As a law student, Sarah Shullman observed 10 Supreme Court arguments and tallied the number and tenor of justices' questions. In a new study, she reports her system to be a surprisingly simple and accurate way of predicting high court outcomes. In all the cases, the justices in aggregate asked more questions, and more hostile questions, of the party that ultimately lost. Though Shullman's sample was small, her methodology has already been tested since the study -- and found accurate in 86 percent of cases.
Details here from Tony Mauro of The American Lawyer.