Time to kick off my review of the 2012 MIT Sloan Sports Analytics Conference with the highlight of the conference from a soccer perspective: the Soccer Analytics session. In the space of three years, Soccer Analytics has grown from being shoehorned into an Emerging Analytics session in 2010 to a standalone session in 2012, hosted in the largest meeting space of the conference.
Soccer Analytics panel (L-R): Steven Brown, Drew Carey, Alexi Lalas, Steven Houston, Scott MacLachlan, Marc Stein (moderator).
As with last year, the session was chaired by Marc Stein, who reports on the NBA for ESPN and hosts a soccer show in Dallas. On the panel were Steven Houston (Hamburg SV), Scott MacLachlan (Chelsea FC), Steven Brown (Everton FC), Alexi Lalas (ESPN), and Drew Carey (Seattle Sounders FC).
In my view, there were four subtopics of this year’s Soccer Analytics panel:
- Who is doing analytics?
- What motivates the use of analytics?
- What soccer operations is analytics being used for?
- What are the current challenges of analytics?
To a very large degree, the domestic leagues represented by the panelists capture those parts of the world making heavy use of advanced analytics. English clubs such as Tottenham, Manchester City, Chelsea, and Everton are known to use data actively in their decision-making; in Germany, Hamburg, Borussia Dortmund, and Borussia Mönchengladbach have analytics departments. The New England Revolution recently hired a data analyst, and other clubs in MLS are employing in-house or contracted analysts. The panelists didn’t mention any names, but there are clubs in Spain, Italy, the Netherlands, and almost certainly Japan who use advanced analytics to prepare for matches and evaluate prospective players. The reach of statistical analysis in the sport has expanded significantly over the last 24 months.
So why are clubs embracing analytics in their operations? The answers that the panelists gave centered around two broad answers: culture and economics. Drew Carey made an interesting point that the Seattle area, with data driven companies such as Amazon, Microsoft, and Boeing having such a large presence, encourages a culture of data analysis. I see Carey’s point when I look at the volume of pageviews and unique visits I receive from the Seattle area. The more common answer concerned economics, particularly in a salary-capped league like MLS. Stein missed an opportunity here when he asked whether analytics would be more of an issue if European leagues had a salary cap, when in fact a soft salary cap is coming to European football via Financial Fair Play (FFP). None of the European-based panelists brought up FFP and its impact on the use of analytics in the near and medium-range future, which is unfortunate as I believe that it will force clubs to maximize the value of their expenditures. It would have been beneficial to have heard a discussion on this topic.
The panelists addressed the current uses of advanced analytics in football. Most of the uses concern player scouting, due diligence of prospective players, and performance analysis, but when you reduce the analysis to its basic element, it’s video analysis with tagged data. Steven Brown gave an example of Landon Donovan who prepared for matches by observing tendency data of the players who would mark him. Alexi Lalas said that when he played professionally, he wanted to track the number of possessions he lost in defense to gauge his play. True performance analysis is in its infancy. All of the panelists talked about the importance of context and spatial coordinates in assessing the value of individual plays on the pitch, and the problem of assigning (or calculating) the value of a touch. These issues hit upon the nature of the game and it was here that the panel could have used an analyst who could say something about context-based analytics or the incorporation of spatial information in advanced metrics. At this stage, however, the analysis tools are still quite immature, so perhaps such an analyst would have ended up confusing everyone.
The panelists moved on to the challenges soccer analytics faces toward wider acceptance. It’s not necessarily a generational issue; Sam Allardyce is one of the senior managers in the English Premier League yet was a major advocate of advanced analytics when he was at Bolton. Having someone to champion the idea, which Lalas mentioned, is important, but such a person has to believe that analytics will make their job easier. All too often analytics is viewed in a “scouts vs geeks” scenario, but if the scouts are led to recognize that these advanced metrics could make their jobs easier by filtering the pool of target players, they might embrace the approach as well. The most critical challenges are relevance and credibility. A performance metric that causes an end-user to say “So what?” is a bad metric, as is one that is a beautiful academic exercise but irrelevant to the needs of the technical staff. Scott MacLachlan made a tart comment on the need for credibility (paraphrase): “Would a 19-year-old out of Uni tell someone like Harry Redknapp that his formation is wrong?” You could probably replace the 19-year-old fresh out of Uni with the 30-something with advanced technical degrees with no coaching badges, and both would draw some strong industrial language in response. I think the answer to that statement is to show some humility about the efficacy of statistical analysis in sport, and employ small “buy-ins” in results that build on themselves. Coaching badges would be very beneficial as well, or at least including coaching opinion from the start.
I came into the Soccer Analytics panel with some trepidation about the makeup of the panelists, but those feeling subsided after listening to Drew Carey in the Franchises in Transition panel earlier in the morning. All of the panelists represented various segments of the soccer ecosystem and gave comments that reflected their experiences in those areas which proved to be quite valuable. The soccer analytics landscape looks very different now than it did 12 months ago, and it sure to change significantly between now and next year.