Fishery collapse has been defined as a fishery with annual landings less than 10% of the historic maximum observed catch. However, this 10% rule is not grounded in bioeconomic theory despite being widely used in empirical economic studies of fisheries. We assess the 10% rule by simulating fisheries under pure open access, open access with cost changes, open access with critical depensation, optimal management (both deterministic and stochastic cases), and rebuilding plans. We show that the 10% rule generates false negatives and false positives, and that the prevalence of these problems varies under different institutional configurations, economic incentives, and biological conditions. We urge researchers to abandon this outcome measure for comparative empirical tests and encourage more research on collapse that attends to human agency and institutions.