Subjective probability calibration: A mathematical model

Abstract A mathematical model is offered representing phenomenologically how the brain may subconsciously make subjective probability estimates based upon memories of prior experiences it deems similar, and how those estimates should be calibrated or corrected. The model differs from former theories in that it proposes a new approach based upon a new concept of estimator accuracy or expertise reflecting a subconsious partitioning of the estimator's memory trace of his prior similar experiences, and upon identification of an actual subjective probability of which a particular judgment is an estimate. It is hypothesized that the brain on the average estimates a particular subjective probability to be the midpoint of the range of values permitted such estimates. Various related statistical distributions are derived, and formulae are presented relating the accuracy, the estimated and actual subjective probabilities, and the expected actual subjective probability to be associated with a particular judgment.