I argue that Lewis's counterfactual theory of causation, given his treatment of counterfactuals in terms of world-comparative similarity faces insuperable problems in the form of the problem of effects and the problem of epiphenomena.