The classical approach to probability is based on the relative frequency: number of favorable outcomes/number of all possible outcomes. It works if and only if the outcomes arerepeatable. However, some event is not repeatable (e.g. origin of life, big bang...etc.). As a remedy, the Bayesian school conceptualizes probability as adegree of subjective belief.

Bayesian starts with a prior probability.The estimate of the prior is based on background information or the researcher'sexperience, and thus it is verysubjective. For example, in the book entitled "The probability of God" Steve Unwin used the Bayesian approach to compute how likely God exists. He used 0.5 as the prior probability. But Richard Dawkins was opposed to Unwin's computation. Dawkins argued that the prior should be lower than 0.5. Tufts University physicist Larry Ford has examined Unwin's calculation and has made his own estimate using the same formula. Ford's result is a much smaller probability because he assigned very low scores to the evidence of God's existence. How could we address the problem of different estimations by different people in Bayesian probability?