Bayesball: Update

For those of you who enjoyed my brief review on the Bayesball paper, you may also enjoy the following two articles that came out in the last couple of days:

Statistical evidence consistent with performance-enhancing drugs in professional baseball: The authors compare players statistics with the Mitchell Report, an anonymous blood sample study that concluded 5% of MLB took performance drugs in 2003. This study examines “approximately 10,000 players who ended their careers between the years 1920 and 2000.” They separate 1920-1979 performance with 1980-2000, noting a significant increase in home-run likelihood, but see no difference in hits, runs batted in, strikeouts pitched, team wins, and player career length. They conclude:

One might expect that performance-enhancing drugs would raise the level of competition across the board, for pitchers as well as batters, since both increased strength and speedy recovery can contribute to high career tallies. However, we see evidence of competitive advantage mainly in the case of home runs. This indicates that the level of competition between pitcher and batter are tipping in the favor of the batter, possibly as a result of widespread performance-enhancing drug use.

I find this statement truly weird: One might expect A would cause B, however we see evidence for ~B, possibly as a result of A. *narrows eyes* If you could somehow subtract the expected 5% doper distribution and reproduce the 1920-1979 distribution, I’d be willing to buy it. Then again, aren’t there records of steroid abuse before 1979?

There’s also this paper: Paradoxical popups: Why are they hard to catch?: Which is a more or less qualitative analysis of ball dynamics where the aerodynamic effect of ball spin causes frequent fielder misjudgment. No real data, but they do cite conversation with 6 MLB players. Fun if your into baseball and physics.


3 responses to “Bayesball: Update

  1. Note that the “strikeouts” graph (lowerleft, fig. 2) indicates improved performance for pitchers as well (not as striking, but it looks noteworthy to me) as the batters. What’s funny is that although there are many more home runs, and a few more strikeouts, the “wins” graph hasn’t changed much. So homeruns are just a vanity thing? Or they still don’t occur often enough to matter?

    This 5% thing is a lowerbound they say, and a pretty loose one…

    Yeah, a latent variables analysis which tries to impute which players are the dopers would be fun to do…

  2. The authors do mention that, specifically in 1920 the league added restrictions on pitchers to increase the frequency of homeruns, and increased strikeouts could very well indicate that batters are taking more chances. I’m no baseball fan, so I can’t guess…

    I’m not sure how the participants in the Mitchell report were chosen, so the 5% could be uninformative.

  3. Pingback: Live Fast, Publish Hard « Imaginary Potential

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s