Players' Perception Versus Players' Physiology: Differences in Acute and Chronic Internal Training Load Between Starters and Nonstarters in Elite Soccer
Related Experiment Video
Updated: Oct 31, 2026
An Inertial Measurement Unit Based Method to Estimate Hip and Knee Joint Kinematics in Team Sport Athletes on the Field
Published on: May 26, 2020
Damiano Li Volsi1,2, Paolo Riccardo Brustio2,3, Rocco Perrotta4,5
1Department of Medical Sciences,University of Turin, Turin, Italy.
Purpose:
The aims of this study were to analyze acute and chronic training-session internal training load (ITL) in starters (ie, playing the previous official match) and nonstarters Italian Serie A soccer players, and to examine whether player status moderated the relationship between session rating of perceived exertion (s-RPE), and players' heart rate-based training impulse (TRIMP). Acute and chronic workload were also examined using exponentially weighted moving averages (EWMAacute and EWMAchronic) and their workload ratio.
Methods:
The s-RPE and TRIMP data from 32 male players were collected during the 2024-2025 season, including 157 team training sessions and 2258 individual training session observations (948 starter and 1310 nonstarter). Linear mixed-effects models were applied to test differences between starters and nonstarters for s-RPE, TRIMP, EWMAacute, and EWMAchronic, with acute-to-chronic workload ratio retained as a secondary comparative metric. In addition, we assessed whether player status moderated the association between perceived and physiological ITL indicators.
Results:
Nonstarters showed significantly higher values than starters for s-RPE, TRIMP, and their respective EWMAacute, EWMAchronic, and acute-to-chronic workload ratio indicators (P ≤ .001; ηp2 range = .010-.076), although the difference for EWMAchronic s-RPE was trivial in magnitude (ηp2=.002). TRIMP was positively associated with s-RPE (β = 1.94; 95% confidence interval [CI] = 1.84 to 2.04; P < .001), and the TRIMP × player status interaction indicated a stronger association in starters than in nonstarters (β = 0.181; 95% CI = 0.053 to 0.309; P = .006). Similar moderation effects were observed for EWMAacute (β = 0.210; 95% CI = 0.036 to 0.384; P = .018) and EWMAchronic (β = 0.293; 95% CI = 0.129 to 0.456; P < .001).
Conclusions:
Training-session ITL differed between starters and nonstarters, and player status moderated the relationship between s-RPE and TRIMP-derived indicators, suggesting that starters and nonstarters may perceive physiological training load differently. These findings highlight the importance of considering player status when interpreting perceived and physiological ITL responses.
