Exercise regimen is an integral part of sports coaching, the goal of which is to improve an individual's athletic performance. In an effort to better understand and optimize athletic performance, exercise physiologists have developed mathematical models of physical training and performance, but the forecasting ability of such models has been proven to be poor. We argue that the main reason for poor forecasting ability is in training load quantification methods, as all that currently appear in scientific literature quantify load with a single numeric value and thus models cannot distinguish between different types of training.
We present a new method for quantifying training load with which we determine the load for each type of training independently, we also present a new prediction model. The prediction ability of the new model is better. The average absolute prediction error of the most commonly used model is 50.1W, while the average absolute prediction error of our model is 19.1W.
|
