A scoring rule assigns a reward or loss to an announced probability distribution and realized outcome. In the reward convention, a proper scoring rule makes truthful reporting maximize expected reward; the loss convention reverses the comparison.
New to topics? Read the docs here!