There tend to be a lot of things that are very low probability of success, but also to within rounding error zero cost of failure, and very very high pay off if they go well.
You should obviously doing lots of these. You're probably not doing enough of them.
12
29
217
