Creative Commons source reading · Load, readiness and fatigue
Advancing training effectiveness prediction in mass sport through longitudinal data: A mathematical model approach based on the Fitness-Fatigue Model
Wang W; Zhao Y; Hou X; Yu W; Ren H; Ren H
Originally published in PLoS One · 2025-12-03
From the article
Selected source text. JournalOS removes markup and shortens excerpts where indicated. The full article provides the complete context.
Abstract
AbstractDespite the critical need for scientific training load assessment in mass sports, the Fitness-Fatigue Model (FFM) requires further mathematical optimization and practical output indicators. The aim of this study was to optimize the mathematical relationship between “adaptation” and “fatigue” in the FFM, identify generalizable model output indicators, and evaluate its performance in predicting training effectiveness in mass sport. To account for the nonlinear and time-varying characteristics of training effectiveness, this study proposed new mathematical assumptions and optimized parameters against individual longitudinal data. The external load (speed and wattage) and internal load (wearable-compatible heart rate variability [HRV] and heart rate recovery [HRR] related indicators) of each training day were collected for 28–42…
3.2 Fitting effect evaluation result
Using ΔHRR1, HRr%, and TLHRV to calculate the output indicators for model fitting and parameter estimation separately (the ratio of external load to internal load on the same day), each individual obtained their own three fitting models parameters using their longitudinal data (S2-S4 Tables). When utilizing ΔHRR1 as a factor in calculating the output indicators of the model, the evaluation of the model fitting effect showed that all 13 subjects obtained small SSE and RMSE, with R2 values ranging from 0.54 to 0.84 (54% of subjects exhibited R2 values between 0.50 and 0.70, and 46% of subjects exhibited R2 values between 0.70 and 0.90). As for HRr%, the majority of subjects achieved good model fitting effect and low error, but there were also four subjects with R2 values less than 0.50. When it comes to TLHRV, most subjects achieved…
3.3 Model prediction ability evaluation results
Although the prediction error of some subjects’ fitting models is slightly higher than that of the original model, the overall predictive ability of the fitting model is still better. Temporal dependency analysis showed there were no significant relationships between prediction horizon and absolute percentage error for most subjects, indicating no population-wide consistency in temporal dependency of prediction accuracy. Significant but directionally inconsistent correlations were observed in a few cases, suggesting a highly individualized impact of temporal factors on prediction accuracy. The evaluation results of predictive ability are shown in Fig 3. Specific values for each subject, including those from the temporal dependency analysis, are provided in S8–S13 Tables. Taking subject 4 as an example, the 3D images of the fitted…
Editorial reading guide
- How did the authors define training exposure and performance response?
- What alternative explanations, confounders or individual differences limit the conclusion?
