Chance, Vol. 23, No. 4, pp. 52-57, 2010
6 Pages Posted: 9 Sep 2010 Last revised: 31 Jan 2011
Date Written: December 8, 2010
There have been more hitting streaks in Major League Baseball than we would expect if hitting streaks were just the result of random chance (ie, if the independence assumption were true). All batting lines of MLB hitters from 1957-2006 were randomly permuted 10,000 times and the number of hitting streaks of each length from 2 to 100 was measured. The average count of each length streak was then compared to the corresponding total from real life, when the games were in chronological order.
The number of streaks in real life was significantly higher than the average over the random permutations, and the spread between real life and the permutations increased as streak length increased.
Non-starts (such as pinch-hitting appearances) were removed since these may be unduly reducing the number of streaks in the permutations; the number of streaks in the permutations more closely matched real life but longer streaks (30 games) were still significantly lower than real-life totals.
Possible explanations are given for why more streaks have appeared in real-life than we would expect, including weather, stadium/ballpark, opposing pitching, changes in strategy, and possibly the hot hand idea.
Keywords: Baseball, Hitting Streak, Random, Permutation, Persistence, Skills, Skill, Streak, Luck, Hot Hand, Sports, Streakiness
JEL Classification: C15, L83
Suggested Citation: Suggested Citation
McCotter, Trent, Hitting Streaks Don't Obey Your Rules: Evidence that Hitting Streaks Aren't Just By-Products of Random Variations (December 8, 2010). Chance, Vol. 23, No. 4, pp. 52-57, 2010. Available at SSRN: https://ssrn.com/abstract=1674152