How Can AI Improve Performance In Football?

Introduction

I recently wrote about the role of AI in revolutionising football clubs. This week, I wanted to dig deeper into the most important part - how AI could influence a team's success on the pitch.

"Football is a game of fine margins" was one of Ole Gunnar Solksjaer's famous clichés in his time managing Manchester United (and he certainly knew a thing or two about those).

Match events like offsides, red cards and last-minute winners have always been the margins that have dictated games. But as the game continues to evolve, competition is getting fiercer. Clubs are looking more and more at the fine margins they can control to gain those extra one percents over their rivals. The answer is now proving to be within cutting-edge technology like Machine Learning and AI.

In this article, I'll look into some of the ways these technologies could change different aspects of the game like how a team prepares, how players recover and how to mitigate an opponent's threats.

First of all...

I think it's important to address the question on everyone's mind, as with every other industry, is AI going to take our jobs?

I don't believe it will in football. This is because the game relies on individual genius and intuition. Although AI can simulate certain aspects of a human's intuitive abilities, intuition involves complex context comprehension and emotional understanding that AI cannot replicate. Moreover, Football is a game full of ambiguity and subtle complexities and AI doesn't like uncertainty.

Instead, I believe the technology will serve as a tool to make clubs more efficient and improve decision-making. AI will help fix the costly errors clubs keep making and free up coaching resources to focus on continuous innovation and improvement.

Nobel prize-winning psychologist and author of 'Thinking: Fast and Slow', Daniel Kahneman, shares an interesting perspective on the future of AI. He refers to the point I make that AI has its limitations, but so do humans. Kahneman says that combining human judgement with AI would be the most powerful approach, as this can eliminate some of the cognitive biases that the human brain is prone to.

Potential sporting applications of AI in football

Data Analytics

Since the story of the Oakland Athletics and the release of Moneyball, the world of sports has begun to realise the benefits of using data analytics to improve performance. Football has been no exception, with the likes of Brentford Football Club and Liverpool Football Club becoming real pioneers in their uses of data. Data science has now become widespread across the game globally, but it's still very much in its infancy.

Whilst clubs are collecting and paying for vast amounts of data, much of that data isn't being utilised effectively, if at all. This is where AI and ML could step in to advance the field and the game. Here are some examples of how...

Decision-Making

One of the most crucial elements of football is decision-making. Think back to when Harry Kane should have squared it to Raheem Sterling to send England to the World Cup final in 2018. The split-second decisions of individuals can have huge impacts, on the game and further.

I think that AI could play a major role in evolving how we approach in-game decision-making. Imagine a system that assesses different game scenarios and predicts the most effective actions to take. For example, individual left-wingers could become more informed on whether to shoot or pass when approaching the box, based on their own personal historical data as well as their opponents.

AI algorithms are capable of analysing millions of data points collected during training and matches, such as positional data, ball movement, and scoring probabilities. A deeper understanding of the information generated by these algorithms could equip coaching teams with new knowledge to train players to make better, more informed decisions during games.

Diet and Nutrition:

Long gone are the days of the steak and Guinness diet in Football - player diets have come a long way in recent years. Nevertheless, there's still a real opportunity for AI to progress nutrition to the next level.

Look at the examples being set by Bryan Johnson. At 45 years of age, his data-driven experiment has allowed him to reverse his fitness and organ health to the levels of an 18-year-old.

AI can be used to tailor dietary plans to individual players based on their genetic, dietary and performance data. By taking into account their individual needs, nutritional requirements, and performance goals, clubs can use these insights to optimise a player's diet and improve their overall fitness, power, and recovery periods.

Injury Prevention

If you think about the costs involved with an injury, then there's a big case to invest significantly into solutions that prevent them. Clubs can waste millions on transfer fees, wages and rehabilitation costs. Not to mention the impact on the team's performance when key players are missing.

AI can play a crucial role in preventing injuries by monitoring players' movement patterns and identifying potential risks. By assessing factors like fatigue levels, muscle imbalances and relative hardness or traction of the pitch, clubs can calculate and monitor the injury propensities of individual players at a given time.

AI could be used to design personalised training programs and rest schedules for each player. These algorithms could also alert coaches and medical staff of potential oncoming injuries, leading to quicker interventions and reducing chances of long-term harm.

Scouting and Recruitment:

Perhaps one of the most relevant use cases for AI in football would be around recruitment. The meteoric rise in transfer fees and player wages of recent years has been clear for all to see. Just look at how much money Manchester United has spent on transfer failures over the past 10 years. They're not alone though, most clubs have been guilty of wasting fortunes on failed transfers.

A key reason for this is that most recruitment decisions have typically been made on intuition and gut feeling. Imagine a solution that could minimise the risk of these nightmares by vetting a player/staff member to ensure they're the right fit for both the club's culture and performance needs at that moment.

AI can analyse vast amounts of player data from various sources, such as scouting reports, video footage, and a whole range of other statistics. By assessing these data points, clubs can identify the players and staff out there that align with their playing style and team requirements the most.

Further to this, they could also evaluate the future potential of a player by considering their physical characteristics, technical skills, and performance indicators - presenting an exciting prospect for academies to look at. Clubs could start making more informed decisions on when to make a big money signing or invest more into an alternative academy prospect.

Certain clubs like Liverpool, Burnley and Brentford have been showing clever uses of data in their recruitment. This explains why last season, Brentford managed to beat Manchester United, Manchester City and Liverpool with squad values 3-4x that of their own. But even then, their clever uses of data are still in their infancy and could be elevated further with AI.

Training and Coaching:

One of the most interesting applications of AI in coaching could be its ability to tailor training to each player's unique needs and goals. If a player wanted to develop in certain ways physically or technically, AI and ML solutions could adapt training regimes to help them reach these goals far more effectively.

Predictive modelling is posing an attractive prospect for coaches too. ML-powered tools can simulate game scenarios, allowing them to develop game strategies and inform decision-making. Imagine that you could simulate the possibilities of an opponent's set-plays or build-up plays for instance. It would be like having a virtual testing ground where they can devise, test and fine-tune winning tactics without the risk of a real game. Advanced AI algorithms can also understand match dynamics and provide tactical recommendations, ensuring that coaches have the most relevant information at their fingertips.

I anticipate that clever uses of coaching AI will make the game even more competitive, leading to more underdog upsets against the more talented sides.

Player Development:

In football, a lot can fall on development. Whether it's developing talent into more valuable assets or ensuring already valuable talent doesn't go to waste, the financial implications tied to player development are significant. By investing in AI, clubs can begin to eliminate the human error that's led to so much unfulfilled potential over recent years.

Teams can utilise AI here to identify players' strengths and weaknesses, track their progress, and provide personalised training and development plans. By assessing a player's potential and simulating the outcomes of different development paths, clubs can nurture talent effectively and maximise their chances of success.

Conclusion

Looking at some of the cases laid out here, the impact AI can have on a team's sporting performance is clear. With that in mind, I've found it surprising to see how seriously (or not) the majority of clubs are taking it.

Perhaps there's a lack of education in the area, or the off-the-shelf solutions out there are limited. Or maybe there's a general attitude of risk aversion across clubs towards investing in new technology. Either way, the examples set by Brentford and Brighton prove that investments in data science could be a very cost-effective alternative to big money transfers.

It may come down to the environment and ownership of the club. Brentford and Brighton's owners come from the world of gambling and have morphed their club's identities around what they know - data. Most clubs are run by business or football people who aren't quite as adept to the new age of data like these.

Regardless, whether there's an underperforming team that needs to change its fortunes, or a successful side looking to stay ahead of the competition, all clubs should be looking to improve their performance with AI.