We provide a general framework for quantifying the effects of genotypic selection prediction accuracy and varying cost ratios of phenotyping to genotyping on the economic performance of genotypic selection relative to traditional phenotypic selection. Economic performance is measured using normalized average cost per unit of genetic gain. The economic performance of genotypic selection declines with (1) trait heritability, (2) relative cost of genotyping, and (3) the number of QTL (genes) affecting the trait. The benefits of increasing the training population size tend to be higher for traits with low heritability and traits affected by a larger number of QTL. The economically optimal sizes of the training population tend to be larger than the sizes that are typically used in current plant breeding programs.