Here’s where football analytics is at in 2023: with enough money, time and brainpower, clubs can not only know the type and outcome of every touch in leagues around the world, they can also buy off-ball data for most of those leagues to see how teams and players move, how they interrupt defensive lines or find pockets of space.
Advertisement
Models that go way beyond expected goals can calculate goal probabilities in any situation on the pitch and assess which players are making dangerous runs or picking smart passing options. There’s been a lot of research done on using this data to predict how a player will perform in different circumstances — the billion-dollar question at the heart of the transfer business.
Here’s how much all of that is worth if it doesn’t impact decisions: exactly as much as doing no data analytics at all.
Sarah Rudd ought to know. Until recently she was head of analytics at Arsenal, one of the handful of clubs in the world that has invested heavily in data. Rudd joined the club in 2012 when Arsenal acquired an analytics company she worked for called StatDNA, and by the time she left in 2021, the club was doing cutting-edge work with off-ball tracking data. But you might not have known it from the way Arsenal muddled through long stretches of the post-Wenger years, showing little evidence of whatever insights they were getting from their analytics investment.
On a recent Tuesday in Texas, far away from north London, Rudd huddled with a group of people around a small black conference table, trying to do football data differently. Their approach would be faster, leaner, and more efficient than what they’d done before — trying to do Moneyball, more or less, on analytics itself.
Rudd sat next to her husband, Ravi Ramineni, a fellow analytics pioneer and former Microsoft employee who spent a decade working for the Seattle Sounders. Across the table was Cole Grossman (below), who played in midfield in Major League Soccer and the Norwegian top flight before earning an MBA. Together, the three founded src ftbl (pronounced “source football”), a consulting start-up that aims to help clubs not just gather data but actually use it.
The conference room belonged to the Houston Dynamo, a club that until recently spent years gathering dust at the bottom of the MLS table. Houston had none of the expensive data infrastructure of a Premier League behemoth like Arsenal. Technically, they still don’t have a single employee dedicated to football analytics, but what they have in src ftbl is access to some of the field’s leading practitioners and a plan to make the most of them.
Advertisement
“Asher said, ‘I don’t want to be the best analytics department in MLS,’” Rudd explained, “‘I want to make the best decisions.’”
Asher Mendelsohn is Houston’s technical director, a calm, bearded 42-year-old Ivy League graduate hired last year by new ownership to reshape the club’s football operations. Like just about anyone who has launched a career in sports in the past 20 years, he believes data can help a team be more efficient. That may be especially true at a mid-tier football club, where the list of potentially useful signings from leagues around the world is way too long to scout.
Mendelsohn has also learned, though, that simply having data isn’t enough. “When we did a tour of the Premier League, we went and visited some of the biggest clubs in the world. We would talk to the people doing their team performance reviews and the scouting departments,” he said. “And then we would talk to the decision-makers and there was a huge disconnect.”
Last winter, as Mendelsohn and general manager Pat Onstad geared up for their first offseason transfer window with Houston, they had a lot of decisions to make — the club would overhaul its whole squad in a few months, bringing in 16 new players — and no analytics operation to help them. “We were trying to get a scouting department up and running, basically from the grassroots,” Onstad said. If they were going to try to be systematic about it, they needed data to narrow things down in a hurry.
Most clubs dream of building an analytics department that will give them a secretive edge over their rivals, but hiring the right people and getting going can take years. One popular shortcut is to sign up for a scouting platform that offers off-the-shelf metrics and data visualisations to help clubs analyse opponents and transfer targets. That’s easier than building out the basics yourself, but a browser tab can’t talk to coaches and scouts and structure how a club makes decisions.
Advertisement
“The question is, can we partner with somebody who could look at a dataset and add more value than just regurgitating what a data provider is giving us?” Mendelsohn said. “What value are they going to add to the analysis?”
Consulting firms like src ftbl offer another way for clubs to approach data, one that’s becoming more popular as the first generation of football analytics veterans leave club front offices for broader roles. Luke Bornn and Javier Fernandez worked for Roma and Barcelona, respectively, and collaborated on some of the best academic football research; now they work for Zelus Analytics, consulting across sports and working with European football clubs owned by RedBird Capital. Ian Graham and Michael Edwards, who oversaw Liverpool’s vaunted analytics department, recently announced a new consultancy called Ludonautics.
These start-ups face some of the same challenges as club analytics departments — hiring people and developing tools to work with a growing number of data sources still takes time — but they have the experience to guide them. “Ravi and I left behind 10 years of intellectual property at our clubs,” Rudd said. “We need to catch up as quickly as possible.”
When they took on the Dynamo as their first client, Rudd and Ramineni knew they didn’t have time to build all the models and validate all the metrics they’d like to have before the club started signing players. They had to work fast and focus on what would matter most to Onstad and Mendelsohn’s decisions.
Houston knew they wanted a winger-striker hybrid who could inject some pace into their attack and stretch opposing back lines. That’s a common profile that could have sent the search in any number of directions, but src ftbl had done enough preliminary league-style analysis to have some idea of where they might want to look.
“A lot of people think MLS is a very physical and athletic league. We have evidence that it’s not that physical or athletic compared to a lot of leagues in Europe. There’s a lot of space in MLS games,” Ramineni said. “So we’re trying to understand the intrinsic qualities of the league, what type of football, things like time and space.”
The Croatian top flight was one of the leagues Ramineni and Rudd believed would translate well, and they had event data to populate player radars with basic metrics and start sorting prospects from the league. But without expensive tracking data, they couldn’t directly measure the sorts of off-ball runs that Houston wanted from their new attacker. Instead, they used events such as high-value receptions, through-ball targets and cutbacks from an area Ramineni calls the optimal assist zone as proxies for effective off-ball movement.
Advertisement
“We’re always going to have clients that don’t have tracking data, that just can’t afford it or choose to spend their money elsewhere. So we started building this framework in terms of, ‘What are the core competencies — the skills or attributes — that a player needs regardless of their position?’” Rudd said. “We want to make sure that, regardless of the dataset, we have a way to measure those. Maybe this dataset is an imperfect way of measuring it, but maybe that means that team needs to do some extra video scouting in that area.”
When Houston’s scouts looked at a young Croatian league attacker named Ibrahim Aliyu who looked promising in src ftbl’s data, they liked what they saw. But when Onstad watched a handful of clips, he was less impressed at first.
A hesitant reaction from the sporting director might be enough to get a player tossed out of a fast-moving search at a lot of other clubs. “But what happened is they said, ‘Listen, data says this is right. Scouts like him — this is why. Take a deeper look,’” Onstad remembered. He went back to watch more of Aliyu’s games.
“You’re like, ‘Oh, I see it! Now I see what you guys are talking about,’” he said.
Ibrahim Aliyu (right) (Rick Ulreich/Icon Sportswire via Getty Images)At Arsenal and Seattle, Rudd and Ramineni had worked for years to get that kind of buy-in. Now they had the client’s attention and could use their experience to influence decisions, the holy grail of analytics operations everywhere.
Houston started the season slowly, but as they integrated their new players into the team, performances began to turn around. In his first six months in MLS, Aliyu has so far played mostly at striker instead of on the wing, but even in the middle his off-ball movement has made a difference.
“It’s dramatically changed our ability to generate chances,” Mendelsohn said. “And through things that are very hard to measure. Even just the threat of space in behind and being able to exploit that changes the way teams have to play against us, which creates more space for our attacking mids.”
Advertisement
Studying complex interactions between players is the sort of thing Rudd worked on at Arsenal and, as src ftbl continues to add data scientists, they may one day catch up. The list of big technical projects already on their plate, such as developing off-ball metrics from tracking data and modelling how a player’s performance will translate to a new league or team, is daunting enough already. But in conversation around the conference table, their curiosity was already pushing them toward even more distant horizons.
“In an ideal world, our metrics are good enough that hip fluidity and ball striking technique and arm position and things like that, if they matter, they’re showing up in our metrics,” the resident ex-pro Grossman said. “I guess the reason it’s interesting for me is if we don’t need scouts to look at event data, what are things they could look at? If there are holes in our metrics — maybe a scout is unbelievable at noticing hip fluidity in centre midfielders, which might be predictive of ability to play in tight spaces.”
Because they spend so much time checking their own data and models against video, src ftbl can resemble a scouting outfit themselves. “If you’re going to build a model and you haven’t looked at it on video, you cannot present that to anybody inside a club,” Rudd said, “because, you know, they’re just not going to believe it.”
The goal is always to get buy-in, that magical pixie dust of belief that transforms data into decisions. Former Seattle Sounders sporting director Garth Lagerwey watched Ramineni work for years to get the whole club invested in analytics. “If you’re using data across every platform within the club, that means you have to have relationships across the club with the scouts, the coaching staff, the other analysts, the sports scientists, with the medical team,” he said.
When he moved to Atlanta United ahead of this season, Lagerwey hired src ftbl to try to replicate that culture at his new club. “Even with a really established group in Seattle, it took five or six years. We need to fast-track that here in Atlanta,” he said. “That’s what we’re trying to do with a consulting role.”
Eventually, clients may still aim to take their analytics in-house, hoping for that private edge. “What we always ask ourselves is: alright, what are the advantages we can create that are hard for other teams to copy?” Mendelsohn explained. “It’s a lot faster to bring a group in that’s already done this than just building from scratch. But there’s nothing to say that one of the bigger clubs out there, whether it’s in MLS or somewhere else, doesn’t come in and hire them for way more money.”
In recent months, src ftbl have started working with clubs in England and Spain. They plan to mentor club employees to take over for them in Houston and Atlanta as they someday move on to new clients. But unlike when they left their clubs, this time their tools and models will come with them, allowing them to build on past work as they lay the groundwork for a better understanding of the game.
Advertisement
“Without sounding romantic or anything like that, that is what drives us,” Rudd said. “I think we want to figure out football, and working with clubs gives us an opportunity to do so much more effectively.”
(Top picture: Sarah Rudd and Ravi Ramineni with Santi Cazorla during his time at Arsenal. Credit: Sarah Rudd)
ncG1vNJzZmismJqutbTLnquim16YvK57k3Ftbm9janxzfJFsZmlxX2aEcK3RrJynmZxiwLOvjJ%2Brm6RdlruiuNisoKxn