A reasonable alternative to the Wbal models when maximal mean power profiling is used instead of critical power-based models

Abstract

As an alternative to the W_bal models (Skiba and Clarke, IJSPP, 2021), a methodology based on maximum mean power profiling (MMP) is presented here to predict intermittent exercise performance potential. The methodology consists of 1) collecting the maximum values of exponentially weighted moving averages (EWM) with different time characteristics in a recent exercise history and 2) comparing the different EWM of a new session with the historical EWM. The methodology is applied here to six weeks of training data in a professional cyclist, using a set of EWM with time characteristics ranging from 12 seconds to 1 hour to create the MMP. During a new training session, the maximum ratio between each EWM and the corresponding historical maximum value provides the residual performance potential. A parallel can be drawn between the variations of the residual performance potential in this model and those of the W_bal model. On the one hand, the main advantage of the proposed method is that the accurate estimations of W' and CP are not needed, which allows wide applicability to field/outdoor data. On the other hand, the methodology presented here does not provide a direct physiological explanation for the variation in residual performance potential during exercise.