In a number of papers, Marc Lange and others have argued that statistical theorems non-causally explain empirical matters of fact. A typical (toy-)example would be an explanation of the average outcome of throwing dice in terms of the law of large numbers. Lange calls those explanations Really Statistical Explanations (RS Explanations). In my talk, I will argue that viewing RS explanations as non-causal is mistaken and misrepresents their function. I will show that in spite of the fact that information about the explanandum’s causal history is not explicitly mentioned when stating RS explanations, such information is pragmatically conveyed due to features of the conversational context. Next, I will argue that the thus conveyed information is essential to account for why statistical theorems explain empirical matters of fact in the first place. In the final part of the talk, I sketch a novel interpretation of RS explanations as a particular kind of coarse-graining explanations.