[maest]

You know, I've never found a rigorous definition of what "causality" actually is.

Like, I know what "correlation" is: slap a regression on A and B and see what comes out (after considering heteroscedasticity and friends). But, for causation, how does one find it?

Is causation even a well defined concept? In most disaster analysis situations you see that failure wasn't caused by a single factor but it came as a result of a combination of different factors which, on their own, are benign.

If someone is crossing the road while checking thir phone (so not paying attention) and a drunk driver hits them with their car, what "caused" the accident?

Do phones "causes" accidents? Does drunk driving "cause" accidents?