with whom The Chief Justice, Justice Thomas, and Justice Alito join, dissenting.
This is a companion case to Lafler v. Cooper, post, p. 156. The principal difference between the cases is that the fairness of the defendant’s conviction in Lafler was established by a full trial and jury verdict, whereas Frye’s conviction *152here was established by his own admission of guilt, received by the court after the usual colloquy that ensured it was voluntary and truthful. In Lafler all that could be said (and as I discuss there it was quite enough) is that the fairness of the conviction was clear, though a unanimous jury finding beyond a reasonable doubt can sometimes be wrong. Here it can be said not only that the process was fair, but that the defendant acknowledged the correctness of his conviction. Thus, as far as the reasons for my dissent are concerned, this is an a fortiori case. I will not repeat here the constitutional points that I discuss at length in Lafler, but I will briefly apply those points to the facts here and comment upon a few statements in the Court’s analysis.
* * *
Galin Frye’s attorney failed to inform him about a plea offer, and Frye ultimately pleaded guilty without the benefit of a deal. Counsel’s mistake did not deprive Frye of any substantive or procedural right; only of the opportunity to accept a plea bargain to which he had no entitlement in the first place. So little entitlement that, had he known of and accepted the bargain, the prosecution would have been able to withdraw it right up to the point that his guilty plea pursuant to the bargain was accepted. See 311 S. W. 3d 350, 359, and n. 4 (Mo. App. 2010).
The Court acknowledges, moreover, that Frye’s conviction was untainted by attorney error: “[T]he guilty plea that was accepted, and the plea proceedings concerning it in court, were all based on accurate advice and information from counsel.” Ante, at 141. Given the “ultimate focus” of our ineffective-assistance cases on “the fundamental fairness of the proceeding whose result is being challenged,” Strickland v. Washington, 466 U. S. 668, 696 (1984), that should be the end of the matter. Instead, here, as in Lafler, the Court mechanically applies an outcome-based test for prejudice, and mistakes the possibility of a different result for constitu*153tional injustice. As I explain in Lafler, post, p. 175 (dissenting opinion), that approach is contrary to our precedents on the right to effective counsel, and for good reason.
The Court announces its holding that “as a general rule, defense counsel has the duty to communicate formal offers from the prosecution” as though that resolves a disputed point; in reality, however, neither the State nor the Solicitor General argued that counsel’s performance here was adequate. Ante, at 145. The only issue was whether the inadequacy deprived Frye of his constitutional right to a fair trial. In other cases, however, it will not be so clear that counsel’s plea-bargaining skills, which must now meet a constitutional minimum, are adequate. “[H]ow to define the duty and responsibilities of defense counsel in the plea bargain process,” the Court acknowledges, “is a difficult question,” since “[bargaining is, by its nature, defined to a substantial degree by personal style.” Ante, at 144-145. Indeed. What if an attorney’s “personal style” is to establish a reputation as a hard bargainer by, for example, advising clients to proceed to trial rather than accept anything but the most favorable plea offers? It seems inconceivable that a lawyer could compromise his client’s constitutional rights so that he can secure better deals for other clients in the future; does a hard-bargaining “personal style” now violate the Sixth Amendment? The Court ignores such difficulties, however, since “[t]his case presents neither the necessity nor the occasion to define the duties of defense counsel in ’those respects.” Ante, at 145. Perhaps not. But it does present the necessity of confronting the serious difficulties that will be created by constitutionalization of the plea-bargaining process. It will not do simply to announce that they will be solved in the sweet by-and-by.
While the inadequacy of counsel’s performance in this ease is clear enough, whether it was prejudicial (in the sense that the Court’s new version of Strickland requires) is not. The Court’s description of how that question is to be answered *154on remand is alone enough to show how unwise it is to consti-tutionalize the plea-bargaining process. Prejudice is to be determined, the Court tells us, by a process of retrospective crystal-ball gazing posing as legal analysis. First of all, of course, we must estimate whether the defendant would have accepted the earlier plea bargain. Here that seems an easy question, but as the Court acknowledges, ante, at 150, it will not always be. Next, since Missouri, like other States, permits accepted plea offers to be withdrawn by the prosecution (a reality which alone should suffice, one would think, to demonstrate that Frye had no entitlement to the plea bargain), we must estimate whether the prosecution would have withdrawn the plea offer. And finally, we must estimate whether the trial court would have approved the plea agreement. These last two estimations may seem easy in the present case, since Frye committed a new infraction before the hearing at which the agreement would have been presented; but they assuredly will not be easy in the mine run of cases.
The Court says “[i]t can be assumed that in most jurisdictions prosecutors and judges are familiar with the boundaries of acceptable plea bargains and sentences.” Ante, at 149. Assuredly it can, just as it can be assumed that the sun rises in the west; but I know of no basis for the assumption. Virtually no cases deal with the standards for a prosecutor’s withdrawal from a plea agreement beyond stating the general rule that a prosecutor may withdraw any time prior to, but not after, the entry of a guilty plea or other action constituting detrimental reliance on the defendant’s part. See, e. g., United States v. Kuchinski, 469 F. 3d 853, 857-858 (CA9 2006). And cases addressing trial courts’ authority to accept or reject plea agreements almost universally observe that a trial court enjoys broad discretion in this regard. See, e.g., State v. Banks, 135 S. W. 3d 497, 500 (Mo. App. 2004) (trial court abuses its discretion in rejecting a plea only if the decision “is so arbitrary and unreasonable that *155it shocks the sense of justice and indicates a lack of careful consideration” (internal quotation marks omitted)). Of course after today’s opinions there will be cases galore, so the Court’s assumption would better be cast as an optimistic prediction of the certainty that will emerge, many years hence, from our newly created constitutional field of plea-bargaining law. Whatever the “boundaries” ultimately devised (if that were possible), a vast amount of discretion will still remain, and it is extraordinary to make a defendant’s constitutional rights depend upon a series of retrospective mind-readings as to how that discretion, in prosecutors and trial judges, would have been exercised.
The plea-bargaining process is a subject worthy of regulation, since it is the means by which most criminal convictions are obtained. It happens not to be, however, a subject covered by the Sixth Amendment, which is concerned not with the fairness of bargaining but with the fairness of conviction. “The Constitution ... is not an all-purpose tool for judicial construction of a perfect world; and when we ignore its text in order to make it that, we often find ourselves swinging a sledge where a tack hammer is heeded.” Padilla v. Kentucky, 559 U. S. 356, 388 (2010) (Scalia, J., dissenting). In this case and its companion, the Court’s sledge may require the reversal of perfectly valid, eminently just, convictions. A legislature could solve the problems presented by these cases in a much more precise and efficient manner. It might begin, for example, by penalizing the attorneys who made such grievous errors. That type of sub-constitutional remedy is not available to the Court, which is limited to penalizing (almost) everyone else by reversing valid convictions or sentences. Because that result is inconsistent with the Sixth Amendment and decades of our precedent, I respectfully dissent.