Tell me the match context, and I’ll tell you how much the team will run, or should have run: an update and divisional extension of a predictive running model for Spanish professional football
- 1University of the Basque Country, (UPV/EHU), Vitoria-Gasteiz, Spain
- 2Department of Competitions and Mediacoach, LaLiga, Madrid, Spain
- 3INEFC Lleida. University of Lleida, Spain
Abstract
To update the predictive model of a team’s total distance (TotDisTea) proposed by Castellano et al. (2024) for the Spanish First Division (LaLiga1), using five seasons (2021–22 to 2025–26), and to extend it to the Second Division (LaLiga2), from a dual perspective: a retrospective (evaluative) approach to assess performance already produced, and a prospective (predictive) approach usable before a match. Team physical performances from official matches were analysed (LaLiga1: 2,894 team-match records from 1,447 matches; LaLiga2: 3,542 from 1,771), recorded with a multi-camera system. For each division, an all-possible-regressions procedure selected the best linear model for TotDisTea from seven situational predictors: match location, match outcome, team and opponent level, effective-playing-time (EffPlaTim), ball-possession (BalPos) and the opponent’s total distance (TotDisOpp). A reduced model excluding TotDisOpp was additionally estimated for prospective use. Models were cross-validated at the match level and fitted with cluster-robust standard errors. In both divisions TotDisOpp was the strongest predictor (b= 0.72 and 0.79). The LaLiga1 model additionally retained team and opponent levels, EffPlaTim and BalPos (R2adj= 0.86); the more parsimonious LaLiga2 model retained TotDisOpp, EffPlaTim and BalPos (R2adj= 0.85). Match location and match outcome were not selected in either league. The reduced models retained EffPlaTim and BalPos, with lower accuracy (R2adj= 0.71 and 0.61). Opponent running is the most stable predictor in both divisions, but the predictive structure is league-specific and coefficients may change over time, warranting periodic updating. The equations let practitioners evaluate a team’s response after a match, and enable anticipation beforehand.
Related
- Does Preseason Training Load Shape Body Composition? External Load, Internal Load, and Fatigue in Semi-Professional Football Players.
- Isokinetic Knee Strength and Multidirectional Jump-Landing Asymmetry in Professional Football Players After Anterior Cruciate Ligament Reconstruction: An Exploratory Cross-Sectional Study
- Aerobic Fitness Is the Main Determinant of Match-Related Physical Performance Decline in Professional Football Players (Part 2)
- Comparing Match-Related Physical Performance Using Arbitrary and Individual Running Speed, Acceleration, and Deceleration Zones in Professional Football Players (Part 1)
- Long‐term market value after meniscal injury in professional football players largely follows expected career trajectories: A matched cohort study
BibTeX
@article{castellano2026tell,
title = {{Tell me the match context, and I’ll tell you how much the team will run, or should have run: an update and divisional extension of a predictive running model for Spanish professional football}},
author = {Julen Castellano and Roberto López-Del Campo and Ricardo Resta and Raul Hileno},
journal = {Biology of Sport},
year = {2026},
month = sep,
doi = {10.5114/biolsport.2026.164806},
url = {https://doi.org/10.5114/biolsport.2026.164806},
note = {CC BY 4.0},
}
RIS
TY - JOUR
TI - Tell me the match context, and I’ll tell you how much the team will run, or should have run: an update and divisional extension of a predictive running model for Spanish professional football
AU - Julen Castellano
AU - Roberto López-Del Campo
AU - Ricardo Resta
AU - Raul Hileno
PY - 2026
DA - 2026/09/11
JO - Biology of Sport
DO - 10.5114/biolsport.2026.164806
UR - https://doi.org/10.5114/biolsport.2026.164806
AB - To update the predictive model of a team’s total distance (TotDisTea) proposed by Castellano et al. (2024) for the Spanish First Division (LaLiga1), using five seasons (2021–22 to 2025–26), and to extend it to the Second Division (LaLiga2), from a dual perspective: a retrospective (evaluative) approach to assess performance already produced, and a prospective (predictive) approach usable before a match. Team physical performances from official matches were analysed (LaLiga1: 2,894 team-match records from 1,447 matches; LaLiga2: 3,542 from 1,771), recorded with a multi-camera system. For each division, an all-possible-regressions procedure selected the best linear model for TotDisTea from seven situational predictors: match location, match outcome, team and opponent level, effective-playing-time (EffPlaTim), ball-possession (BalPos) and the opponent’s total distance (TotDisOpp). A reduced model excluding TotDisOpp was additionally estimated for prospective use. Models were cross-validated at the match level and fitted with cluster-robust standard errors. In both divisions TotDisOpp was the strongest predictor (b= 0.72 and 0.79). The LaLiga1 model additionally retained team and opponent levels, EffPlaTim and BalPos (R2adj= 0.86); the more parsimonious LaLiga2 model retained TotDisOpp, EffPlaTim and BalPos (R2adj= 0.85). Match location and match outcome were not selected in either league. The reduced models retained EffPlaTim and BalPos, with lower accuracy (R2adj= 0.71 and 0.61). Opponent running is the most stable predictor in both divisions, but the predictive structure is league-specific and coefficients may change over time, warranting periodic updating. The equations let practitioners evaluate a team’s response after a match, and enable anticipation beforehand.
SN - 0860-021X
SN - 2083-1862
ER -