Tracking key metrics: fluctuations in external and internal load across game quarters in collegiate basketball...

This study aimed to explore the quarter-by-quarter variations in external and internal load and their interrelationships throughout basketball games among collegiate basketball players.

This longitudinal observational study analyzed data from 18 official games, yielding a total of 470 data points (quarter 1 (Q1): 112, Q2: 128, Q3: 122, Q4: 108) from 14 male players of the Chinese University Basketball Association. Only players without injury records who completed at least 15 min per game and 5 min per quarter in at least five games were included in the final analyses. Catapult S7 devices and Ratings of Perceived Exertion (RPE) were employed to assess both external and internal load variables. Linear mixed-effects models and repeated-measures correlations were utilized to assess differences in external and internal load between backcourt and frontcourt players, as well as to examine the relationship between RPE and objective load metrics across game quarters. Decision tree visualizations were employed to analyze load thresholds and identify key quarter-specific features that differentiate player performance.

Backcourt players exhibited higher RPE (Q1: p < 0.01, ES = 0.72; Q2: p < 0.001, ES = 1.3), PlayerLoad (PL)·min^- 1 (Q1: p < 0.001, ES = 1.0; Q2: p < 0.001, ES = 1.0), and IMA COD (Q1: p < 0.05, ES = 0.69; Q2: p < 0.01, ES = 0.67) in the first and second quarters compared to frontcourt players. In contrast, frontcourt players showed increased IMA Accel in the third (p < 0.01, ES = -0.65) and fourth quarters (p < 0.05, ES = -0.90) compared to backcourt players. Additionally, significant correlations were found between RPE and PL (Q1: r = 0.51, p < 0.001; Q2: r = 0.46, p < 0.001; Q3: r = 0.57, p < 0.001; Q4: r = 0.61, p < 0.01) and explosive efforts (Q1: r = 0.61, p < 0.01; Q2: r = 0.46, p < 0.001; Q3: r = 0.44, p < 0.001; Q4: r = 0.45, p < 0.01) across all quarters, underscoring the connection between physical exertion and athlete perception. Finally, decision tree with heatmap visualizations identified RPE, PL, and PL·min^- 1 as critical feature distinguishing the physical demands of each quarter.

This study highlights position-specific physical demands across game quarters in basketball, with backcourt players showing higher intensity (RPE, PL·min^- 1, and IMA COD) early in games and frontcourt players exhibiting greater IMA Accel later, supporting tailored training and recovery strategies based on RPE and objective load metrics.