This paper is an overview of an approach in the philosophy of science of constructing causal models of the history of science. Units of scientific knowledge, called "advances", are taken to be related by causal connections, which are modeled in computers by probability distribution functions. Advances are taken to have varying "causal strengths' through time. The approach suggests that it would be interesting to develop a causal model for scientific reasoning. A discussion of counterfactual histories of science is made, with a classification of three types of counterfactual analyses: (i) in economic and technologic history, (ii) in the history of science and mathematics, and (iii) in social history and evolutionary biology.