Predicting Tour de France stage-winning times with continuous power and drag area models and high speeds in 2013

Abstract

We present comparisons between our predicted stage-winning times and actual stage-winning times for the 2012 and 2013 Tour de France races. The former race represented our last use of a decade-old discrete power and drag area models; the continuous power and drag area models used for the latter race represent significant changes in our modeling. Although our new model worked well when applied to the 2012 Tour de France, it did not fare so well in 2013, especially during the second week of the race when speeds were quite high.