Sure, but you only know that after the fact. Basically, talking about `risk' after the fact is a murky business. (And can invite a lot of philosophical debate about frequentist vs Bayesian interpretations of probability. And determinism.)