Brier score
From Wikipedia, the free encyclopedia
A Brier score measures the accuracy of a set of probability assessments. Proposed by Brier (1950), it is the average deviation between predicted probabilities for a set of events and their outcomes, so a lower score represents higher accuracy.