|
Activity Number:
|
272
|
|
Type:
|
Contributed
|
|
Date/Time:
|
Tuesday, August 4, 2009 : 8:30 AM to 10:20 AM
|
|
Sponsor:
|
Section on Statistics in Sports
|
| Abstract - #304917 |
|
Title:
|
Accounting for Home Ice Advantage in Women's College Hockey Rankings: A Bayesian Approach
|
|
Author(s):
|
Michael A. Rutter*+
|
|
Companies:
|
Penn State University Erie
|
|
Address:
|
4205 College Drive , Erie, PA, 16428,
|
|
Keywords:
|
Sports ; Ranking ; Bayesian ; Hockey
|
|
Abstract:
|
To determine the at-large bids to the eight team NCAA Division I Women's Ice Hockey tournament, four factors are considered. The Ratings Percentage Index (RPI) is arguably the most important of the four as it is used as a tie breaker. This paper proposes an alternative to the RPI that utilizes individual game results as opposed to a team's record and the records of its opponents. The proposed method also improves on RPI by explicitly allowing for a tie as an outcome and accounting for a possible home ice advantage. While the home team may have limited tactical advantage, the schedule and type of travel may have an effect. A typical weekend road trip consists of consecutive games played in a 24 hour (or less) window preceded by a long bus trip. A Bayesian approach is used to incorporate prior information on home ice advantage and the probability of a tie.
|
