Thanks. I hadn't read much on the subject before. Being able to go from that observational data and those causal assumptions to calculating precisely the expected results of an experiment you haven't carried out is amazing! I second your recommendation.