Abstract:
|
Games and sports in which all participants compete simultaneously (e.g. marathons, ice skating, etc.) result in all players being ranked against one another. Oftentimes these players compete many times against each other over the course of their careers. It is reasonable to assume that the player's ability will change over time as well as be effected by different characteristics unique to the athlete (e.g. age, gender etc.). We modify the rank order logit (ROL) model created by Glickman and Hennessy (2015) to include covariates in a time varying parameter that is representative of the player's skill. The results of the ROL model are combined with a latent trajectory model in order to categorize the progression of atheletes' careers. We apply this approach to international cyclists over a 15 year time period. The database is provided by the Cycling Quotient Ranking website and consists of male cyclists who competed internationally in races from 2000 to 2015.
|