Abstract The origin of angiosperms is a classic macroevolutionary problem, because of their rapid rise in the Early Cretaceous fossil record, beginning about 139 Ma ago, and the conflict this creates with older crown-group ages based on molecular clock dating1. Silvestro et al.2 use a novel methodology to model past angiosperm diversity based on a Bayesian Brownian Bridge model of fossil finds assigned to extant families, concluding that a Cretaceous origin is vanishingly unlikely. However, their results strongly conflict with the known temporal distribution of angiosperm fossils, and, while we agree that statistical analysis aids interpretation of the fossil record, here we show the conclusions of Silvestro et al.2 are unsound.