Training load (TL) and recovery should be in optimal balance to obtain maximal performance gains. We aimed to study sleep as a recovery technique and its relationship with TL and performance in elite athletes. Twenty‐six elite female artistic gymnasts were divided into an under 13 (n = 6), an under 14 (n = 6), a junior (n = 7; 14–15y) and a senior (=World Championship (WC) competitors, n = 7; ≥16y) category. Sleep, through sleep logs, and training parameters, using the session Rate of Perceived Exertion (sRPE) scale, were monitored to calculate total sleep time (TST), sleep efficiency (SE), TL, monotony and strain. Performance of WC competitors was evaluated through coach and WC qualification ranking. For the entire group, TST (effect sizes (ES) = −1.12, confidence intervals (CI) = −60:−47, P < .05) and SE (ES = −0.13, CI = −1.40:−0.10, P = .022) were shorter during week than weekend nights. TST and SE were highest in youngest gymnasts (P < .05). TL was lowest in under 13 and senior gymnasts (P < .05), while TL, monotony and strain were highest in junior gymnasts (P < .05). A negative regression was found between TST and TL the day after, while higher TL also led to lower TST the following night (P < .001). For the WC competitors, TST the night before the qualifications was shorter than the mean TST of the WC period (ES = −0.95, CI = −170:24, P = .030). TST correlated with coach ranking (r = −0.857, P = .014). Higher TL correlated with worse WC (r = 0.829, P = .042) and coach (r = 0.893, P = .007) ranking. This research in elite gymnasts indicated associations between decreased TST, augmented TL and inferior performance. Optimizing sleep and TL may therefore represent strategies to enhance performance.

This study aimed to identify the strongest predictive variables of winning and losing in the highest Belgian soccer division. A predictive machine learning model based on a broad range of variables (n = 100) was constructed, using a dataset consisting of 576 games. To avoid multicollinearity and reduce dimensionality, Variance Inflation Factor (threshold of 5) and BorutaShap were respectively applied. A total of 13 variables remained and were used to predict winning or losing using Extreme Gradient Boosting. TreeExplainer was applied to determine feature importance on a global and local level. The model showed an accuracy of 89.6% ± 3.1% (precision: 88.9%; recall: 90.1%, f1-score: 89.5%), correctly classifying 516 out of 576 games. Shots on target from the attacking penalty box showed to be the best predictor. Several physical indicators are amongst the best predictors, as well as contextual variables such as ELO -ratings, added transfers value of the benched players and match location. The results show the added value of the inclusion of a broad spectrum of variables when predicting and evaluating game outcomes. Similar modelling approaches can be used by clubs to identify the strongest predictive variables for their leagues, and evaluate and improve their current quantitative analyses.

Training-intensity distribution (TID), or the intensity of training and its distribution over time, has been considered an important determinant of the outcome of a training program in elite endurance athletes. The polarized and pyramidal TID, both characterized by a high amount of low-intensity training (below the first lactate or ventilatory threshold), but with different contributions of threshold training (between the first and second lactate or ventilatory threshold) and high-intensity training (above the second lactate or ventilatory threshold), have been reported most frequently in elite endurance athletes. However, the choice between these 2 TIDs is not straightforward. This article describes the historical, evolutionary, and physiological perspectives of the success of the polarized and pyramidal TID and proposes determinants that should be taken into account when choosing the most appropriate TID.