In Lafler v. Cooper (analysis here), the Supreme Court (I would argue) recognized a new constitutional right to effective assistance of counsel during the plea-bargaining phase when a plea is rejected (or something like that; the 9th Circuit will flesh that out).

One interesting aspect of this new-fangled right is that attorneys must be able to predict whether it is better to go to trial, or better to accept a plea bargain. The Court noted:

Respondent has satisfied Strickland’s two-part test.Regarding performance, perhaps it could be accepted thatit is unclear whether respondent’s counsel believed respondent could not be convicted for assault with intent tomurder as a matter of law because the shots hit Mundybelow the waist, or whether he simply thought this would be a persuasive argument to make to the jury to show lackof specific intent. And, as the Court of Appeals for the Sixth Circuit suggested, an erroneous strategic prediction about the outcome of a trial is not necessarily deficient performance. Here, however, the fact of deficient performance has been conceded by all parties. The case comes to us on that assumption, so there is no need to address this question.

“Erroneous strategic prediction is not necessarily deficient” can be read (hello 9th Circuit) as a really “erroneous strategic prediction can be deficient.”

So what can enable an attorney to make a non-erroneous, or preferably a good, prediction of what would happen at trial?

Say hello to the value of prediction markets. In my recently published article on FantasySCOTUS, I addressed just this scenario, in the case that the efficiency of a plea bargain depends on the outcome of a currently-pending Supreme Court case.

Imagine that during an interrogation a suspect was read her Miranda rights, did not affirmatively invoke her right to remain silent, and subsequently made an incriminating statement. Assume that Berghuis v. Thompkins, which presented just this issue, has been argued before the Court, but not yet decided.195 The prosecutor offers the defendant a favorable plea bargain that is only on the table for a limited duration; if not accepted, the prosecutor will take the case to trial. If the defendant accepts the plea agreement, she waives all appeal rights.

The defense attorney is faced with a choice. If her client accepts the plea bargain, and the Supreme Court subsequently finds that this interrogation did not result in a violation of Miranda, her client will have secured a short sentence, less than what she likely would have received at trial. Alternatively, if her client accepts the plea bargain, and the Supreme Court finds this interrogation did result in a violation of Miranda, her client cannot challenge the confession on appeal, and she is stuck in jail; had she gone to trial, the court would have suppressed the evidence, and she would have likely been acquitted without the confession.

If FantasySCOTUS shows that the Court will find a violation of Miranda rights in Berghuis v. Thompkins, perhaps the attorney should roll the dice and go to trial, hoping the judge will ultimately suppress the evidence, or perhaps her client could challenge it on appeal. If FantasySCOTUS shows that the Court will not find a violation of Miranda (the actual outcome of this 5–4 decision), perhaps the attorney should accept the favorable plea bargain, and not risk it. These are real decisions defense attorneys have to make. With the FantasySCOTUS of the future, this decision could be aided by informed predictions and their accompanying statistical measures of certainty.

This predictive power will, ultimately, help a lawyer augur the outcome of any trial, irrespective of the Supreme Court.

Perhaps one day checking the predicted outcome of a case, should it go to trial, will be part of this new-fangled constitutional right to effective representation during plea-bargaining!

Just think. Harlan propping up the Constitution! I can’t wait.