Earthquakes partner with German FA to bring machine learning and artificial intelligence to soccer
Friday morning, the San Jose Earthquakes announced an agreement with the German FA, the world’s most successful footballing nation in recent years. Germany won the world cup in 2014, and also won the U21 European Championship and the Confederations Cup in 2017.
The agreement entails “a multi-year collaboration focused on knowledge exchange, game development and machine learning,” according to the Earthquakes press release.
Markus Weise, Head of the DFB Academy’s conceptual development unit, states that the collaboration “is a wonderful opportunity for the DFB to develop training content and diagnostics further.” Weise continues, “capitalizing on game insight through machine learning and knowledge exchange are great examples of how we want to drive developments proactively forward through co-innovation – and this collaboration with San Jose allows us to do just that.”
There is a lot to unpack from this relatively vague release, mostly because there has not been a lot of progress in the soccer world with regards to machine learning. The Verge wrote an article on an interesting startup called Stratagem that is using artificial intelligence to bet on soccer matches, and there are a number of similar fringe companies, but there is still plenty of work to be done with machine learning in the soccer world especially. Machine learning applies artificial intelligence to analyze big data. A good definition for artificial intelligence is “the scientific understanding of the mechanisms underlying thought and intelligent behavior and their embodiment in machines.” (AAAI) The goal of using machine learning would be to reason about the situations, make predictions or decisions on what to do next, and learn from experience.
Soccer can be broken up into a number of decisions or events, and from these moments in the span of a game, different forms of statistics can be gathered. Companies like Opta have mastered collecting all different kinds of data from matches. The current big question in the soccer analytics realm is finding out what to do with all the data.
Jesse Fioranelli mentions in the press release that one of the goals of the partnership is to “deliver an insight-driven message to players and coaches.” The Earthquakes and the German FA are most likely looking to do is to use this machine learning to make adjustments to training, create game plans, develop scouting reports, and even make in game changes based on data. For the Earthquakes, the German FA is an ideal partner because they are arguably the most successful FA today. For Germany, the Earthquakes are an ideal partner because they have undergone a complete “forward-thinking” transformation in the front office, and they are located in the worldwide hub for computer science and innovation.
Ultimately, it is impossible to know for sure what will come of this partnership, but it certainly does show a level of initiative that other clubs in MLS have been hesitant to take. Former LA Galaxy head coach and current USMNT head coach Bruce Arena said “analytics in soccer, if no one here has figured it out, doesn’t mean a whole lot” in 2016. Colorado’s Pablo Masteroni recently went on an analytics rant stating “stats will lose to the human spirit every day.” On the other hand, clubs like Toronto FC and Seattle Sounders FC have made investments in data analytics and sports science regarding the physical performances of athletes.
In the past, the Earthquakes would certainly be placed on the Bruce Arena or Pablo Masteroni side of the analytics spectrum, but that is not the case under new GM Jesse Fioranelli. This January, the Earthquakes also partnered with Recentive Analytics, who are a machine learning company that essentially study data to help Earthquakes’ sales, marketing, and services teams maximize and optimize ticket sales.
Overall, this is yet another exciting step for the Earthquakes, which they seem to be making every other week at this point.