One is tempted to fully agree with you if the probability is associated to the knowledge or method of the person who tries to assess the probability.Predicting that it will rain tomorrow with probability 40%, and predicting that it will rain tomorrow with probability 60% are both correct, no matter whether it rains tomorrow.
On the other hand one is tempted to say that the maximum knowledge would lead to a single "correct" probability. For example if we assume that ergodicity holds (I'm probably not using the right word here). Let's say that we have infinitely many times the same system and check whether tomorrow it will rain and make the statistics. We would know with absolute certainty the probability that it will rain tomorrow, i.e. we would get a percentage. It would probably be different from 40% and 60%. People seem to believe that this percentage is the ultimate one, I think. But now it's clear to me that none is the ultimate one, at least when the percentage reflects the knowledge of the observer.