I've written a baseball simulator. This is not a tool like a game where you play a game of baseball on the computer. It is a baseball season simulator. Required data includes
n = number of teams
g = number of games per season
s = skill levels for each of the teams
N = number of skill plays per game
The simulator works as a Bernoulli trial between each of the teams for the required number of games per team. Within each game, there is an independent Bernoulli trial of N plays where the skill makes a difference.
Conclusions: With the model we can very accurately simulate a baseball season extremely closely with an actual season - by chi-squared. We can also show that just the slightest improvement of a team can dramatically affect its season's outcome.
Details to follow.
No comments:
Post a Comment