Using Bayes' Rule to Update an Event's Probabilities Based on the Outcomes of Partially Similar Events

There is a widely known Bayesian solution to the problem of updating the probability of an event occurring given information on the outcome of n completely similar events. But in many, if not most, cases, we only have information on partially similar events. For example, firms must assess the probability of a new product being successful given information on past products that are only partially similar to the new product. This paper shows how the well-known Bayesian solution for completely similar events can be extended to solve the problem with partially similar past events.