|
Activity Number:
|
280
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 8, 2006 : 10:30 AM to 12:20 PM
|
|
Sponsor:
|
Section on Statistics in Sports
|
| Abstract - #307073 |
|
Title:
|
Regression Modeling Using Multiple Sources To Detect Group-Level Differences in Pitching Performance
|
|
Author(s):
|
Liam O'Brien*+ and Matthew Aschaffenburg
|
|
Companies:
|
Colby College and Colby College
|
|
Address:
|
Department of Mathematics, Waterville, ME, 04901,
|
|
Keywords:
|
covariance modeling ; pitching ; multiple sources
|
|
Abstract:
|
There are many different statistics used as metrics of pitching performance - both at the individual level and at the team level. We use three metrics as benchmarks of pitching performance for starting pitchers over a five-year period. Of interest is whether there are group level (i.e., league or handedness) differences in how these outcomes vary through time. Difficulty arises in modeling the association amount the metrics within player. To circumvent this, we use a Kronecker product covariance structure to reduce the number of association parameters. This makes estimation of the effects of interest more efficient and detection of league differences through the years easier.
|
