The touchline pretenders among us aren’t the only ones scouring Football Manager 2017’s database for the perfect signing. From the big dogs in the Champions League, to the relative pups of the USA’s Major League Soccer, football clubs across the planet tap into the same database to help them make crucial decisions about the players that could take them to the next level.
“We have some clubs that will take every single aspect of data,” Sports Interactive’s head of business development, Tom Markham, tells me. “One of the Champions League clubs we are working with will evaluate the composition of teams. They’ll look back at older data and see what is the actual makeup—I’m talking nationalities, I’m talking characters—and they will use the attributes in terms of their evaluation.
“There’s another team in the Premier League that uses our data, and all they wanted was height and weight data. We’ve got clubs that will come to us and they’re looking for player earnings and contract expiry dates. We’ve had Ray Houghton come in when he was working for the FAI, trying to find anyone who had an Irish passport who could potentially play for Ireland.”
It’s somewhat surreal that data collected to support a simulation has ended up influencing the reality it’s based on, but it shouldn’t be that surprising. After all, the ‘big data’ bandwagon has rolled through almost every industry over the last decade or so, making it one of the most valuable commodities on the planet, as apparent in the rise of data-hoarding juggernauts such as Google and Facebook. It is only natural that it should come to football too.
Data analytics departments within clubs have grown exponentially in recent years and data-driven terms such as xG (expected goals) are starting to weave their way into the lexicon of the average football fan. The value of data remains a hotly debated topic in the sport, however, and even advocates like Markham think there are dangers in the data revolution. He gives the example of players who have told him that their decision-making process has changed because their clubs are too focused on evaluating passing accuracy.
“Players that might have played, in the past, a through ball that might only have a 25% chance of coming off, but will result in a goal, now will just play a lateral pass because they know that their stats are being evaluated in terms of them giving the ball away. They’re not doing what naturally they should be doing and it could result in real problems in terms of their development.”
Caution is warranted when it comes to how you collect and evaluate data, then, but few would argue against it having a role to play. You only need look to MLS in the US—where Football Manager has been embraced by a number of clubs—to see how just how valuable it can be. “They’ve got very rigid budgets,” Markham points out. “You have your marquee players, but the difference between being successful or unsuccessful is the players you put around those. They’re shopping in bargain basement markets—the likes of Central America, Costa Rica, Honduras—where there is talent, but they’re very good value.“
The problem is that information on these players is very thin on the ground. Directors of football at these MLS clubs, they get a DVD from an agent and, let’s face it, you can make anyone look amazing in a five-minute edit. So they’ve been coming to us; we have independent evaluations of all of these players, but we also have our scouts on the ground in these different locations.”
To be more precise, Sports Interactive has an average of 1,300 scouts around the world compiling the data for the game. “To put that in perspective,” Markham says, “Manchester City, or the City Football Group as they’re now known, with their four clubs and sharing intelligence with China, they’ve got 40 scouts globally. That really shows the depth of the network.”
This network of scouts—mostly composed of professional scouts, but it includes fans who do it for fun—operates in 51 countries, collecting data on players that’s fed back into a database with information on over 650,000 individuals.
“When you’re playing the game there’s around 40 attributes that are visible, where you can see how a player compares to another in terms of 0 to 20,” Markham says. “Behind that there’s probably another 230 attributes that aren’t visible. They’ll evaluate even the mental state of a player; a lot of the stuff that’s hidden, if we had it on display we could have a legal case on our hands.”
Many of these stats, then, have an element of subjectivity to them. That doesn’t mean they’re not valuable—as Markham points out, when a player spends a couple of years on the bench, subjective evaluation of their attributes is all you’ve got to go on. It does, however, highlight how data doesn’t stop you from having to make a judgement call at some point. Data, and by extension Football Manager, does and will continue to provide clubs with valuable insights, but it will never eliminate the capacity for the sport and its players to surprise us.