If you have ever watched the movie Moneyball, you will have an idea of how the world of data is being used in Football. Beane, who is played by Brad Pitt, aims to reinvent his team by outsmarting the richer ball clubs by challenging the old-school traditions. He does this by recruiting bargain-bin players who the scouts have labeled as flawed, but are known to have winning potential.
Similar to anything related to data and AI right? Many people believe that it’s flawed due to being dependent on old-school methods, but do not realise the level of opportunities and further understanding it provides us.
In the last 20 years, the implementation of using data in the football sector has exploded, with nearly every football club using tools such as machine learning. Tech experts and football clubs have come together to find new methods how to scout footballers, known as technical scouting. The tech experts use maths, statistics, and optimization in their decision-making process.
We have learned about all of this from our newest Data Scientist on the team; Stephane Moreau. Read more about Stephane in our ‘Meet the Team’ edition by clicking this link. Stephan has been working with Football and Data for a year and a half, where he particularly focused on the recruitment side. He was part of an agency that brought data from data providers and created tools to produce analyses for football clubs and players.
Data Providers for Football
Data providers do exactly what it says; they provide data. Though new computer vision techniques are now automating the process, the vast majority of data has been collected manually by teams of people logging ‘Event Data’ such as shots, passes, tackles, etc. The other type of data is ‘Tracking Data’, which gives the location of the ball and players at every instance of the game, thus providing the crucial extra layer of contextual information for all the event data.
Some of the big data providers that are known in the football world are wyscout which provides video links, statistics about the players, and tools. Other known data providers are Opta, statsbomb, and instat. There are also third-party data providers that buy data and provide methods and analysis.
xG (eXpected Goals) Model
One of the new methods is xG, an abbreviation for eXpected Goals has been booming! It has been used to help with the decision-making process. It looks at the expected goals and uses probability models to predict the probability of a shot being scored by a team and then compares one team’s xG against another.
Below is an xG timeline visualisation of an expected goals (xG) model output example. It shows a game between Real Madrid vs Sheriff Tiraspol. It shows how Real Madrid dominated the game by a huge margin with a greater xG, stating that they had a greater probability of winning; which was a major shock result earlier this season. However, this failed to convert those chances and Real Madrid lost the game. The xG graph shows a different picture of the match than just seeing the 2–1 final scoreline.
Pitch Control
One of the most recent research surrounding football analytics is creating predictive models which explore the probability of each team controlling the ball at every point on the pitch at every instance of a game. This type of analysis is used in the performance aspect of a team’s defensive shape or development of attacking strategies to create space for shooting opportunities.
This is shown in the visualisation below:
However, there is still a lot of randomness in football as it is very fluid. This leads to endless interpretations allowing for new tools to be created as its new grounds. The market is still very competitive, as it is fairly new to the industry making it harder to find the best tool as there are many untouched areas.
Some facts about football and data:
Ian Graham is the physicist who leads Liverpool’s data department. Graham’s first experience of handling football data analysts was in late 2006 when Damien Comolli who at the time was the director of football at Tottenham Hotspur commissioned Decision Technology to work with a variety of one-off projects.
However, it is not only Ian behind the success of Liverpool in the Jurgen Klopp’s era; William Spearman; a Harvard University graduate who worked at the European Organisation for Nuclear Research, Tim Waskett; an astrophysicist who specialises in coding, and Dafydd Steele; a chess champion have all been working together.
