Nancy Cartwright relies upon an inference pattern known as inference to the best causal explanation (IBCE) to support a limited form of entity realism, according to which we are warranted in believing in entities that purportively cause observable effects. IBCE, as usually understood, is valid, even though all other forms of inference to the best explanation (IBE) are usually understood to be invalid. We argue that IBCE and IBE are in the same boat with respect to their ability to support realist conclusions. Either rule can be interpreted as valid, this is a matter of semantic convention. However, doing so deprives the rule of the empirical content the realist needs, requiring the realist to find independent warrant for a strong (theoretical or causal) premise. We then examine the proposed means of obtaining this warrant, and find them as inadequate in the case of IBCE as they are in the case of IBE.