Entertainment Computing Journal, Volume 52 - January 2025
Chengjie Liu, Hongbing Liu
https://www.sciencedirect.com/science/article/pii/S1875952124002817
This Journal articles aim was to develop and evaluate an AI-based framework that enhances the tactical and cognitive training of football players through intelligent data analysis, predictive modelling and adaptive learning. The authors want to bridge the gap between artificial intelligence and tactical coaching by creating an adaptive training system capable of learning from match data, predicting player performance trends and improving both preparation and tactical execution in modern football.
This study proposes a Performance-Focused Strategic Training Module (PFSTM) that applies Artificial Intelligence (AI) specifically recurrent learning to help improve the tactical and cognitive training of football players. The model integrates data from previous matches and training sessions to identify “lagging features” such as slow decision-making, inaccurate passing, or low tactical awareness. It then uses iterative learning to personalize feedback and design targeted training sessions that maximize individual and team performance
The paper’s results demonstrate that PFSTM achieved 95% accuracy showing that it has a strong predictive reliability. AI can be used to combine performance attributes such as passing accuracy, awareness, defensive capacity and teamwork into adaptive learning cycles that will continually evaluates these through recurrent learning to refine strategy and decision-making in real time.
“The concatenation between the features overlaps under multiple strategy evaluation sessions to maximize player performance,” (Liu & Liu, 2025).
PFSTM can be used in pre-match preparation to simulate opponent tactics and assess how individual players respond to various in-game scenarios. By analysing individual players “performance-focused player features… from previous games,” coaches can design specific tactical drills before a match, addressing weaknesses identified through data modelling. AI can also forecast team formations or pressing efficiency based on historical patterns.
In the future coaches could apply PFSTM outputs to guide substitutions, formation changes or adjustments to passing strategies depending on detected fatigue or reduced reaction speed. This enhances mid-match adaptability and situational awareness.
After matches, PFSTM provides detailed performance diagnostics, identifying which skills or tactical elements didn't work. By evaluating “expected and current performance,” it helps create data-driven feedback loops to refine future training and strategy. Post-match integration of PFSTM enables trend tracking, ensuring players continuously improve cognitive, physical and strategic components.
In essence, Liu & Liu’s work represents a next-generation AI system that transforms traditional football coaching into a dynamic, data-informed process linking tactical intelligence with continuous performance optimisation.