The United States Supreme Court is one of the pinnacles of human wisdom and judgment. But when it comes to predicting the High Court’s decisions, can a machine do better than the wisdom of crowds?

The interesting new participant in this term’s FantasySCOTUS tournament is {Marshall}+ – a sophisticated predictive algorithm that will match “his” predictive skills against the players of FantasySCOTUS. Developed by law professors Josh Blackman, Daniel Martin Katz and Michael J. Bommarito II, {Marshall}+ analyzes almost 100 variables to construct very precise decision weights. Each Justice’s vote is then predicted by creating thousands of decision trees using these weights.

As Bommarito explains, “Have you seen some of those big flow charts on the Internet where you’re supposed to answer ‘Yes,’ or ‘No,’ to a bunch of questions and there’s a sassy response at the bottom? The model is composed of similar decision trees, only we employ thousands of them.”

In applying the model against all Supreme Court decisions from 1953 to the present, {Marshall}+ correctly identified 70.9 percent of the votes of individual justices across 7,700 cases and more than 68,000 justice votes.

For the record, that is slightly better – but only just so – than the 68.4 percent accuracy from FantasySCOTUS participants last year, promising a neck-and-neck competition.

Follow how {Marshall}+ does with his predictions during the Court’s term here. Will he outperform your league? And watch for our upcoming “interview” with {Marshall}+.

For previous articles on FantasySCOTUS, click here.