AI in soccer

I really wish ChatGPT could write this newsletter sometimes, but it can’t. Its database isn’t up-to-date, and it just doesn’t have the same experience watching soccer as I do, most of which I assure you has been totally sober.

This got me thinking: What is AI’s role in soccer?

Can players use AI-generated feedback to improve their game? Can coaches have ChatGPT generate a game plan for them? Can we introduce robot referees to reduce errors and call a fairer game? How can we as fans harness AI to enhance the viewing experience?

While it may be years before other cyborgs join Erling Haaland on the pitch, soccer-specific technological advancements have pushed the game to new heights in recent years, and – much like Haaland’s goalscoring form – are showing no signs of slowing down.

Officiating

From Diego Maradona's “hand of God” in 1986 to the controversial non-call in last weekend's Manchester Derby, officiating decisions have and will always have a massive impact on the sport. No matter the referee, they will not get every decision correct nor will their decisions be well-received by everyone.

Offside is the most frequently missed call in the sport, but might be no longer with the recent introduction of limb-tracking technology, which combines AI with cameras placed around the stadium to generate an image with virtual offside lines.

When offside is detected, the referee receives an alert from the system and can refer to this image to enhance his or her own judgement in making their final decision.

This effectively removes incorrect offside decisions from the game, and was such a smashing success in Qatar that the Premier League is considering adopting the technology in the near future.

Player performance

It's no secret that footballers will go to extreme means to enhance their performance. Just ask Zlatan Imbrahimovic, who apparently imposed a pre-match sex ban with his wife after one particularly poor performance on the pitch followed a long bedroom session the night before.

The advancement of sport-specific AI means that players will need to find other reasons to not have sex with their wives, as (soccer) performance-enhancing gadgets, software, and training techniques are popping up left and right.

ISOTechne is an intelligent analysis tool that provides individual and team feedback across a variety of technical and tactical areas, like passing, first touch, spin rate, and defensive structure.

The tool combines motion-detecting sensors with synchronized 4k cameras to capture play, outputting player and team-specific insights in a user-friendly dashboard. Pretty cool.

Players are increasingly adopting wearable technology like GPS “bras” and boot sensors, which provide instant access to performance metrics like distance covered, heart rate, touches, max speed, and leg use.

Products like these are widely implemented by professionals, and also offer up-and-coming players the opportunity to identify areas for improvement by comparing their metrics with other players on the app(s), such as Harry Kane and Tyler Adams.

Coaching

Coaches have access to the above technologies on a team scale, but unlike players are presented with the challenge of building cohesive game plans tailored to their players' strengths and opponents' weaknesses.

This accounts for tons of variables, requiring extensive knowledge of the game and nonstop studying; Pep Guardiola is a famous film junkie, but what if AI like ChatGPT could make a game plan for him?

So, as any good scientist would, I asked ChatGPT, "Make a soccer game plan for Manchester City facing Chelsea, including specific player roles and instructions." Here's what it said:

Pretty incredible, huh? The tactics and instructions are a bit obvious, but the fact that a computer can generate this in a matter of seconds is mind-blowing. 🤯

Hey, Spurs – we heard you need a coach? 

Fan experience

It's hard to pinpoint how exactly AI enhances the fan experience, since the customer journey of a fan – whether attending a game or watching at home -- is so varied for everyone. That said, we've decided to highlight two areas in which the soccer fan experience is enhanced by AI.

Anti-discrimination: Racial discrimination is unfortunately still prevalent in soccer, so the Dutch government is piloting the use of AI technology to put an end to discriminatory chants in Eredivisie stadiums.

The pilot is equipping the likes of PSV Eindhoven's and Feyenoord's stadiums with smart sound cameras that can detect abnormal noise like racial chanting. The system is manned by a human operator that can take action if necessary, but is also capable of providing live feedback to fans to encourage them to get louder (in a positive way).

Player statistics: bUt wE hAd hiGHeR xG tHaN yOU is probably a phrase you've heard before, often parroted by internet commenters who feel their team outperformed yours yet that wasn't reflected in the scoreline.

xG (expected goals) and xA (expected assists) are statistics that are regularly reported in today's world, but few fans fully understand them. Developed by Stats Perform, xG and xA leverage AI to indicate "the likelihood of a goal being scored" and "the likelihood that a given pass will become a goal assist," respectively.

Each statistic takes into account a myriad of contextual factors, many of which are overlooked by fans; xG accounts for 35 factors, including the player's position on the field, their view of the goal, and the position of the opposing goalkeeper. All of these factors combined yield an xG for that shot, which can be added up over the course of a game or season.