just because a small part of the engine of max is in there it does not even come close to real world perfomace.
If you can remember the good ol Kyro2. In 3dmark2001 it lost big time to the GF2 cards (usally about 1000 pts behind stock GF2). Yet in Max Payne it usally was faster. How in the world does a card using a same engine provide such different results?
Becase the way the game is coded responds does not equal to the way the bench mark responds thus its not a good indication of how games run in even its own engine!