Long Hitting Streaks and Probability

Julius Su (of capacitor fame) e-mailed me to tell me about an opinion piece in the New York Times about the probability of there being a hitting streak of at least 56 games, which baseball players had that streak, and in which season it was most likely to occur.

When I looked at the url for the article, I noticed that the name of one member of my thesis committee (Steve Strogatz) was in it (because he coauthored the piece).

The basic premise was to simulate the history of baseball 10000 times using a Monte Carlo method (including all real end-of-season statistics from which to sample) and to determine the longest streak in each of those alternate histories, who accomplished the feat, and the year in which he did it. The result was that there was a streak at least 53 games long more than half the time. The longest streak was between 50 and 64 games 2/3 of the time. The most likely people to have the longest streak were Hugh Duffy and Willie Keeler, and the most likely year for the streak to occur was 1894. Now, here comes the eerie part: Joltin' Joe DiMaggio was the 56th most likely player to have accomplished the feat. Curiouser and curiouser...

I asked Steve by e-mail whether there is a technical article, and he just responded to confirm that there isn't any (and wasn't sure if he and his coauthor were going to write one).