This paper adapts the dose-response research tool, well established in pharmacological studies, to an exercise and performance setting. Training is measured in quantitative units as the dosage inputs, and their effects on fitness, fatigue, overtraining and performance responses are modelled. In this way, one can answer such questions as 'what performance level would be predicted if a certain amount of training was undertaken?' More specifically, athletes and their coaches are interested in designing a training schedule to maximize performance potential at some future date and minimize the risk of overtraining during that time, for some minimal training inputs. This approach leads to the practical recommendation that athletes should train more intensely but only on alternate days, for a 5 month session, in a triangular-shaped training profile, with the heaviest training occurring between weeks 12 and 4 before competition.