Thursday, March 31, 2005

A thousand simulated seasons, summarized

In physics, especially in high energy particle physics, one does a lot of Monte Carlo simulations, so I was delighted to see this 1000-event projection of the 2005 baseball season using two different models. SG sent me the Diamond Mind projection system data from which I extracted the following:

The Gaussian fit has a good chi-squared per degree of freedom and has the Yankees over Boston by about 2 games on average. There's a little spike at Boston winning by 20 games, but somehow I suspect the Boss would throw money at the problem if NY got that far behind. On the other hand the Yankees are a, well, mature team - Crosby apparently will be the youngest guy on the 25-man roster at age 28.

In the same spirit, this would be an exciting season. The actual thing is just around the corner - what could be better than real data?



Post a Comment

<< Home