The output of statistical decision theory is not a degree of credence but an actionable decision: to have surgery or not, to convict or acquit. In coming down on one side or the other, we are not deciding what to believe about the state of the world. We’re committing to an action in expectation of its likely costs and benefits.
