There is another sense in which a probability estimate can be cautious: by not being too close to 0 or 1.
It isn’t caution to overestimate.
Overestimating the probability of X is just underestimating the probability of not(X).
There is another sense in which a probability estimate can be cautious: by not being too close to 0 or 1.
Overestimating the probability of X is just underestimating the probability of not(X).