This is a defence of the statistical interpretation of fitness. I distinguish three interpretations of the evolutionary explanations in which fitness figures: the Two Factor Model, The Single Factor Model and the Statistical Interpretation. They differ in their degrees of causal commitment. The first two are committed to fitness distribution being a cause of population change. The last maintains that fitness distribution correlates with population change, but does not cause it. The defence of the statistical interpretation relies upon a distinctive feature of fitness--that fitness is a function of the mean and variance of reproductive output and population size--and a distinctive feature of causation--that it conforms to the Sure Thing Principle.