The present study proposed two models (visual and mathematical) to determine the three phases of "U"-pacing profile during a cycling time trial. The reliability of visual model was tested and models were compared. Fifteen cyclists performed a maximal incremental test and two 4-km TT. For the visual model, four experienced evaluators analysed twice the pacing, seven days apart. The mathematical model consisted on the mean of power output during phase 2 (1- until 3 km) plus two standard deviations, to distinguish phase 2 change points between phases 1 (CP1) and 3 (CP2). The CP1 occurred at 419 ± 186 and 415 ± 178 m for visual and mathematical model and CP2 occurred at 3646 ± 228 and 3809 ± 213 m, respectively. There was no difference between models for both CP (p <0.05). The within-evaluator visual model reliability for CP1 was ICC >0.87 and CP2 was ICC >0.96 (p <0.05), and between-evaluator reliability was ICC > 0.89 (p <0.05). Bland-Altman plots showed agreement between models, most the difference was <5%. The visual and mathematical models are reliable and produce similar values for determining main phases of the "U"-pacing profile during a cycling TT.
© Copyright 2017 International Journal of Performance Analysis in Sport. Taylor & Francis. All rights reserved.