Sunday, March 30, 2008

Baseball, Monte Carlo Style

How fun is this?: a Monte Carlo study of the history of baseball done by two Cornell applied mathematicans, using historical statistics, player-by-player.
Then again, this is the first time anyone has tried this?

This is exactly what we do in particle physics when we're trying to see how well a measurement will work -- we "try" (i.e. simulate) the measurement many times (i.e. often way more times than we actually do the measurement), and see how often we hit or miss (which tells us the efficiency of the procedure). We also like to calculate the probability of something looking like our measurement, just by a chance fluctuation of our backgrounds (which tells the the "purity" of our data). Maybe Bob Adair should have thrown some physicists at understanding the relative importance of various landmarks baseball's history, instead of just looking at the equipment?...

No comments: