[lupusreal]

If a ship's captain is asleep, as he logically must be some of the time, and the officers on watch crash the ship into a reef, that captain's career is over even though by normal civilian standards it might be surmised that he did nothing wrong. Even if he's got the receipts to prove he trained his men properly according to the book and they passed all their tests. Civilian intuition about blame doesn't really apply.

[m4rtink]

Oh, so you need people who are lucky ?

And no need to worry about loosing talented people due to circumstances outside of their control (and preventing other from even consideringto join) - they attract bad luck and that is more dangerous in the long term!

[vladms]

There are some situations where you try to reduce type II errors (false negative) even if you increase type I errors (false positive).

The hypothesis is "the captain was to blame". If this is true but you reject it (type I error) you will probably loose another ship. If this is false but you reject it (type II error) you need to find another qualified captain (which, depending on the field might or might not be a problem).

My guess is that there are more pilots/captains trained that planes/ships.

[lupusreal]

There are all kinds of subtle factors which may allude quantification, but which nonetheless matter because they contribute to outcomes which matter. When things really matter you can't afford to ignore these factors just because you can't pin them down. So yes, you favor those who are "lucky".

[potato3732842]

I hate this sort of "everything that requires subjective judgement is luck" trope that is pervasive on HN and similar parts of the internet.

The line between luck and skill is blurry. The captain's job is to avoid the preconditions for failure. Maybe that means a different route, maybe that means scheduling such that you're awake for the sketchy stuff. Yes, all of this is subjective and has tradeoffs, that's why it's skilled work people take years to develop the skills for.