Modeling performance of elite cyclists

Modeling performance of elite cyclists: The effect of training on performance

Rozendaal, Romke (TU Delft Electrical Engineering, Mathematics and Computer Science)

Jongbloed, Geurt (mentor)
van Erp, T. (mentor)

Delft University of Technology

Applied Mathematics

2017-11-01

Both coaches and cyclists nowadays determine the training schedules for important races based on experience
and gut-feeling. Data are available for the evaluation of training schedules, but a sound scientific
method of accomplishing this does not yet exist. The biggest hurdle in developing such a method for cycling
is the absence of an objective performance metric. This report uses the data of six cyclists in order to introduce
and analyze a new performance metric: the cycled critical power. Subsequently, this metric is used to fit
the parameters of the Fitness-Fatigue model; a model that is successfully applied in other sports to model the
performance of an athlete. Several training attributes are included in the analysis, such as duration, intensity,
and load; with which the performance of the cyclists is modeled. The model, applied to the data, leads to
correlations for the six cyclists with R2-values within the range of 0.13 - 0.56. Compared to other scientific
research, this range indicates that the Fitness-Fatigue model could be a valuable tool for elite cyclists and
coaches when constructing their training schedules, but further research into the performance metric and
the application of the Fitness-Fatigue model is still required.

http://resolver.tudelft.nl/uuid:06d0d081-f9b2-463f-85eb-096d54a4dcbd

Student theses

master thesis

© 2017 Romke Rozendaal