|
Activity Number:
|
69
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Sunday, July 29, 2007 : 4:00 PM to 5:50 PM
|
|
Sponsor:
|
Section on Statistics in Sports
|
| Abstract - #308445 |
|
Title:
|
The Distribution and Simulation of Baseball Winning Percentages
|
|
Author(s):
|
Richard Auer*+
|
|
Companies:
|
Loyola College in Maryland
|
|
Address:
|
Loyola College in Maryland, Baltimore, MD, 21210,
|
|
Keywords:
|
Major League Baseball ; Multivariate Normal Distribution ; Simulation ; Conditional Normal Densities
|
|
Abstract:
|
Many major league baseball purists state that the wild card team receives an easy ride to the playoffs. After all, the wild card team is not even the champion of their own division. Using real major league baseball data and modeling a single season's worth of winning percentages as a multivariate normal random vector (with first place to last as the variates), a simulation study suggests that that the wild card team is typically much stronger than the weakest of the divisional winners. A goodness of fit study of the multivariate model is described along with a detailed look at how conditional normal densities are used to simulate many seasons of data, one place of finish at a time.
|
