Page:United States Reports, Volume 542.djvu/125

86 

We reverse the judgment of the Court of Appeals and remand the case for further proceedings consistent with this opinion.

It is so ordered.  , concurring in the judgment.

I agree with much of the Court's opinion and concur in its disposition of the case. I do not, however, agree with its holding that respondent need not show prejudice by a preponderance of the evidence. Ante, at 83, n. 9.

By my count, this Court has adopted no fewer than four assertedly different standards of probability relating to the assessment of whether the outcome of trial would have been different if error had not occurred, or if omitted evidence had been included. See Chapman v. California,, 24 (1967) (adopting "harmless beyond a reasonable doubt" standard for preserving, on direct review, conviction obtained in a trial where constitutional error occurred); Brecht v. Abrahamson, , 637 (1993) (rejecting Chapman in favor of the less defendant-friendly "'substantial and injurious effect or inﬂuence'" standard of Kotteakos v. United States, , for overturning conviction on collateral review); United States v. Agurs, , 111–113 (1976) (rejecting Kotteakos for overturning conviction on the basis of violations of Brady v. Maryland, , in favor of an even less defendant-friendly standard later described in Strickland v. Washington, , 694 (1984), as a "reasonable probability"); id., at 693–694 (distinguishing the "reasonable probability" standard from the still yet less defendant-friendly "more likely than not" standard applicable to claims of newly discovered evidence). See generally Kyles v. Whitley,, 434–436 (1995). Such ineffable gradations of probability seem to me quite beyond the ability of the judicial mind (or any mind) to grasp, and thus harmful rather than helpful 