That's not a rigorous test; it actually seems very sketchy. In the last game, the performance is better before the Denuvo removal, which doesn't make much sense. In another game, they're testing the manual run of a game, as the game (it seems ) doesn't provide an automated benchmark.