined the Commonwealth's two fingerprint experts both at voir dire and at trial. Bennett challenged whether the paper bag had been treated for prints at the appropriate time and in the appropriate manner; he exposed that the expert worked only from photographs, rather than from the paper bag itself; and he raised questions about the chain of custody. Bennett also spent a significant amount of time questioning the methodology of one of the Commonwealth experts (Shiflett), particularly with regard to discrepancies between the paper bag print and the comparison print. In particular, his cross-examination prompted Shiflett to testify that the quality of the print had limited his observations and that the print on the paper bag was "similar to" but not "identical" to the print taken from Phoenix used for purposes of comparison.

Defense expert MacDonnell submitted an affidavit stating that, in his opinion, the "quality of the fingerprint evidence is inadequate . . . to make a positive ID," and as a result he could not "conclude with certainty that the unknown [print] is not from Phoenix." MacDonnell could only testify in response to the prosecution's experts that identification of the fingerprint was impossible, in his opinion. He could not identify the fingerprint as that of another potential suspect. He could not say, for certain, that the fingerprint did not belong to Phoenix. Although Bennett had not thoroughly discredited the Commonwealth's fingerprint experts at trial, he had made substantial progress in this direction. It is possible that, given his progress during cross-examination, the limitations of MacDonnell's testimony, and the threat of exhaustive cross-examination of his own expert, Bennett made a strategic decision not to call him. Judge Moriarty of the Massachusetts Superior Court so concluded. We cannot say that such a finding was objectively unreasonable, even if we might have found differently. See Williams, 120 S. Ct. at 1522 (distinction between "incorrect" and "unreasonable"); see also Tucker, 221 F.3d at 614 (finding that on a close issue, the appellate court can disagree with the state court without the state court decision having been unreasonable); Barnabei, 214 F.3d at 469 (same).

We also cannot conclude that Bennett's failure to send MacDonnell additional photographs for review during and after trial constituted ineffective assistance of counsel. Although it turned out that MacDonnell's opinion of the additional photographs would probably have been helpful at trial, Bennett's choice not to pursue further investigation on this front was a reasonable one at trial. MacDonnell's partially provided expert opinion had lowered Bennett's expectations and Bennett had adequately cross-examined the Commonwealth witnesses. Even with the stronger evidence, MacDonnell could still not testify that the print did not belong to Phoenix. Defense counsel is allowed to make strategic decisions, within the wide bounds of professional competence, as to which leads to follow up, and on which areas to focus his energies. This is especially true during trial, when time is short. See, e.g., Genius v. Pepe, 147 F.3d 64, 67 (1st Cir. 1998). Given our understanding of Bennett's failure to send MacDonnell additional photographs as a strategic choice at trial, we cannot find the state court's decision on this issue an unreasonable application of Strickland.

The blood evidence presents a more difficult case, simply because the defense expert affidavit promised much more. Wraxall offered to testify that: (1) the blood on the paper bag was not the result of "blowback" (i.e., blood splattered from the victim after he was shot); (2) the results of the May 11 test were "scientifically meaningless" and "not the proper basis for a scientific opinion;" (3) prosecution expert Schanfield's methodology was unreliable in important aspects; (4) taken as a whole, the blood evidence was not consistent with the victim, and Schanfield's statement to the contrary was misleading; and (5) the blood could have been taken from anyone in the population. Wraxall also noted that he would not have been able to perform independent testing on the samples, because Schanfield had irreparably altered them and the prosecutors had requested their return.

At the voir dire, Bennett spent several hours cross-examining Schanfield and posed significant questions to the reliability of his testimony. As a result, the trial judge only allowed Schanfield to make conclusions based on the May 11 tests. At trial, Bennett engaged in further cross-examination of Schanfield. Bennett was able to (1) expose problems with the control samples; (2) suggest what the problems with controls might mean for the reliability of the test as a whole; (3) prompt Schanfield to testify that he could not say, based on the May 11 test, that the blood was that of Green; (4) clarify that the test had yielded little in the way of interpretable results; (5) indicate inconsistencies in the testing of one potentially relevant blood allotype (the "N" allotype); (6) get Schanfield to describe certain inconsistencies as "false positives;" and (7) get Schanfield to admit inconsistencies among tests taken just 30 minutes apart. At the very least, even upon reading a cold record, Bennett's cross-examination at trial posed significant questions to the accuracy and integrity of Schanfield's test and his interpretation of the test results. The mere fact that Bennett's cross-examination failed to persuade the jury of Phoenix's innocence is not enough to establish ineffective assistance.

At best, Wraxall's testimony could have added to Bennett's cross-examination by further discrediting Schanfield. Wraxall could have provided additional evidence challenging Schanfield's scientific methods and testing. He could have given the jury a reason to believe that the blood was not consistent with Green's. The jury would have heard that, because the blood was not the result of "blowback," it had not been deposited on the paper bag in the manner Schanfield suggested. However, Wraxall could not testify that the blood was not Green's; in fact, he would not testify that the blood belonged to or did not belong to any specific person. In short, he would support Bennett's cross-examination attempt, but not provide any significant new facts or evidence that might lead the jury to an alternate explanation. As such, we can see that Bennett might have made a strategic decision not to call him: to call Wraxall would reinforce the lack of an alternative explanation, would open his own expert to cross-examination, and might simply do no good. See United States v. McGill, 11 F.3d 223, 227-28 (1st Cir. 1993) (no ineffective assistance claim based on counsel's decision not to call witness, after "skillful cross-examination" had elicited much the same opinion evidence that counsel had hoped to establish through defense witness). Upon consideration of the record and Wraxall's affidavit, we agree with the district court's determination that Bennett's action was within the bounds of permissible strategic choice. Given this independent assessement, we cannot say that the same conclusion by the state court was objectively unreasonable, even if it may not have satisfied the formal requirements of a Strickland analysis. Thus we may not grant a writ of habeas corpus on this basis.

We also cannot conclude that the state court's refusal to grant a new trial based on ineffective assistance of counsel was unreasonable based on Bennett's so-called "promise" to the jury. Phoenix bases this claim on a sentence of the opening statement in which Bennett noted that "if anything, tests will show that . . . some of the blood that was on the bag . . . could not have been the blood of Ray Green." It is true that no testimony explicitly showed this "promised" fact. However, given that (i) Bennett used the "if anything" disclaimer; (ii) Wraxall would not have testified as supposedly promised, but only that Schanfield's testing was fatally flawed; and (iii) Bennett's cross-examination of Schanfield had attempted, and potentially achieved, the same result, we cannot conclude that Bennett either made a promise or that any promise he made went unfulfilled. Moreover, our cases that premise a habeas writ on an unfulfilled promise during opening argument generally require greater specificity in the promise and greater contemporaneousness between the promise and jury deliberations. Compare Anderson v. Butler, 858 F.2d 16, 19 (1st Cir. 1988) (ineffective assistance found upon explicit promise to call psychiatric witness made a day prior to jury deliberations, where voir dire had focused on jury willingness to accept such testimony), with McGill, 11 F.3d at 227-28 (decision not to call previously promised witness who had "feet of clay" not ineffective assistance). The promise here was neither "dramatic" nor was the indicated testimony "strikingly significant." Anderson, 858 F.2d at 17.

We note that in the pre-Williams world, we would occasionally remand to the district court for hearings on whether an action was within the allowable bounds of strategic choice or not. See, e.g., United States v. Rodriguez-Rodriguez, 929 F.2d 747, 752 (1st Cir. 1991). The holding of Williams dictates that we do not do so here. The district court thoroughly and carefully considered Phoenix's claim of ineffective assistance of counsel. We agree with its analysis and conclusion, and therefore affirm the denial of the writ of habeas corpus.

Affirmed.