Fuzzy rationality in the elicitation of subjective probabilities

The paper focuses on the application of a modification of the classical bisection method, called triple bisection, for elicitation of subjective probabilities. The following is formulated: a theoretical algorithm for probability elicitation by an ideal observer (with infinite discriminating abilities), based on bisection; a practical algorithm for probability elicitation by a real observer (without infinite discriminating abilities), based on the triple bisection. The results are discussed from the point of view of fuzzy logic. The arguments support the idea that people are imperfect in their estimates, which is why the combination of systematic algorithms and fuzzy logic provides results that Tit best to the observer's opinion, and the precision reflects the observer's abilities.