Not actually that tricky
It looks like a Georgetown Law student, Sarah Levien Shullman, cracked the Supreme Court code. A few of years ago, while in her second year at GU Law, Shullman posited that the number of questions the justices act can serve as a predictor of which side will win the case, according to the New York Times. Looking at 10 cases, she determined that the lawyer who was asked more questions was less likely to win.
Her study made it all the way up to Chief Justice John Roberts, who did some follow-up research of his own with a somewhat larger sample of cases and found the hypothesis worked.
Now, Shullman’s theory has been corroborated by a in-depth study from the Washington University Journal of Law and Policy which looked at 2,000. The new study found Shullman’s observation about the relative number of questions asked was indeed a solid predictor of the Court’s ultimate ruling, finding furthermore that the greater the difference between the number of questions asked to the two sides, the more solidly the theory held.