Friday, April 8, 2011

A lack of understanding of probabilities

Tom Verducci wrote a recent article for primarily about why Red Sox fans should be worried (the three word summary: old, expensive pitchers). He also went on to discuss the amazing success of home teams this year (55-35 record, good for a .611 winning percentage), and how it would be the biggest home field advantage ever! Unlike in other games, where home field advantage is a result of crowd noise, the structure of a baseball game will always favor a home team, even if the teams are identical.

That might be a project for another time...

Over the last ten years the home team has been good for a .545 wining percentage. Using that as the "true" winning percentage, I ran a simulation of 1,000 sets of 90 games. In 113 of them, the home teams won 55 games or more, clearly demonstrating that the hot start for home teams is well within the boundaries of random variation. Just think, we were equally likely to ending up with an article discussing the death of home field advantage, as home teams won 43 or fewer of the 90 games (0.478 winning percentage! Worst ever!) 120 times out of the 1,000 simulations.

Verducci does get it half right - acknowledging that he thinks it is a random development - but then manages to undermine that statement by then attributing the winning percentage to "cycles in the game". If I can show that you're wrong in a couple commercial breaks between innings, your idea probably shouldn't be a major feature of your article.

No comments:

Post a Comment