All of us make judgments of probability, and we rely on them for our decision-making. This paper argues that such judgments are trustworthy only to the extent that one has good reasons to think that they are produced by maximally inclusive, well calibrated cognitive processes. A cognitive process is maximally inclusive when it takes into account all the evidence which one regards as relevant, and it is well calibrated when anything it would assign, say, an 80% probability to would be (...) true 80% of the time. We further have good reasons to think these judgments are trustworthy when, inter alia, they are produced by processes that have good track records of calibration. Call this inclusive calibrationism—or just “calibrationism” for short. In arguing for calibrationism, I also appeal to various empirical results, including research into probabilistic reasoning funded by the US intelligence community. Together, these ideas and results have implications for some important philosophical problems: the problem of the priors, the problem of unique events and the use of intuition in probabilistic reasoning. These theses and results also imply that our judgments are often less trustworthy than we might hope for potentially many domains, including law, medicine and others—barring good track records, that is. (shrink)