Statistically speaking? I have read somewhere that you are more likely to be killed by a meteor impact than winning the jackpot in ⟨some lottery⟩. Statistically speaking, that is nonsense. Nobody is known to have been killed by meteor impact, but many people have won the jackpot. I don’t know if the claim is true or not, but if it is, it is because some possible meteor impacts could kill millions, possibly even billions. They are just too rare to show up in a statistically meaningful way yet (fortunately). I hope the relevance of this example to the discussion of nuclear safety is obvious. I believe the technical term is “long tails” (of a probability distribution). Long tails are notoriously difficult to reason about.

This is not to say that I am absolutely opposed to nuclear power, but I don’t think the debate is easily settled one way or the other.

see my answer to rogerkirkness