I meant to do this yesterday, when it would have been more believable, but here I am on a regular old day writing about baseball. In case you missed it, non-linear dynamics guru Steven Strogatz and graduate student Samuel Arbesman had a paper in the New York Times on using a game simulator to simulate every at bat for all of baseball 10,000 times. They then used this artificial data set to look for trends, including a really cool distribution of hitting streak records (the real record of 56 games is held by Joe DiMaggio). Apparently DiMaggio’s record — according to Wikipedia, “the top American sports feat of all time” — was actually to be expected. In the 10,000 samples the median record was 53 games. Additionally, judging from the way most of the streaks seemed to occur in the 1890’s , the sport must have been even easier in the gay nineties than the steroid nineties. Maybe they were using the performance-enhancing drugs of the time — cocaine and heroin? Actually, I might watch a baseball game if everyone was coked up. Seems like it would have to be exciting even to a philistine like me. Anyhow, it’s a neat study and very similar to some of the exposure approaches that they use at EPA. They simulate the day-to-day life for a bunch of individuals (ala The Sims) in order to assess how well we understand the origin of chemicals found in blood.
Incidentally, former Duke physics (or was it math?) undergrad Lauren is now a graduate student with Strogatz. I ran into her last year at a talk on the Gillespie algorithm by Dan Gillespie himself. As he put it: The Gillespie algorithm was created by Dan… There are many copies, and we have a plan. Of course, his plan is a more modest one involiving simulating chemical reactions.