Throughout the sports analytics boom, those involved have had very few complaints. A new way of looking at the game, analytics has given coaching staff and medical staff alike the tools to make informed decisions based on empirical data rather than truisms or gut feeling. Data in sports can be harnessed to prevent injuries, discover otherwise hidden talent, and guide a coach’s hand.
Despite its success and widespread usage, though, data’s role in sports has been far from perfected. The sheer volume of data being collected on a daily basis by analytics teams has, in many cases, been overwhelming. As a result, the most common complaint from professionals within the industry has been the difficulty in making sense of the vast swathes of data at their disposal. Developments in data-collecting technology have meant that new metrics are being recorded regularly - clearly, the industry needs to find efficient ways of analyzing what is an incredible number of data points.
Enter artificial intelligence. Much like in other industries, a wave of AI-driven software is looking to solve the issue of data overload, with a number of companies rising to prominence with solutions for different areas of sports analytics.
A perfect example of a technology designed to solve a specific data question is Disney Research’s paper entitled Data-Driven Ghosting using Deep Imitation Learning, which focuses on soccer defending. By analyzing season-wide positional data from individual players on the field (provided by STATS), the system can predict where the ‘average’ player may move in any given situation. It can also predict where the very best players would have gone in any given situation.
This data, when overlaid on positional data from a team’s actual players, can save coaches hours explaining where their defenders should have been in any given situation, and allows them to simply show them instead. The system analyzes some 2,200 data points per 10 seconds of play, a scale unfathomable without the use of advanced artificial intelligence. The technology can also be used to predict the probability of a goal being scored against any particular defense from any particular play, a tool most coaches would be unable to say no to.
Another company looking to demystify the sports analytics is Dutch-based DashTag. The company’s latest product, Sports Bot, enables soccer players from any level to engage with their personal statistics in a refreshingly inventive way. The wearable Dash collects the player’s data throughout a given game or training session and, when the session is over, it messages the user personally on their favorite messaging platform.
There is an element of gamification to Sports Bot, too. Users are automatically compared to their peers, their own training history, and against weekly challenges. The bot is, of course, AI-driven, and it gets more intelligent every time it engages with the user, bringing a conversational tone (it even uses emojis) to an industry that can be overwhelmingly numerical. The gamification also extends to the user’s stats, which are given in terms of games like Fifa. The next wave of sports analytics programs, most of which will be driven by AI in some way, will be all about simplification. The data is there, it’s the ability to make sense of it both on a professional and individual level that needs work.