There's also the fact that we usually cannot estimate the probabilities of events with that kind of precision.

I'd say that most people rarely have enough data to split probabilities into more than about five bins: ~0%, fairly unlikely, coin flip (~50%), fairly likely, and ~100%.