Journal Article - 11 January 2024
Chaoyi Gu, Varuna De Silva, Mike Caine
https://www.sciencedirect.com/science/article/pii/S0950705123008730?via%3Dihub
This article focuses on how artificial intelligence and analytics methods can be used to support understanding processes especially in more complex systems. Focusing mainly on process mining technique and examines methods, challenges and applications of process mining in these more complex contexts.
In this study the aim was to evaluate football teams by measuring how well they control open space on the pitch. It uses tracking and event data to create pitch control map sequences which outlines the areas each team controls during possession. Then based on average league performances, a machine learning model can then be trained to predict how space would be controlled. The teams actual performance is then compared to the expected performance and a metric showing how likely a situation leads to a goal. The difference quantifies team efficiency.
This paper also outlined how this model would help enhance anayltics for football:
Team Performance: The top four teams from the 2019–20 Premier League showed higher attacking and defensive efficiency than other teams, creating more valuable space when in possession and restricting space better when defending.
Match Analysis: When comparing teams it outlined how Team 2 had a strong first half but faded in the second, while Team 1 improved and capitalised on space as Team 2 pushed players forward. The model revealed how tactical shifts impacted possession quality and space control throughout the match.
Successful Counterattack Case Study: Team 1 scored via a counterattack where they effectively exploited space. The model showed how a defence with poor coordination and decision-making led to Team 2 losing valuable space, allowing Team 1 to outperform expected possession value (EPV) and score.
Individual Performance: The model also evaluated how players contibuted to the match. Player 50 (best passer) and Player 193 (most effective dribbler) ranked highest in space-creation through passing and dribbling. This application helps coaches assess their line ups and formations based on their real impact on space control.