The study of elections is such a domain within political analysis, of which the conceptual-methodological basis and the set of investigative instruments, as the authors of the article maintain, are yet far from being elaborated to the full. Thus, they point to the still lacking standard criteria that would enable typology, classification and estimation of elections. In this article, the authors make an attempt to offer universally applicable criteria of estimation of elections, proceeding from their two main characteristics: predictability and competitiveness.