We compared the magnitude of the hindsight bias in individuals and groups with the prediction that the plausibility of an outcome would affect the magnitude of the group–individual difference. We provided groups and individuals with outcomes of scientific studies, and asked them to predict the probability of those outcomes as if they did not know the given outcomes and to report their level of surprise at the outcomes. Overall, groups were more prone to hindsight bias than were individuals, but the group–individual difference was present only when the given outcomes were relatively implausible (Study 1). Moreover, this difference was not eliminated even when participants were asked to consider alternative outcomes (Study 2). Implications are discussed.