Author(s): THRON, M., BUCHHEIT, M., RUF, L., SONDERMANN, Y., WOLL, A., ALTMANN, S., Institution: KARLSRUHE INSTITUTE OF TECHNOLOGY, Country: GERMANY, Abstract-ID: 1428

INTRODUCTION:
Peak speed achieved during incremental running tests (vpeak) as a proxy of maximal aerobic performance is widely used in professional soccer to evaluate aerobic fitness and to prescribe high-intensity training. However, maximal testing is not always feasible in congested competition schedules or during return-to-play phases. Therefore, the present study examined whether commonly assessed submaximal parameters can be used to estimate vpeak.
METHODS:
A total of 2,648 test observations from 889 highly trained (tier three and four) male and female soccer players (12.8–37.6 years) were analyzed after filtering for exhaustion criteria. Athletes performed incremental running tests either on a treadmill or on a Tartan™ track. Lactate-derived thresholds (e.g., running speed at 4 mmol·L⁻¹; vL4) and submaximal heart rate responses (fractional utilization of true HRmax and estimated HRmax (220 minus age) at 12 km·h⁻¹; %HRmax12km·h⁻¹ and %estHRmax12km·h⁻¹) were collected. Different linear mixed-effects models were applied to identify relevant predictors of vpeak, considering test mode, age, sex, vL4, and %HRmax12km·h⁻¹. Model robustness and practical accuracy were evaluated using k-fold cross-validation.
RESULTS:
vL4, age, and sex already provided high predictive value (conditional R² = 0.80; RMSE of cross-validation = 0.64 km·h⁻¹), while the inclusion of %estHRmax12km·h⁻¹ further reduced estimation error (conditional R² = 0.81; RMSE of cross-validation = 0.63 km·h⁻¹). The most accurate model included age, sex, vL4, and %HRmax12km·h⁻¹, explaining a large proportion of variance in vpeak (conditional R² = 0.83) with a cross-validated RMSE of 0.60 km·h⁻¹, whereas including test mode did not improve the model fit further. An online calculator can be assessed here: https://vpeak.shinyapps.io/vpeak_prediction/
CONCLUSION:
From a practical perspective, these findings indicate that routinely collected submaximal markers such as vL4 allow practitioners to estimate vpeak with an error below 0.64 km·h⁻¹ and with %estHRmax12km·h⁻¹ or %HRmax12km·h⁻¹ further improving accuracy. While estHRmax can be easily calculated, true HRmax can be assessed in high-intensity training sessions or matches for improving estimation accuracy. This approach provides a time-efficient and less demanding alternative to maximal testing and may support training prescription, performance monitoring, and load management in elite soccer environments.