External Load Indices Predict Technical–Tactical Performance Gains in Collegiate Basketball: A Controlled Intervention Study
Abstract
Background: External-load monitoring in basketball is widespread, yet its direct translation into technical–tactical performance gains remains poorly understood. This study investigated whether a constraints-led, game-based training program would lead to superior adaptations compared with a traditional technique-then-scrimmage model in collegiate basketball players. Methods: Forty-eight male collegiate athletes (aged 18–22) from two intact training units were assigned to either an experimental (game-based) or control (traditional) condition over eight weeks. Both groups trained three times per week (75–90 min / session) under matched volume and intensity (sRPE 6–7; 70–85% HR max ). Outcomes were assessed at Pre and Post: external load (distance, PlayerLoad™, accelerations, jumps), internal load (mean HR, session RPE), perceptual responses (fatigue, enjoyment via PACES), and technical–tactical performance (field-goal percentage [FG%], assist-to-turnover ratio [AST/TO]). Primary effects were tested using baseline-adjusted ANCOVA models with HC3 robust standard errors; effect sizes are reported as partial η². Results: Groups were equivalent at baseline (all p > 0.05). Compared with control, the experimental group showed greater Post improvements across external-load measures: distance +191 m, PlayerLoad +22.53 AU, accelerations +3.84, jumps +4.51 (all p < 0.01; η² = 0.32-0.75). Technical–tactical outcomes improved: FG% increased by 3.3 percentage points (~7% relative) and AST/TO by +0.19 (~18% relative) (both p < 0.05). Mean HR and session RPE did not differ between groups (p > 0.05). Perceptually, the experimental group reported lower fatigue (−0.52 points; 95% CI [−1.20, −0.20]; p < 0.01) and higher enjoyment (+7.27 points; 95% CI [2.90, 10.60]; p < 0.01). Holm–Bonferroni correction for multiple testing did not change primary inferences. Conclusions: In a collegiate basketball context, a constraints-led, game-based training model enhanced technical accuracy, tactical efficiency, external mechanical load, and enjoyment—without increasing physiological strain or perceived exertion. This evidence supports the integration of representative learning design with load monitoring as a viable approach to optimize performance transfer.