Monday, October 02, 2006

2006 Playoff Sims

I will be using the baseball game simulator program that I wrote to simulate playoff game outcomes. I will run a sample size of 25 (now up to 1,000 games for the World Series) games for each game. I will run the first two games of each series, then I will simulate each remaining game the day of or the night before that game is to be played. How does the software work? A text file of hitting and pitching stats is used to determine the likeliehood of each at bat outcome. Hitting outcomes are determined on an at-bat by at-bat basis and not a pitch by pitch basis. The software uses a random number generator to determine if the batter gets on base. If his OBP is .400 and the random number is less than .400 then he gets on base. Once he is on base then this same random number is used to determine what kind of hit it is, or if it is a walk based on single, double, triple, HR and walk rates. A constant vs LHP vs RHP split is used to account for lefty pitcher vs lefty hitter type of situations. Good pitchers are given an advantage over bad pitchers in that they begin the game with a number that is subtracted off of the hitters OBP/batting average etc... This number is decreased for each batter the pitcher faces, thus simulating a pitcher getting tired the more pitches he throws. Artificial intelligence algorithms determine when to steal, sac bunt, pinch hit, bring in a new pitcher or throw a beach ball onto the field. Lineups and pitching matchups are selectable and there is an option for using a DH instead of having the pitcher bat, for American League simulated games. This is all I can think of as far as a description goes for now. If you have any questions, please send them to: xeifrank@yahoo.com vr, Xei

No comments: