concurring.
I concur in the judgment of the Court. I write separately to set forth my reasoning.
This is an application for post-conviction relief in a capital case based in part on a claim of ineffective assistance of counsel at trial. The claim relates to counsel’s failure to obtain and present at trial certain confidential information concerning Martini’s character. I am satisfied that reasonably competent capital counsel would have obtained the information. The question is whether “the jury’s penalty-phase deliberations would have been affected substantially” if the information had been obtained and presented at trial. State v. Marshall, 148 N.J. 89, 250, 690 A.2d 1 (1997) (Marshall III). Under a constitution that guarantees public trials it would be impossible to conduct a partially-closed penalty-phase hearing. Even if such a closed hearing were not illegal, it would be unwise.
In Marshall III, supra, the Court adopted a different test than the Strickland/Fritz1 test for application in penalty trials to *277measure ineffective assistance of counsel claims. We departed from United States Supreme Court precedent:
[ R]ecaU that the Strickland court rejected as too severe a standard for prejudice a requirement that “counsel’s deficient conduct more likely than not altered the outcome in the case.” Strickland, supra, 466 U.S. at 693, 104 S.Ct. at 2068, 80 L.Ed.2d at 697-98. The Court instead adopted what it described as a “somewhat lower” standard, requiring a showing of “a reasonable probability that, but for counsel’s unprofessional errors, the result of the proceeding would have been different,” and observing that “[a] reasonable probability is a probability sufficient to undermine confidence in the outcome.” Id. at 694, 104 S.Ct. at 2068, 80 L.Ed.2d at 698. The Court emphasized, however, that the standards it had adopted “do not establish mechanical rules.” Id. at 696, 104 S.Ct. at 2069, 80 L.Ed.2d at 699.
[Id. at 248, 690 A.2d 1.]
In contrast, the Marshall III Court held:
Our recognition of the profound distinction between our circumscribed appellate-review function and the capital jury’s significantly less-restricted role in deciding between life and death informs our application of the prejudice prong of Strickland /Fritz to penalty-phase proceedings. That distinction demonstrates that a reviewing court strays from its traditional function if it attempts to predict the probability that a penalty-phase jury would have changed its verdict if counsel had not been deficient. In our view, an adaptation of the Strickland /Fritz prejudice test to capital-case penalty-phase proceedings that more faithfully reflects our appellate function would require courts to determine whether there is a reasonable probability that, but for counsel’s unprofessional errors, the jury’s penalty-phase deliberations would have been affected substantially. That standard is, in our view, more consistent with the Strickland Court’s admonition that a “reasonable probability that the result of the proceeding would have been different” is “a probability sufficient to undermine confidence in the outcome.” Strickland, supra 466 U.S. at 694, 104 S.Ct. at 2068, 80 L.Ed.2d at 698. The reasonable probability that ineffective assistance of counsel in the penalty phase of a capital case substantially affected the jury’s penalty-phase deliberation equates with “a probability sufficient to undermine confidence in the outcome.”
[Id at 250, 690 A.2d 1.]
I am satisfied that some of the undisclosed information “would have affected substantially” the jury’s deliberations. To explain my reasoning is difficult when the reader does not know the contents of the undisclosed information. The reader will have to take my word for it that much of the information contained in the undisclosed files might have been considered mitigating by a jury. I certainly learned things about defendant that I did not know from his original appeal.
Hence, for me, the question comes down to whether Martini’s interest in preserving the confidentiality of the undisclosed infer-*278mation outweighs society’s interest in seeing that the death penalty is imposed fairly and rationally. See State v. Koedatich, 112 N.J. 225, 548 A.2d 939 (1988) (Koedatich II), cert. denied, 488 U.S. 1017, 109 S.Ct. 813, 102 L.Ed.2d 803 (1989), and appeal after remand, 118 N.J. 513, 572 A.2d 622 (1990). In Koedatich, the Court explained that our procedures for trial and appeal are
established not only to protect the interests of the accused, but also to enable a state to enact a constitutional death penalty statute____ A defendant who prevents the presentation of mitigating evidence “withholds from the trier of fact potentially crucial information bearing on the penalty decision no less than if the defendant was himself prevented from introducing such evidence----”
[Id. at 331-32, 548 A.2d 939 (citation omitted).]
A jury verdict is not irrational when the defendant’s reasons for excluding the evidence are not irrational. That is plainly the case here. In addition, as the Court explains, the information might cut both ways, lending further rationality to a decision to withhold the evidence.
Justice STEIN joins in this opinion.
Strickland v. Washington, 466 U.S. 668, 687, 104 S.Ct. 2052, 2064, 80 L.Ed.2d 674, 693 (1984); State v. Fritz, 105 N.J. 42, 58, 519 A.2d 336 (1987).