How Football's Relationship with AI is Changing
Technology and football have a complicated relationship. Historically, football has been notoriously slow to embrace new technologies. Goal-Line technology was only implemented after the infamous England-Germany match at the 2010 World Cup, where England footballer Frank Lampard scored after the ball clearly bounced inside the German goal, only for the goal to be disallowed as the referee did not see it cross the line. It took another 6 years for the “Video Assistant Referee” (VAR) technology to follow, with VAR first being used in a friendly between France and Italy in June 2016. Traditionalists feel that too much technology usage could ruin the ‘beautiful game’ and as such there has been substantial resistance to the introduction of new technology.
However in recent years, the relationship between football and technology has changed and continues to change. Many clubs have realised that there is great potential for Artificial Intelligence to be used in a variety of ways within the world of football, both on and off the pitch.
In a way, the Lampard incident paved the way for the introduction of AI into football matches. Starting with the 2013-14 season, the Premier League chose to use Hawk-Eye computer systems, which were already in use in other sports such as cricket and tennis, to help with officiating by providing goal-line technology in football matches. Hawk-Eye uses cameras in many positions around the stadium and artificial intelligence is used to locate and identify the ball at all times. The ball only needs to be visible in two of the cameras to be found. If the ball is close to the goal line, AI is used to abstract the image of the goal-line and remove all obstacles, including players, before scanning to see if the ball has passed the line. Immediately after processing the scan, a signal is sent to the referee’s watch to inform the referee of the outcome. The technology is not perfect, as evidenced by an incident during last week’s Sheffield United-Aston Villa game, in which the technology failed to detect a ball crossing the line. However, this is the only recorded incident of the technology failing in the Premier League since its implementation in the 2013/14 season, with over 3000 matches played in total. This technology has helped to ensure fair officiating when determining whether a ball has crossed the line, removing all the ambiguity that previously existed.
Goal-line technology in action. This technology means there is no longer any room for ambiguity as to whether a ball has crossed the line.
Credit: Hawk-Eye
In addition to officiating, AI has seen usage on the pitch by football clubs. Non-league side Wingate & Finchley Football Club, in partnership with The Big Bang Fair, employed a “robot football coach” in 2019. The “coach” uses real-time data to make on field decisions during matches, including substitutions and formation changes. The technology employs a spoken interface through Amazon Echo, making it easily accessible for all coaches, regardless of their technological background. Its capabilities aren’t just limited to matches. The “coach” can generate pre-match reports assessing players’ fitness and motivation and the strengths, weaknesses and formation of the opposition, and use this data to make pre-match decisions which it feels gives the team the best chance of success.
Similar technologies were used by Leatherhead Football Club in the 2018/19 season. IBM’s tool, IBM Watson, took a variety of data from matches, including pass completion, build-up play, positioning and other aspects of matches, processing this data in order to build a picture of both the team’s collective performance and individual player performances, allowing the club to identify weaknesses and strengths and adapt their style of play and strategy going forward. The club, which were predicted by many to be relegated and endured dire form before the introduction of Watson, turned around their fortunes and almost managed to qualify for the promotion playoffs.
IBM Watson saw Leatherhead, who were predicted by many to be relegated, finish 5 points away from the playoffs.
Source: BetVictorIsthmian
Although AI might be useful for on-pitch analysis, a key area of interest to football clubs has been the potential for AI assistance with off-the-pitch decision making. In 2015, former Real Madrid footballer Esteban Granero founded Olocip, after identifying data analysis as a weakness in the footballing industry. Olocip uses AI to look at large data sets, taking the information in those data sets and processing it in order to form predictions and offer advice on the actions football clubs should take in a number of areas including scouting, physiotherapy and training. Whereas traditionally football clubs will make decisions based on historical data, for example, when signing a player, a football club might look at past performances and goal figures. Olocip instead focuses on the future, using AI to create a picture of how a player might perform at their new team or how a player might best fit into that team’s system. This focus on the future is highly attractive to football teams, where many decisions made involve large sums of money. Having an idea of what a certain tactic or purchase could produce provides more certainty moving forward for those clubs and allows them to make better informed decisions.
Moreover, Italian side AS Roma announced last month an “Artificial Intelligence Partnership” with the technology company Acronis, in which Acronis will provide “AI and Machine Learning solutions” to process a variety of data in many areas, including both the footballing and business side of Roma’s operations, as well as cyber security operations. This agreement isn’t the first of its kind. Acronis also has partnerships with other high profile clubs, among them the likes of Manchester City and Liverpool FC.
Certainly, the relationship is changing between football and technology. Almost every league which can afford it has purchased goal-line technology and with the move to VAR, technological assistant for match-day officials is likely to increase in usage. This changing relationship is also true for football clubs. Clubs from around the world and across divisions are showing an interest in new technologies and artificial intelligence, which can maximise their performance and ensure they make the best decisions for their teams. Perhaps clubs will start to move away from traditional scouting methods, instead using data driven analysis by AI which is able to take a more fact based and less subjective approach to choosing the right players for a team.