This research presents a framework that supports usability experts in determining which method to use when evaluating Mobile Mixed Reality Games (MMRGs). These are games that combine the real and virtual world by means of e.g. a smartphone and require the player to change their geographical location. As some different styles of MMRGs exist, e.g. running versus cunning or multiplayer versus single player, not every method is suitable for each style. The results of the methods are benchmarked against a heuristic evaluation and it is shown that using Instant Data Analysis (IDA), Diary, interaction logs combined with audio diary and retrospective think aloud combined with IDA perform statistically comparable, but that the latter is not favorable based on qualitative merits.