The Value of Studying Subjective Evaluations of Probability

The evaluation of probabilities, or the art of forecasting (in the judicious sense of weighing the chances of possible outcomes: not as sheer prophecy), is neither a question of taste (descriptive attitude) nor a mathematically determined question (normative attitude). All evaluations are (logically) admissible, provided only that coherence is satisfied; among these, everybody may judge one or the other more or less ‘reasonable’ (that is, in short, more or less in agreement with his own evaluation, or with the most popular one). The major aspect of coherence consists in conforming “learning from experience” to Bayes’ theorem.