And Pearl's point is that the "real world" logic you just applied is both important and not statistical—you need to augment your analysis with these causal assu

And Pearl's point is that the "real world" logic you just applied is both important and not statistical—you need to augment your analysis with these causal assumptions in order to translate probability into meaningful causal statements.