Saturday, June 26, 2010
Remembering the importance of small sample size
When doing any sort of statistical analysis, it can be really, really easy to fall prey to drawing conclusions based on a small sample, particularly when it fits with the overall story you are trying to tell about a player. Our minds try to create a pattern out of everything, even if it's more than likely (statistically speaking) just noise. Baseball statistics are no different, and in my curiosity I began looking in to it, figuring baseball statisticians that were both smarter and harder working than me had already figured it out. Lo and behold, I was right. Many statistics very quickly (swing rate for batters stabilizes within 50 PAs), while others, such as isolated power for a hitter or a pitchers walk rate, take most of a season. Here is a summary for when statistics for batters and pitchers stabilize, and for the more adventurous, the full, more detailed article is available as well, which goes in to the methods used to calculate these numbers. So, whether you are looking at a players stats trying to figure out if they'd be a good trade target in fantasy baseball, or trying to figure out why David Ortiz gets off to a slow start every April, keep these numbers in mind before drawing any conclusions.