concurring in the denial of rehearing in banc.
The dissenters have identified possible challenges posed by New York’s constitutional standard for ineffective assistance of counsel claims. As they note, the New York standard could leave room for New York courts to find a lawyer effective by focusing on the “fairness of the process as a whole,” People v. Benevento, 91 N.Y.2d 708, 714, 674 N.Y.S.2d 629, 697 N.E.2d 584 (1998), rather than on whether “there is a reasonable probability that ... the result of the proceeding would have been different” absent defense counsel’s mistakes, Strickland v. Washington, 466 U.S. 668, 694, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). See Henry v. Poole, 409 F.3d 48, 70-72 (2d Cir.2005) (“paus[ing] to question whether the New York standard is not contrary to Strickland ”).
As both Chief Judge Jacobs’ dissent and Judge Wesley’s concurrence observe, however, such difficulties can be avoided by separate consideration of counsel’s performance under the federal standard when a federal challenge is presented in the New York courts. For the reasons set forth in the panel’s decision, see Rosario v. Ercole, 601 F.3d 118, 127 (2d Cir.2010), I am satisfied that the trial court here engaged in such an inquiry, albeit “not delivered in Strickland terminology,” id. (quoting Rosario v. Ercole, 582 F.Supp.2d 541, 553 (S.D.N.Y.2008)). Accordingly, this case does not require us to review New York’s standard. Thus, I concur in the decision of the Court to deny rehearing in banc.
DENNIS JACOBS, Chief Judge, joined by ROSEMARY S. POOLER, GERARD E. LYNCH, and DENNY CHIN, Circuit Judges, dissenting from the denial of rehearing in banc.I agree with the panel majority that the New York standard for ineffective assistance of counsel is more lenient to defendants generally, lacking as it does a “but for” prejudice requirement. See People v. Turner, 5 N.Y.3d 476, 480, 806 N.Y.S.2d 154, 840 N.E.2d 123 (2005). But it is nevertheless contrary to the standard set forth in Strickland v. Washington, 466 U.S. 668, 104 S.Ct. 2052, 80 L.Ed.2d 674 (1984). I respectfully dissent from the order denying in banc review because this defect likely will give rise to more cases that will bedevil the district courts, which are left to sort out case-by-case a problem that is systemic.1
*686I
Under federal law, a lawyer is ineffective when conduct that falls “below an objective standard of reasonableness,” Strickland, 466 U.S. at 688, 104 S.Ct. 2052, creates “a reasonable probability that ... the result of the proceeding would have been different,” id. at 694, 104 S.Ct. 2052. “[U]nder New York law the focus of the inquiry is ultimately whether the error affected the ‘fairness of the process as a whole.’ ” Rosario v. Ercole, 601 F.3d 118, 124 (2d Cir.2010) (quoting People v. Benevento, 91 N.Y.2d 708, 714, 674 N.Y.S.2d 629, 697 N.E.2d 584 (1998)). The test articulated by the New York Court of Appeals thus allows a lawyer whose overall performance is adequate to be deemed constitutionally effective notwithstanding an isolated lapse that calls the result into question—the very scenario that triggers relief under Strickland. The New York standard is fairly unambiguous:
Two of our decisions have rejected ineffective assistance claims despite significant mistakes by defense counsel (People v. Hobot, 84 N.Y.2d 1021, 622 N.Y.S.2d 675, 646 N.E.2d 1102 (1995); People v. Flores, 84 N.Y.2d 184, 615 N.Y.S.2d 662, 639 N.E.2d 19 (1994)). Those cases hold, and we reaffirm today, that such errors as overlooking a useful piece of evidence {Hobot), or failing to take maximum advantage of a Rosario violation (.Flores), do not in themselves render counsel constitutionally ineffective where his or her overall performance is adequate. But neither Hobot nor Flores involved the failure to raise a defense as clear-cut and completely dis-positive as a statute of limitations. Such a failure, in the absence of a reasonable explanation for it, is hard to reconcile with a defendant’s constitutional right to the effective assistance of counsel.
Turner, 5 N.Y.3d at 480-81, 806 N.Y.S.2d 154, 840 N.E.2d 123 (emphases added). Thus the New York test averages out the lawyer’s performance while Strickland focuses on any serious error and its consequences.
In the passage quoted above, the Turner court relies on Flores. That is dubious precedent. In Flores, a case involving a single serious error, the New York Court of Appeals relied on the “totality of representation” to decide that defense counsel’s waiver of a Rosario claim did not constitute ineffectiveness. People v. Flores, 84 N.Y.2d 184, 187, 615 N.Y.S.2d 662, 639 N.E.2d 19 (1994). Years later, this Court granted habeas relief, finding “at least a reasonable probability ... that had that Rosario claim been pressed, Flores would have been granted a new trial by the trial court or on appeal.” Flores v. Demskie, 215 F.3d 293, 305 (2d Cir.2000).
Because the New York standard allows the gravity of individual errors to be discounted indulgently by a broader view of counsel’s overall performance, it is contrary to Strickland.
II
The present case illustrates the constitutional defect in the New York standard. Rosario’s pre-trial and trial counsel admitted an uncommonly bad mistake in believing that the state court had denied an application for funds to send an investigator to find and interview about a dozen alibi witnesses who would swear that the defendant was in Florida when the charged crime was committed in New York. This failure was not a misfire of strategy or tactics; it was conceded error. In denying Rosario’s collateral challenge, the Supreme Court of the State of New York acknowledged as much, but then shifted the focus:
*687The best and most reasonable explanation, then, is that there was a misunderstanding or mistake which persisted through the case and which the parties simply cannot explain. But it was not deliberate. And that does not alter the fact that both attorneys represented defendant skillfully, and with integrity and in accordance with the standards of “meaningful representation” defined by our appellate courts.
It is this shift—from the specific mistake to the broader performance—that concerns me and should concern the entire Court.
Ill
I recognize that some colleagues may not consider this case an ideal vehicle for deciding the issue, in view of the state court’s alternative ruling that “an alibi defense was presented through the two witnesses who had the best reason for remembering why” Rosario was elsewhere when the crime was committed, and that the prospective additional alibi witnesses “were, for the most part, questionable and certainly not as persuasive as the two witnesses who did testify.”
The panel majority seizes on this alternative ground to assert that the state court “considered the prejudicial effect of the errors, and concluded that the outcome of the trial would not have been different but for those errors.” Rosario, 601 F.3d at 128.
The state court’s findings of fact may bear on whether the state court unreasonably applied the correct federal standard, but they do not obviate the need to start with the correct standard; a finding on a mixed question of law and fact (such as prejudice) is suspect (at least) if it is guided by a defective understanding of the law. Moreover, although the state court conducted a hearing that included testimony from seven prospective alibi witnesses, I am unimpressed by the finding that they were “for the most part, questionable,” and that the two who testified at trial were the best of the lot. First, if a witness is without flaw, I tend to suspect perjury; second, corroboration matters. As the panel dissent forcefully explains, Rosario was seriously prejudiced by the absence of more alibi witnesses. Rosario, 601 F.3d at 131-37, 140-42 (Straub, J., dissenting).
IV
The full Court took an in banc poll in this case and decided not to revisit the panel’s ruling. But this should not be construed as an imprimatur.2
I acknowledge that in most instances the state standard is more solicitous of the Sixth Amendment right to counsel than ours, and I respect the measures taken by the New York courts to administer cases in a way that seeks to accommodate a federal standard that is not congruent. That said, members of this Court entertain serious disquiet that in the courts of New York the gravity of individual mistakes may be submerged in an overall assessment of effectiveness, in a way that violates the federal Constitution.
Unneeded conflict can be avoided by separate consideration of counsel’s performance under the Strickland standard in *688the New York courts when a federal challenge is presented. No doubt, there are other ways to the same end. But without some further vigilance in the state courts, the issue will be presented to us one day in a case in which fact-findings do not blur focus on the constitutional question, and an in banc panel of this Court may be convened to deal with it.
. Senior Circuit Judge Chester J. Straub, the author of the panel’s minority opinion concurring in part and dissenting in part, was not authorized to participate in the in banc poll, but has endorsed the views expressed in this opinion.
. At different times, this Court has been of different minds on the question. Some opinions have said (albeit in dicta or in following binding precedent) that the New York test is not contrary to Strickland. See Eze v. Senkowski, 321 F.3d 110, 124 (2d Cir.2003); Loliscio v. Goord, 263 F.3d 178, 192-93 (2d Cir.2001); Lindstadt v. Keane, 239 F.3d 191, 198 (2d Cir.2001). A later panel voiced doubt. See Henry v. Poole, 409 F.3d 48, 70-71 (2d Cir.2005) (pausing “to question whether the New York standard is not contrary to Strickland," but granting habeas relief on the unreasonable application ground).