A useful first split is whether the prototype tests the game loop or the production pipeline. For gameplay prototypes, we open the shortest playable loop, record the first moment a tester hesitates, and defer polish unless it changes the decision. What people skip is often a clearer signal than stated preference.