(58) USING EXTERNAL LOAD METRICS TO PREDICT GAME PERFORMANCE IN HIGH SCHOOL BASKETBALL ATHLETES

Purpose: External load monitoring has become a valuable tool for practitioners and sports scientists to monitor and regulate practice and training loads. However, limited research exists investigating their usefulness as predictors of game performance in high school basketball athletes. Our aim was to understand if external load metrics provide predictive insight to performance, and if so, which are the most predictive. Methods: Catapult External load data from 239 youth basketball athletes [M/F: 202/36; 16.1 ± 1.6 yrs; 183.0 ± 11.3 cm; 76.8 ± 14.4 kgs] were collected during practices and games from the 2022-2023 and 2023-2024 school years. Potential predictor variables included overall averages from 50 different Catapult external load outcomes that are commonly used by Sports Scientists. Game performance was quantified using Hudl Game Stats (including value-point-system (VPS), points, rebounds, blocks, and steals). VPS is intended to be an indicator overall performance that is comprised of several in-game statistics (i.e. points, field goal percentage, turnovers, and more). To identify the optimal prediction model for game statistics partial least squares regressions (PLS) were used to identify possible predictor variables from the external load metrics. The PLS was followed by a stepwise regression to further identify associated predictors. Results: Average inertial movement analysis (IMA) high band jump count (p< 0.001), average practice session duration (p< 0.001) and their interaction (p=0.001)were significant predictors for VPS (R² = 22.9%). IMA high band jump count (p=0.003) and average practice session duration (p=0.01) were significant predictors for points (R² = 10.4%). Average time spent in Player Load Band 3 (moderate workloads) was inversely associated with blocks (R² = 6.0%, p = 0.004). Average time spent in Player Load Band 1 (low workloads) was inversely associated with steals (R² = 7.7%, p = 0.001). Conclusions: Average frequency of high intensity jumps and average practice session time were the largest predictors of overall game performance in high school basketball athletes, as defined by the value-point-system. Average frequency of high intensity jumps and average practice session time were the largest predictors of overall game performance in high school basketball athletes, as defined by the value-point-system. However, average external load metrics provided minimal insight into individual game statistics. Future research should integrate related covariates that may provide additional predictive insight such as, anthropometrics, age, directional external load metrics, and fitness. Practical Applications: While external load monitors have a role in managing acute and chronic training loads, these data provide coaches and sports scientists with context to the limitations of using external load metrics to predict performance in youth athletes. Acknowledgements: This study was funded by PepsiCo, Inc and International Management Group (IMG) Academy. The views expressed in this abstract are those of the authors and do not necessarily reflect the position or policy of PepsiCo, Inc or IMG Academy.