In February, Matthew Benham, the owner of the English club Brentford, did an astonishing thing. Even though Brentford had earned promotion to the second-tier Championship just last season, and even though the club was in fifth place in the league, Benham effectively fired the manager (Mark Warburton), assistant manager (David Weir) and sporting director (Frank McParland), announcing their contracts would not be renewed at season’s end.
Why on earth would he do a thing like that? The answer, Benham argues, lay in the data. His data. Benham, you see, is not your typical sports owner. Over the years, Benham made a fortune betting on soccer, more than enough to buy Brentford (his childhood team) in 2012 and a Danish club called Midtjylland last year. His betting success was based on a mathematical model that he believes is far more reflective of a team’s strength than the league table itself.
“The table is notoriously inaccurate because there’s just way more randomness than people understand,” Benham explains. “One of the best examples is over halfway through the season in Germany, Dortmund were bottom of the league. We actually still made them the second-best team in Germany. They were just horrifically unlucky. Usually the way the fans and especially the media look at it, randomness is quite an unsatisfactory explanation, so they like to look for stories and a narrative. But very often randomness is the main explanation.”
When Brentford was in fifth place in December, Benham said, his mathematical model showed the club was actually the 11th-best team in the Championship. What’s more, Warburton, Weir and McParland were chafing under some of Benham’s data-driven preferences. And so, even if Brentford is able to turn around its Championship playoff semifinal against Middlesbrough on Friday (‘Boro leads 2-1 after the first leg) and eventually earn a remarkable promotion to the Premier League, Warburton, Weir and McParland will be looking for new jobs this summer.
[Editor's note: Middlesbrough wound up winning the second leg 3-0, eliminating Brentford from promotion contention]
“One lesson I’ve learned is if you go to a manager and say you’ve been unlucky and you actually deserve to be higher, that’s actually quite an easy message to give to a manager, naturally enough,” Benham says. “But if you say to a manager, ‘You’ve been lucky, you deserve to be lower,’ that’s actually an incredibly difficult message for the manager to swallow.”
(By contrast, at the end of this season Benham notes the model rated Brentford fifth-best in the league, indicating the team wasn’t lucky all season long, but just in the first half.)
For years, as the Analytics Revolution has come to other sports like baseball and basketball, we’ve been waiting for it to come to soccer. And sure enough, all sorts of companies like Opta and ProZone have sprung up to gather more data than you could ever imagine about the world’s game. Meanwhile, plenty of clubs have set up their own analytics departments to try and identify advantages that their competitors might not see.
But do you want to know the truth? Whenever I ask soccer analytics people which clubs are really using all that data in a big way, the answer is direct: Not many, if any. Managers just don’t want to give up control to the numbers. They don’t trust them.
Benham’s teams are different. As Rasmus Ankersen, his right-hand man and the chairman at Midtjylland, puts it: “The great thing about this project is that the innovation and use of data doesn’t come from an analyst who’s hidden away in the analyst’s room in the basement. It comes from the owner. So everything we do has the flavor of analytics.
“You have to give it a role and you have to believe in it. And Matthew’s success in betting is obviously built on overriding gut feeling in each region and overcoming biases by using data. So he is really confident that if the data is not perfect, it is less imperfect than human eyes and the human ability to make judgments.”
It’s still too early to know if Benham’s grand experiment will work. But the early returns are rather mind-blowing. Midtjylland, which has gone in heavily on Benham’s math model, is about to win its first Danish league title and currently leads that country’s top-flight by nine points. (Midtjylland has never won any kind of trophy before.) And Brentford, which has barely scratched the surface of using Benham’s model, seems destined to climb to the Premier League—if not this season, then soon.
Starting with a new head coach and sporting director this summer, Brentford is set to adopt Benham’s model more fully in the same way that Midtjylland has done. And Ankersen, an energetic Dane, will take on dual roles with both clubs. The goal is to do nothing less than change the way we think about soccer.
“I just think it’s a fascinating project to try to change an industry and try to explain the inefficiencies of an industry with a smaller budget than the competition,” says Ankersen. “That’s what we’re trying to do. Even for Midtjylland, we’d like to push for being at the top of the Danish league against next year, and our ambition is to play in European group play, in Champions League or Europa League, every second year. That’s the next step we want to take.”
For Brentford, their goals are also high.
“We want to get into the Premier League and see how we can go from there,” says Benham. “The interesting thing about the television deal in the Premier League is to some extent it levels the playing field. Because even the sort of bottom club gets so much television money. In most other top leagues the size of the club determines how aggressive they can be, so if you’re a big-name club you get lots of commercial revenue and so on. I think the television deal to an extent levels the playing field. So if we get into the Premier League, let’s see how far we can go.”
What exactly has Midtjylland done this season using analytics?
“We look into all possibilities in football, because there are a lot of inefficiencies,” says Ankersen. “One of the inefficiencies is the transfer market, which we’re trying to exploit by using data in a more intelligent way.”
One player the club identified last year was Tim Sparv, a little-known midfielder, largely because he did well for Greuther Fürth, a second-tier German team that was highly rated by Benham’s model. (The decision to pursue Sparv is explained in this terrific piece by De Correspondent’s Michiel De Hoog).
Another area of inefficiency in soccer is set pieces, says Ankersen, who has made sure the club has regular set-piece meetings between himself, players, coaches and outside consultants. Midtjylland has scored nearly a goal a game from set-pieces this season, an insanely high strike rate.
Ankersen says what Midtjylland is doing with set pieces isn’t rocket science, citing players who are fearless in the box and can deliver accurate balls, as well as a coaching staff that has shown a willingness to improve that part of the game.
What else does Ankersen want his team focusing on?
“We believe a lot in improving the professionalism of players,” he says. “We’ve never understood why a football player shouldn’t be as professional as an individual athlete. An individual athlete optimizes every single part of his game. Nutrition, everything in small details. But football players don’t do that. Why not? We think there’s a big opportunity there.”
Ankersen is quick to add that he thinks it’s still far too early to say if their project will be a massive success, though the early returns make them optimistic about what comes next.
Benham, for his part, had grown frustrated with what he describes with disdain as “the English way” of doing things in soccer.
“If we go back to the Brentford situation, there’s obviously a huge amount of criticism from the English media that I wanted to change the manager and in fact get rid of the role of manager and have a head coach instead,” he says.
“The way the English media were portraying it, they were saying, ‘It’s wrong, the owner wants to be the all-powerful dictator, when it should be the manager who’s the all-powerful dictator.’ And the thing is there shouldn’t be anyone who’s the all-powerful dictator. Really, the idea is just to get lots and lots of bright people involved, have lots of debates and interactions and just always look at how can we improve? How can we be different?
“English football is just dominated by this idea that we have to do what everyone else does. One guy I particularly admire in football is the new Dortmund manager, Thomas Tuchel. I’ve been fortunate enough to meet him a couple times recently. He was talking about the time they won a game and he made six changes for the following game just because he thought that was the right team for the next game. And the media were like, ‘But you can’t do that! You’re not allowed to change the winning team!’
“It’s not actually a law, but in English football people seem to act about a lot of these things as if it’s a law. It’s having that attitude of just because everyone says this is the way to do it, it doesn’t mean it’s the correct way to do it.”
Now the challenge for Brentford and Midtjylland is to take the next steps. For Midtjylland, this summer’s transfer window is a major test of the mathematical model when it comes to identifying players. (The club is selling three or four of its current players to wealthier clubs after this season’s success.) Meanwhile, Benham will hope to make the right hire of Brentford’s new head coach—not a manager, he says, but a head coach.
The project is one of the most fascinating experiments in sports right now. But it’s only beginning.
“Obviously, the next couple years we’ll find out if we’re right or not,” says Benham. “I’ve obviously had an enormous amount of success in sports betting, but that doesn’t guarantee anything whatsoever that I’ll be successful in sports clubs. I think there are a lot of people who are going to watch the project in the next couple years, and a lot of people hope we’ll succeed, and lot of people hope we’ll fail. So it’ll be pretty interesting.”