Academic Journal - 19 March 2024

• Authors

https://www.nature.com/articles/s41467-024-45965-x.pdf

Article Aim

This article aims to look at how AI systems can be created to help assist football tactics (with a focus on corner kicks). The goal of this article is to identify patterns of how teams execute corners, predict outcomes such as who will received the ball and whether a shot occurred and general tactics. TacticAI is currently used in collaboration with Liverpool football club anaylsts and coaches to ensure it works in field.

Key Findings

TacticAi creates a structured representation of match situations and how they will pan out. They could also simulate alternative set ups and test how these changes increase or decrease the dynamics in play. It also uses a lot of predictive modelling to show who received the ball and whether it resulted in a shot occurring and generative modelling which suggests small but meaningful tactical adjustments such as repositioning players on the pitch.

<aside> <img src="/icons/graduate_pink.svg" alt="/icons/graduate_pink.svg" width="40px" />

Summary - Pre-match

• TacticalAi Could be really useful when it comes to pre-match analysis. One use could be scouting out the opponents, looking at similar past scenarios and how the opponents would defend transitions, attacks etc. It is good for identifying taatical weakness in the opponents and then tailoring their training sessions to focus on how to exploit unguarded zone etc.
• This would mean teams could use it to simulate what might happen again their opponent and generate the optimal player set up and also predict and prepare the most effective tactical setups based on historical data.
• It could also be really useful for the what-if scenario, looking at what would happen if they were to change players positions or roles in the team slightly and how it would improve the expected outcomes </aside>