A selection of the literature on the statistical description of electromagnetic fields and complex cavities is concisely reviewed. Some essential concepts, for example, the application of the central limit theorem and the maximum entropy principle, are scrutinized. Implicit assumptions, biased choices of variables and constraints, and some mathematical flaws are identified and possibly corrected. The commonly accepted probability distribution for the electromagnetic field has, therefore, to be considered as a hypothesis, which has not been contradicted by experiments. The alternative promising random coupling model is discussed as well. Ideas on future research are presented.