March 23, 2010

No Pepper

Against the Fence

by Tommy Bennett

Phil Birnbaum has a good reminder of the difference between frequentist and Bayesian statistics.
The central Bayesian insight is that you need information about the world as it exists "prior" to the test. For example, if we want to know how a Triple-A player with certain statistics will perform in the major leagues, we need to know how Triple-A players (or some other appropriate subset) are likely to perform in the major leagues. However, Bayesian statistics are often and easily abused. 

J.C. Bradbury helpfully reminds baseball players to turn that frown upside-down.
Players who smile fully in their pictures tend to live longer. Under the assumptions of the model in the paper (Abel & Kruger), smily intensity explained 35% of the variation in survival rates. They use a hazard model to arrive that this odd result.

Missed the MIT Sloan Sports Analytics Conference? A video of the keynote is available here.
The talk features moderation by Michael Lewis and includes panelists Bill Simmons and Mark Cuban, among others.

Joe Nathan will undergo Tommy John surgery on Friday.
The procedure will be performed by Dr. David Altcheck, whose particular brand of TJ is known as "docking." Altchek also performed the surgery on Nathan's teammate Carl Pavano.