This piece explores potential problems with the focus on unpredictability and nonlinearity within complexity theory. Whilst not completely rejecting the application of ideas of nonlinearity and unpredictability within the social sciences, I argue that greater empirical and conceptual care is needed. The arguments made are illustrated by a critical examination of cases from John Urry’s Global Complexity, including the dominance of the petroleum-fuelled car in the 20th century and the prevalence of wild-fires in Malibu. Empirically speaking, I argue that claims about particular instances of nonlinearity and unpredictability in the social world must be backed up by appropriate evidence, rather than analysts simply assuming that all social phenomena have these characteristics. Conceptually speaking, I suggest that care needs to be taken to distinguish genuinely unpredictable phenomena from those that are simply poorly understood at the present time. I also argue that predictability should be seen as a matter of degree.