Let $ \pi _0,\pi _1 $ be two populations with probability distributions say $ f_0,f_1 $ (with measure $ \mu $ ). Then for testing $ H_0:f=f_0 $ against $ H_1:f=f_1 $ , we can define a test $ \phi $ with a constant $ k $ such that the expectation of the test under the null hypothesis $ H_0:f=f_0 $ denoted as $ E_0 $ is,
$ E_0\left[ \phi \left( x \right) \right] =\alpha $ (level of significance) (‎9.5)
and,
$ \phi \left( x \right) =\begin{cases} 1& when\ f_1\left( x \right) >kf_0\left( x \right)\\ 0& when\ f_1\left( x \right) <kf_0\left( x \right)\\ \end{cases} $ (‎9.6)