Distinguishing two probability distributions with a finite number of data points

Allows the forecasting of how easy it will be to distinguish whether data are drawn from a particular probability distribution or from a different one (both distributions must be known). Determines which range of values is best suited for the test, which assumes only one independent variable.