With information supplied by a large (n = 3393) sample of high school students from Toronto, this paper tests the assumption that three forms of leisure activity--peer, risky, and self-improving leisure--have a relatively independent impact upon patterns of offending and victimization. Although we find significant support for this proposition, we also find that traditional criminal motivations are still strongly related to criminal incidents, particularly offending behavior. The positive association between leisure and victimization includes, counter intuitively, the sort of self-improving leisure that might have been expected to reduce the risk of victimization. We discuss our findings in terms of the relationship between traditional motivational explanations of crime and newer, more situational ones.