Early Identification of Talent in Cyclo-Cross by Estimating Age-Independent Ability via Probit Regression

Early identification of talented athletes is a goal of team managers, sponsors, and other stakeholders so that the benefits of resources devoted to sponsorship and salary can be maximized. A maximum likelihood method is presented that assigns athletes in a chosen set of races an overall ability rating based on their results. Overall ability is expressed as a quadratic function of age so that the relationship between ability and age is estimated simultaneously. This enables relative current ability and age-independent ability to be estimated. After excluding athletes near or beyond the peak age of performance, the most talented young athletes can be identified by a ranking based on the estimates of age-independent ability. The method is demonstrated by analyzing performances in the 2012/2013 cyclo-cross World Cup.