This article presents a longitudinal test of three proposed functions of overconfidence. In a sample of 894 high school boys surveyed across two school years, we examined whether overconfidence in sporting ability and intelligence predicts improved mental health, motivation, and popularity. Both sporting and intelligence overconfidence showed positive cross-sectional associations with mental health outcomes, but there was little evidence that overconfidence predicted improved mental health over time. Some evidence emerged that overconfidence in sporting ability, but not intellectual ability, predicted increased effort, but neither type of overconfidence predicted improvements in ability over time. Finally, sporting but not intellectual overconfidence predicted increased popularity over time. These results suggest that overconfidence is associated with increased social success over time in at least some domains, and contradict the oft-cited possibility that overconfidence leads to increasingly deleterious outcomes over time.