Artificial intelligence is now everywhere, even in the game of soccer: here's how DeepMind (Google) and Liverpool are collaborating to create the coach's virtual assistant.
Technology is bursting into the world of soccer. Whether it will contribute to evolve it in a positive or negative way is still too early to say, but first VAR, then algorithms and artificial intelligence have ended up in a few years among the words of common use associated with the world of football.
And the next step is already ready, even if, weighing the companies involved, we could even talk about two steps in one. Liverpool, then the 2019/2020 Premier League holding team, and DeepMind have joined efforts to explore the benefits that artificial intelligence applied to soccer can give. Having never heard DeepMind before is understandable, but just knowing that it is a subsidiary of Alphabet - i.e. the holding company headed by Google - is enough to be almost certain to hear this name again in the future. The reason why Liverpool was chosen is curious: DeepMind's founder Demis Hassabis, who participated in the project, is a big Reds fan, so it's the first club asked to collaborate.
The behind-the-scenes analysis that would change football
So Liverpool provided every piece of information collected in each of the matches played in the Premier League between 2017 and 2019. An incredible amount of data when you consider that in today's soccer, several tools are used to monitor as many aspects as possible, including GPS trackers and sensors of various kinds, some of which are embedded inside the ball.
From artificial intelligence applied to soccer, we do not expect the same results obtained in 2016 when DeepMind's AlphaGo software defeated one of the strongest players in the world in the game of go, if only because the game of soccer contains an inherent uncertainty. No desire to replace coaches even in the long run, only to develop systems that examine the greatest number of variables - and in soccer, as we know, there are many - and assist the role of men, who seem irreplaceable. At least at the moment.
Artificial intelligence (AI) could, for example, be used to predict the consequences of a tactical variation or how the opponent's game might vary with the absence of a key man. Or even provide more precise information about penalty kicks: researchers at DeepMind analyzed more than 12,000 shots from the penalty spot awarded across Europe in recent seasons, uncovering habits related to role on the field - for example, forwards shoot more often in the lower left corner than midfielders - or strong foot.
In addition, and perhaps most interesting for coaches and team managers, an AI-based model that analyzes certain parameters on players for strength and fitness could give better indications than those inferred by a human, and thus advise team managers on a rest period for certain players in order to reduce the risk of injury.
Developments in the medium term, five years or so
Do not expect revolutions in the short term, warns the DeepMind manager who is following the project closely. "I doubt that the fruits of what we are doing will come in six months or a year, but in about five years some of the tools we are developing will have reached a good level of maturity and could be ready to advise coaches on key mistakes made during a competition or variations that could improve the game."