The benchmark to how well the simulator is doing predicting games is to compare it to other simulators doing the same thing (don't know of any) or Vegas Oddsmakers.  Both systems not only predict winners, but also give a win probability.  If Vegas gives the BoSox a 59% chance of winning, and I give them a 52% chance of winning, and the BoSox win, then Vegas did better than the simulator for that particuliar game, even though we both picked the BoSox to win.  If the Tribe win, then the simulator outperformed the Oddsmakers on that one particuliar game.  I have been doing this for all NL West games this season and comparing the results of the simulator with the Vegas Oddsmakers.  Beating Vegas is an extremely difficult task and the simulator was within 1.12% of doing so this year.  The simulator's margin of error is approximately 1%.  So it did pretty good for it's first year.  I will try to get that number a little lower for the 2008 season.  Alot of it depends on the input data used, not so much the code for the simulator.  I can also use the simulator for various sabermetrical experiments.  For example, I can see if a certain managerial move actually helped or hurt a team's chances of winning.  I can determine a team's win probability at any point during a game, similiar to what Fangraphs does, but there numbers don't take into account who is pitching, hitting, what part of the lineup is up etc...

vr, Xei