The extensive use of performance tests in diet intervention studies mirrors the importance of such a measurement. Although many different endurance performance tests have been used in the past, the majority of these different protocols has never been validated. In this study reproducibility of three different endurance performance tests was evaluated. Thirty well-trained subjects were matched on age, weight, and Wmax and divided into three subgroups. Each group of subjects performed one of three exercise protocols: protocol (A) consisted of cycling at 75% Wmax until exhaustion. In (B) subjects received a preload of 45 min 70% Wmax and then performed as much work as possible in 15 min. (C) consisted of a time trial, in which subjects had to complete a preset amount of work as fast as possible. Each subject performed one of the trials six times. Coefficient of variation (CV) was calculated for each protocol. CV(A) was 26.6%, CV(B) 3.49%, and CV(C) 3.35%. It was concluded that reproducibility of a test at 75% Wmax until exhaustion is poor and these tests are not reliable. Time trial protocols may result in better performance evaluation.