Close Menu
Versa AI hub
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

What's Hot

Salesforce AgentForce 3 brings visibility to AI agents

June 25, 2025

Generated AI Media Production | Deloitte Us

June 24, 2025

6 Key findings from marketing leaders

June 24, 2025
Facebook X (Twitter) Instagram
Versa AI hubVersa AI hub
Wednesday, June 25
Facebook X (Twitter) Instagram
Login
  • AI Ethics
  • AI Legislation
  • Business
  • Cybersecurity
  • Media and Entertainment
  • Content Creation
  • Art Generation
  • Research
  • Tools
Versa AI hub
Home»Tools»TacticAI: AI assistant for soccer tactics
Tools

TacticAI: AI assistant for soccer tactics

By December 26, 2024Updated:February 13, 2025No Comments8 Mins Read
Share Facebook Twitter Pinterest LinkedIn Tumblr Reddit Telegram Email
Share
Facebook Twitter LinkedIn Pinterest Email

the study

Published March 19, 2024 Author

Written by Zhe Wang and Petar Veličković

As part of our multi-year collaboration with Liverpool FC, we developed a complete AI system that can advise coaches on corner kicks.

“The corner was taken right away… Origi!”

Liverpool FC made a historic comeback in the 2019 UEFA Champions League semi-finals. One of the most iconic moments was Trent Alexander-Arnold’s corner kick, which Divock Origi converted into a goal that will go down in history as Liverpool FC’s greatest goal.

Corner kicks have a high chance of scoring, but devising routines requires relying on a combination of human intuition and game design to identify patterns in rival teams and react on the fly.

Today, Nature Communications is introducing TacticAI, an artificial intelligence (AI) system that can provide experts with tactical insights, especially around corner kicks, through predictive and generative AI. Despite the limited availability of gold standard data on corner kicks, TacticAI uses a geometric deep learning approach that helps create more generalizable models, leading to state-of-the-art Achieve results.

We developed and evaluated TacticAI in collaboration with experts at Liverpool Football Club as part of a multi-year research collaboration. TacticAI’s suggestions were preferred by human expert evaluators 90% of the time over the tactical settings seen in real life.

TacticAI demonstrates the potential of assistive AI technology to revolutionize sports for players, coaches, and fans. Sports like soccer are also dynamic areas for AI development because they feature real-world multi-agent interactions with multimodal data. Advances in AI for sports could be applied to a variety of fields on and off the field, from computer games and robotics to traffic coordination.

TacticAI is a complete AI system that combines predictive and generative models to analyze what happened in previous plays and how to adjust to make certain outcomes more likely.

Developing a game plan against Liverpool FC

Five years ago, we began a multi-year collaboration with Liverpool FC to advance AI for sports analytics.

In our first paper, Game Plan, we looked at why AI should be used to support football tactics, focusing on examples such as penalty kick analysis. In 2022, we developed Graph Imputer. This demonstrated how AI can be used in a prototype system for predicting downstream tasks in soccer analysis. The system can predict player movements off-camera even when tracking data is not available. If not, clubs will need to send scouts to watch matches in person.

We have now developed TacticAI as a complete AI system that combines predictive and generative models. Our system allows coaches to sample alternative setups for their athletes for each routine of interest and directly assess the likely outcomes of such alternatives.

TacticAI is built to address three key questions:

What happens in the tactical setup for a particular corner kick? For example, who is most likely to receive the ball and whether there will be a shot attempt. Once setup is complete, can you understand what happened? For example, have similar tactics worked well in the past? How can you adjust the tactics to achieve a specific result? ?For example, how should the position of the defending player be changed to reduce the chance of a shot?

Predicting corner kick results using geometric deep learning

A corner kick is awarded if the ball crosses the byline after touching a player from the defending team. Predicting the outcome of a corner kick is complex due to the randomness of gameplay by individual players and the dynamics between players. This is also difficult for AI to model due to the limited gold standard corner kick data available. In the Premier League, only around 10 corner kicks are played in each game each season.

(A) How to convert the corner kick situation into a graphical representation. Each player is treated as a node in the graph. A graph neural network operates on this graph and uses message passing to update the representation of each node.

(B) How TacticAI handles a given corner kick. All four possible combinations of reflections are applied to the corners and fed into the core TacticAI model. These interact to compute the final player representation and can be used to predict outcomes.

TacticAI successfully predicted corner kick plays by applying a geometric deep learning approach. First, we directly model the implicit relationships between players by representing the corner kick setup as a graph. In this graph, nodes represent players (with characteristics such as position, velocity, and height), and edges represent relationships between players. Next, take advantage of the near symmetry of a soccer pitch. Our geometric architecture is a variant of the group equivariant convolutional network, which generates all four possible reflections for a given situation (original, H-inverted, V-inverted, HV-inverted), and the receiver and shot attempts. Forces the prediction of to be the same in all environments. All 4 people. This approach reduces the search space of possible functions that a neural network can represent to functions that respect reflection symmetry, producing a more generalizable model with less training data.

Provide constructive suggestions to human experts

By leveraging its predictive and generative models, TacticAI can assist coaches with finding similar corner kicks and testing different tactics.

Traditionally, to develop tactics and counter-tactics, analysts would repeatedly watch many match videos, looking for similar examples or studying rival teams. TacticAI automatically calculates numerical representations of players, allowing experts to easily and efficiently examine relevant past routines. We further verified this intuitive observation through extensive qualitative research with soccer experts. As a result, TacticAI’s top 1 searches are related 63% of the time, which is almost twice the benchmark of 33% seen with approaches that directly analyze player positional similarity to suggest pairs. I understand that.

TacticAI’s generative models also allow human coaches to redesign corner kick tactics to optimize the probability of certain outcomes, such as reducing the probability of shot attempts in defensive setups. TacticAI provides tactical recommendations that adjust the positions of all players on a particular team. From these suggested adjustments, coaches can more quickly identify important patterns as well as players who are key to the success or failure of a tactic.

(A) An example of a corner kick where a shot was actually attempted.

(B) TacticAI can generate counterfactual settings with reduced shot probability by adjusting the defender’s positioning and speed.

(C) The proposed defender position reduces the receiver probability for offensive players 2 through 4.

(D) The model can generate multiple such scenarios, allowing the coach to examine different options.

Our quantitative analysis showed that TacticAI accurately predicted corner kick receiver and shot situations, and player position changes were similar to how the play actually unfolded. We also qualitatively evaluated these recommendations in a blind case study in which the evaluators did not know which tactics were involved. Which are from actual gameplay and which are generated by TacticAI. Liverpool FC’s human football experts found that our suggestions were indistinguishable from real corners and were favored over the original situation 90% of the time. This shows that TacticAI’s predictions are not only accurate, but also useful and deployable.

Examples of strategic improvements suggested by TacticAI that raters preferred over the original play:

(A) Most evaluators favor the recommendation of four players.

(B) Improved coverage of the defender furthest from the corner.

(C) Improving the covering runs of the central defender group inside the penalty box

(D) The tracking runs of the two central defenders have been significantly improved, as has the positioning of the other two defenders within the goal area.

Evolution of AI for sports

TacticAI is a complete AI system that can provide coaches with immediate, broad and accurate tactical insight that is actionable on the field. With TacticAI, we have developed a capable AI assistant for soccer tactics and achieved a milestone in the development of useful assistants in sports AI. We hope that future research will help develop assistants that extend to a more diverse range of inputs than player data and can assist experts in more ways.

It shows how AI can be used in soccer, but soccer can teach us a lot about AI. This is a very dynamic and difficult game to analyze, with a lot of human factors from physique to psychology. Even for experts like experienced coaches, it is difficult to detect all patterns. With TacticAI, we hope to learn many lessons in developing broader assistive technologies that blend human expertise and AI analysis to help people in the real world.

Learn more about TacticAI

This project is a collaboration between the Google DeepMind team and Liverpool FC. TacticAI authors include Zhe Wang, Petar Veličković, Daniel Hennes, Nenad Tomašev, Laurel Prince, Michael Kaisers, Yoram Bachrach, Romuald Elie, Li Kevin Wenliang, Federico Piccinini, William Spearman, Ian Graham, Jerome Connor, Yi Yang, Adrià Includes: Recassens, Mina Khan, Natalie Vogellanger, Pablo Sprechmann, Pol Moreno, Nicolas Hess, Michael Bowling, Demis Hassabis, Karl Tuils.

author avatar
See Full Bio
Share. Facebook Twitter Pinterest LinkedIn Tumblr Email
Previous ArticleEthics of advanced AI assistants
Next Article AI exit: Aporia acquired by Coralogix

Related Posts

Tools

Salesforce AgentForce 3 brings visibility to AI agents

June 25, 2025
Tools

Sglang transformer backend integration

June 24, 2025
Tools

Unified interface for robotics zero-shot vision models

June 23, 2025
Add A Comment
Leave A Reply Cancel Reply

Top Posts

New Star: Discover why 보니 is the future of AI art

February 26, 20253 Views

How to build an MCP server with Gradio

April 30, 20251 Views

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20251 Views
Stay In Touch
  • YouTube
  • TikTok
  • Twitter
  • Instagram
  • Threads
Latest Reviews

Subscribe to Updates

Subscribe to our newsletter and stay updated with the latest news and exclusive offers.

Most Popular

New Star: Discover why 보니 is the future of AI art

February 26, 20253 Views

How to build an MCP server with Gradio

April 30, 20251 Views

The UAE announces bold AI-led plans to revolutionize the law

April 22, 20251 Views
Don't Miss

Salesforce AgentForce 3 brings visibility to AI agents

June 25, 2025

Generated AI Media Production | Deloitte Us

June 24, 2025

6 Key findings from marketing leaders

June 24, 2025
Service Area
X (Twitter) Instagram YouTube TikTok Threads RSS
  • About Us
  • Contact Us
  • Privacy Policy
  • Terms and Conditions
  • Disclaimer
© 2025 Versa AI Hub. All Rights Reserved.

Type above and press Enter to search. Press Esc to cancel.

Sign In or Register

Welcome Back!

Login to your account below.

Lost password?