This essay reviews recent work regarding pairwise tests of equal forecast accuracy between nested and non-nested models. While some technical details are given, special emphasis is placed on the practical implementation of the tests.