Even where idiosyncratic factors such as ideology play large and consistent roles in judges' decision making, there are always cases where the patterns of judges' votes confound our expectations. In some ways, these are among the most interesting cases for scholars, raising important questions about judicial behavior and institutions. In the first part of this article, we introduce a quantitative measure of deviations from expected voting patterns intended to allow scholars to systematically study unexpected votes and the questions they raise. In the second part, we illustrate the use of this new measure by developing a preliminary model to predict departures from strict ideological voting in the U.S. Supreme Court in an effort to distinguish alternative explanations for these patterns. Although the conventional wisdom might suggest that disordered voting occurs because of cross-cutting issues, we find little evidence that case complexity is responsible for unexpected coalitions. This result suggests that some alternative explanation exists for such voting behavior; we offer several possibilities—including the influence of legal considerations—in our conclusions.