When Chelsea’s ownership embarked on a comprehensive transformation of the club’s staff and operations shortly after their takeover in 2022, they pinpointed data science as a critical area where Premier League teams were lagging behind their counterparts in American sports. In leagues like the NFL and NBA, franchises typically maintain extensive data departments, comprising numerous experts and analysts dedicated to harnessing vast arrays of statistics to enhance performance.
With Todd Boehly’s extensive experience in sports management as co-owner of the Los Angeles Dodgers and Los Angeles Lakers, alongside Behdad Eghbali and Jose Feliciano, who amassed their fortunes primarily through technology investments, Chelsea’s leadership initiated a radical reimagining of their data department. According to The i Paper, the club is now heavily investing in advancements in artificial intelligence.
Read Next
Cole Palmer reveals why he ‘really didn’t want to go’ to Chelsea
Employing a sophisticated system, Chelsea can evaluate every player across various positions, utilizing hundreds of data points to make informed recommendations on optimal player combinations based on their respective strengths and weaknesses in different formations and tactical setups.
Two pivotal hires have been instrumental in advancing the analytics domain at the club. Sachin Gupta, a 42-year-old with a computer science degree from MIT, joined Chelsea as chief analytics officer last November after nearly two decades of experience with NBA teams. His appointment raised eyebrows within football circles, given his impressive background.
Gupta is best known for developing the NBA Trade Machine, a widely used platform that simplified the complex trade regulations of basketball during his time at ESPN. He later transitioned to basketball teams, where he gained a reputation for merging analytical data with a profound understanding of player and coaching dynamics—areas that have traditionally been at odds. As Minnesota Timberwolves general manager Scott Layden aptly described him, Gupta is regarded as “the smartest guy in the room.”
In January, Javier Fernandez, a leading expert in AI and football analytics, joined the department as director of data science. With a PhD in AI, Fernandez previously held the position of head of analytics at Barcelona for five years before spending the last four years at Zelus Analytics, which is recognized as “the world’s leading sports analytics platform.” His tenure at Barcelona is notable for integrating data and AI into the core of the club, which is considered groundbreaking. He effectively utilized AI to reshape player evaluations and analyses, influencing tactical decisions significantly.
As highlighted on Barcelona’s website, Fernandez was “involved in a continuous process of mapping football concepts and the club’s playing philosophy into actionable algorithms.” His contributions coincided with Barcelona’s achievement of two La Liga titles during his tenure.
At Zelus, Fernandez focused on developing sophisticated models for valuing players and teams. He shared on LinkedIn, “Working alongside some of the sharpest minds in sports analytics was truly transformative. It’s remarkable how much you can learn just by being immersed in a highly creative environment surrounded by brilliant people.”
On Thursday, Fernandez announced two job openings for a full-stack web developer and a data engineer, both of which would play crucial roles in decision-making processes at Chelsea.
AI is making significant strides in transforming the landscape of professional sports, as noted by The i Paper. Industry insiders have revealed some astonishing capabilities of AI applications in football. One AI platform, collaborating with Premier League clubs, claims to have “cracked the code” for successful promotion from the Championship to the Premier League. While skepticism persists throughout the sport, advocates of this technology maintain that it serves as a powerful tool to assist head coaches rather than replace them. “We compare it to an X-ray,” explains Jan Wendt, founder of the AI company PLAIER. “Imagine the time before the first X-ray was delivered to a doctor. We’re not replacing a doctor; we’re simply providing a tool to help them make better diagnoses.”
Our Test: Cole Palmer on the Bench?
Read Next
Cole Palmer is Straining Under the Weight of Chelsea’s Directionless Football
Apart from elite sports supercomputers, I endeavored to assess the knowledge of freely available commercial AI systems. I posed the same question—“What is Chelsea’s strongest starting line-up?”—to three distinct AI platforms.
ChatGPT, which ignited a trillion-pound tech arms race, provided a notably inaccurate response. It included a player who had departed in the summer, another currently on loan, and yet another facing suspension due to a failed drug test. The suggested formation was:
4-3-3:
- Sanchez;
- James, Silva, Colwill, Cucurella;
- Caicedo, Fernandez, Sterling;
- Madueke, Jackson, Mudryk.
Cole Palmer was relegated to the bench, as were Ben Chilwell, who is currently on loan at Crystal Palace, and Kepa Arrizabalaga, who has not played for Chelsea since 2023 and is now on loan at Bournemouth.
In a slightly improved response, Microsoft’s Copilot AI suggested:
4-3-3:
- Sanchez;
- James, Fofana, Badiashile, Cucurella;
- Fernandez, Caicedo, Palmer;
- Madueke, Nkunku, Mudryk.
This AI also failed to recognize that Mudryk is currently suspended.
Meanwhile, Google’s Gemini, after initially hesitating to give an answer, proposed a 4-2-4 formation:
4-2-4:
- Sanchez;
- James, Chalobah or Fofana (it couldn’t decide), Colwill, Cucurella;
- Fernandez, Caicedo;
- Palmer, Madueke, Neto, Jackson.
This exercise was primarily for amusement. The systems and models—often referred to as “tools” by their developers—used by clubs are significantly more sophisticated and targeted than the popular large language models.
Finger Pricks and Transfers
AI is infiltrating various facets of football. For instance, Championship side Norwich City, despite having one of the smallest staff teams in a category one academy, effectively leverages AI to maximize efficiency. They collaborate with Hudl, an AI platform designed to enhance team performance, utilizing their “Focus Flex” cameras that automatically track action during training and upload it in real-time. By combining internet cameras and drones, players can review their performance on their smartphones shortly after returning to the changing room.
Premier League clubs are also partnering with Orreco, a biostatistics company that utilizes simple finger prick blood tests to provide detailed insights into players’ health, rest, and recovery. One club doctor reported a remarkable 74 percent reduction in soft tissue injuries due to this technology. Last year, The i Paper revealed that Liverpool had teamed up with Google DeepMind to predict how opponents would defend during corner kicks, striving to devise optimal tactics for scoring. Meanwhile, Manchester United is believed to be employing AI in their recruitment processes.
Read Next
This Man Utd Stadium Update is Classic PR Guff to Cover Up Ratcliffe’s Blunders
At PLAIER, Wendt elaborated on how his system operates through a three-stage process. He remains discreet about the specific clubs he collaborates with, only disclosing that they include Premier League and Championship teams, as well as others across Europe. He emphasizes that he doesn’t want the accomplishments of the coaches to be overshadowed. However, he indicated the breadth of their impact by stating that during the January transfer window, PLAIER was involved in 163 transfers, 103 of which were successfully completed worldwide.
- 1. They assess a league and club, determining their performance levels in comparison to others and identifying what they need to achieve their sporting goals, including evaluating the coach’s effectiveness.
- 2. They analyze the squad to uncover strengths and weaknesses.
- 3. Based on the findings from the first two stages, they assist in recruiting the right players.
“We make football measurable and then predictable,” Wendt asserts. “In our system, we have over 370,000 players globally, acquiring data from every available source.” He further explains that “a footballer is arguably the most transparent employee in the world, second only to pilots and astronauts.”
“On average, we gather approximately 140 data points per minute for each player, encompassing event and tracking data. We aggregate this information to derive a single score for each player, clearly indicating where they can play globally. The volume of data is so substantial that it exceeds human processing capabilities. An average Premier League player typically has a score of around 5,000, with top players exceeding 9,000.”
Wendt also makes an astounding claim: “We’ve cracked the code. There are five essential factors that must be met to achieve promotion from the Championship to the Premier League. We can explicitly address each of these.” He concludes, “If you follow these five principles and have the basic financial structure in place, promotion to the Premier League is virtually guaranteed. The answers lie within the existing data.”
Silos and Rewards
Currently, clubs are at varying stages of experimentation with AI technology, often trying one aspect while trialing another in the academy, but many continue to operate in isolated silos. It is widely believed that once all these elements are integrated into a comprehensive system that permeates every facet of a club and becomes ingrained in its culture, the results will be transformative.
Several individuals working in various capacities at Premier League clubs have conveyed to The i Paper their conviction that the first team to fully embrace AI will reap significant rewards. “I wholeheartedly agree,” Wendt states. “The first team to adopt this will have a substantial advantage with minimal risk. Ultimately, the effectiveness of the system will quickly become evident, leading to a tipping point where adoption becomes essential. Those who do not embrace it will find themselves at a disadvantage. That moment is approaching.”
PLAIER boasts a team of around 15 individuals, primarily data scientists and AI specialists. “There is skepticism in football, and I completely understand it,” Wendt asserts. “Everyone working in this industry has been trained by traditional systems for decades. However, those who embrace AI now will gain a major advantage, uncovering hidden gems and insights that can shape team strategies.”
“Our aspiration is to help one of the underdog teams in the Premier League qualify for the Champions League and to assist clubs from the Championship in reaching the Premier League. We aim to achieve the unexpected.”
