In a Bayesian framework, can a subjective probability be considered as a proportion of certainty, or does a proportion only make sense if we are, say, counting the number of "successes" out of the number of "trials", as in a frequentist framework? Thank you.
- Anybody can ask a question
- Anybody can answer
- The best answers are voted up and rise to the top
It can. As defined and used, "subjective probability" quantifies in a relative but cardinal way "how much" certain we are, and this number is comparable to other such numbers and the magnitude of their difference is meaningful: thus if I say "I believe outcome A has 0.4 probability and outcome B has 0.2 probability", it is meaningful to also say "outcome A is twice as probable than outcome B". Perhaps this is the strongest (philosophical and anthropological) doubt related to subjective probability: do people really calculate and "experience" such cardinal probabilities, or do they just use "ordinal probability" (which is fully developed theoretically, but it only orders events in terms of their "subjective certainty", without quantifying their distance)... but I am getting carried away, this final part was not in the question.