Our current understanding of the operation of sexual selection is predicated on a sex difference in parental investment, which favours one sex becoming limiting and choosy over mates, the other competitive and nonchoosy. This difference is reflected in the operational sex ratio (OSR), the ratio of sexually receptive males to females, considered to be of fundamental importance in predicting the direction of sexual selection. Difficulties in measuring OSR directly have led to the use of the potential reproductive rates (PRR) as a measure of the level of investment in offspring of males and females. Several recent studies have emphasized that other factors, such as variation in mate quality and sex differences in mortality patterns, also influence the direction of sexual selection. However, as yet there has been no attempt to form a comprehensive theory of sex roles. Here we show that neither OSR nor PRR is the most fundamentally important determinant of sex roles, and that they are not interchangeable. Instead, the cost of a single breeding attempt has a strong direct effect on competition and choosiness as well as consistent relationships to both OSR and PRR. Our life history based approach to mate choice also yields simple, testable predictions on lack of choice in either sex and on mutual mate choice.