A better predictor of marathon race times based on neural networks

Abstract

A novel application of artificial neural networks is presented for the prediction of marathon race times based on performances in races of other distances. For many years Riegel’s formula was used for the prediction of time in running races, given the race time of a person in a different distance. Recently, two different models which perform better than the classic formula in the prediction of marathon times were published in the literature. This work shows how a new approach based on artificial neural networks outperforms significantly these recently published models for marathon time prediction.