I study the consequences of a random exposure to common risk for the purpose of relative performance evaluation (RPE) and find that it significantly affects the usefulness and the empirical measurement of RPE. According to my analysis, the magnitude of the exposure risk not only determines how firms aggregate measures of common risk with measures of firm performance but also the extent to which the firms can control the impact of common risk on their own performance. Simulated regressions of my theoretical model indicate that a high exposure risk can prevent the correct identification of informative performance signals and cause a biased composition of customized peer groups. A high exposure risk also increases the likelihood of a type II error in implicit RPE tests. I evaluate two empirical strategies to control for the magnitude of the exposure risk and find that they significantly reduce the likelihood of a type II error.