The ability of a metallic cavity to sustain an isotropic diffusive regime is a necessary condition in order to be used as a reverberation chamber (RC) in narrow-band settings, for which field samples are expected to behave as complex Gaussian random variables. Yet, no reliable method has been introduced so far that accurately predicts the lowest usable frequency (LUF), starting from which field samples are more likely to pass goodness-of-fit (GoF) tests, statistical tools that can measure how well field samples reproduce Gaussian variables. This letter presents Monte Carlo simulation results that prove that the likelihood of passing GoF tests can be accurately predicted by measuring power transmission through the RC, without any a priori information about its characteristics. These results allow to identify the LUF from which a cavity is fit to generate rich multipath environments, as ideally expected for RCs, e.g., for over-the-air tests of wireless devices, and EMC tests.