Testable forecasts

Predictions about the future are commonly evaluated through statistical tests. As shown by recent literature, many known tests subject to adverse selection problems and cannot discriminate between forecasters who competent uninformed but predict strategically. We consider a framework where forecasters' predictions must be consistent with paradigm , set of candidate probability laws for stochastic process interest. This paper presents necessary sufficient conditions on under which it is possible informed forecasters. show that optimal take form likelihood‐ratio comparing against hypothetical Bayesian outside observer. In addition, illustrates new connection problem testing strategic classical Neyman–Pearson hypothesis testing.