The law of Duverger posits that the single-member plurality electoral system favours the two-party system. The empirical tests of Duvergerian effects often rely on the effective number of parties as the most common measure of party system fragmentation. This study, proceeding from the notion that more focused measures of two-partyness could be of greater utility for this purpose, examines the build-up of two already existing, albeit underused, indices that are intended to operationalize the concept of the two-party system and introduces a new family of party-system indices for the purpose of cross-validation. The quantitative analysis performed on simulated and real-life datasets shows that the measures of two-partyness can improve the methods of inquiry in research on Duvergerian effects, and that some of the indices introduced in this study can serve as useful supplements to the toolkit of political science.