Inside the Data Revolution That’s Transforming Premier League Football
The analytics revolution has transformed the U.S. sports world over the past couple decades. Baseball felt the impact first. Michael Lewis’s blockbuster book Moneyball—which came out in 2003 and examined the success of the small-budget Oakland A’s—popularised the idea that teams could find hidden advantages in the numbers. The data-driven approach to building teams and coaching players has spread beyond the baseball diamond. In basketball, the three-pointer reigns supreme thanks in part to number crunching about the relative value of where shots are taken. American Football coaches, a notoriously risk-averse species, regularly go for it on fourth down these days because the numbers tell them it’s worth the gamble.
By contrast, the most global game has lagged a bit behind. Football has been slower to embrace a Moneyball approach. The difficulties in quantifying such a free-flowing sport combined with the inertia of tradition mean that analytics has taken longer to make an impact in Europe’s powerhouse professional leagues.
One of the first major football clubs to join the analytics movement was Liverpool of the Premier League. In 2010, the team was bought by the Fenway Sports Group, the Boston-based owner of baseball’s Red Sox, led by commodities trader–turned–sports magnate John Henry. With the Red Sox, Henry had hired Bill James—the father of sabermetrics, or the application of analytics to sports—as a consultant and delivered World Series titles. At Liverpool, he wanted to apply a similar data-driven approach.
In 2012, the team hired Ian Graham to lead the effort as director of research. Armed with a Ph.D. in physics from Cambridge, Graham had been working at a consulting firm and doing analytics research for another Premier League team, Tottenham, as well as for bookmakers. He also happened to be a big Liverpool fan. At Liverpool, he launched the first in-house analytics department in the Premier League and was a key part of the management team that led to Premier League and Champions League titles under manager Jürgen Klopp in recent years. Graham left in 2023 to launch an advisory business, Ludonautics, offering the same type of statistical assessments and predictions he oversaw at Liverpool.
Recently, he published How to Win the Premier League: The Inside Story of Football’s Data Revolution. In the book, he shares behind-the-scenes stories about how Liverpool applied his research—including some of the battles with stubborn managers—and explains how the use of analytics has evolved in big-league soccer more broadly. It’s a fascinating read if you’re a sports-plus-data enthusiast (and even more so if you’re a Liverpool fan like, full disclosure, I am). I caught up with Graham to go even deeper on the analytics, as well as to get his pick for the most valuable player in the Premier League and his prediction on whether Manchester City will be the repeat champion for a fifth straight season. The conversation has been edited for length and clarity:
ESQUIRE: There is a mystique to the proprietary edge that analytics can provide for a big football club like Liverpool or for you as a consultant. Why’d you decide to write the book and shed light on how it works?
IAN GRAHAM: Yeah, that’s a good question. I think there’s two reasons, really. It’s very difficult when you work in a club, because you can’t tell the press what you’re doing. There’s a clause in my contract that says you don’t talk to the press unless we explicitly say it’s okay. So having left the club, it was an opportunity to hopefully give an insight for fans into—you know, I’m a Liverpool fan as well—so when fans like me say, “What the hell are they doing? Why didn’t they sign this player?” to try and give, not to give an excuse but to show what the thought processes are and why maybe sometimes not signing a player is the right decision, instead of signing the wrong player.
And then the second reason was, in baseball, right, Moneyball is twenty years old and analytics in baseball has progressed since then. Everyone kind of knows the idea. It’s gone into the mainstream, for better or for worse. Soccer is at least ten years behind baseball, probably more like twenty years behind baseball, and the analysis is just much harder to do. Baseball is a simpler sport to put into statistics and do analysis. So there’s a lot of jargon and, in my opinion, bad explanations about how it all works. There’s also a mystique of Oh, these number crunchers. Can you imagine the data at their disposal? The AI they’re using? And I really wanted to explain to a fan who might not be particularly interested in data analysis that all of the analysis is understandable if you’re just an intelligent fan. If you know the game, good analysis should speak to how the game works. So I really wanted to sort of take that layer of mystery and that layer of jargon off the data-analysis story.
How much would you say that analytics is really being applied in the Premier League now versus five years ago or 2012, when you started with Liverpool?
In 2012, no one was doing it, certainly not in-house and certainly not seriously. Fast-forward to today and most big clubs now have a data department. The differences is in who uses it well and who doesn’t use it well. Arsenal use theirs well. I think Manchester City have got a really good data department. I’m not sure how much they affect Manchester City, but they certainly affect the other City group clubs.
And then the other two big clubs that I talk about in the book, which, again, is instructive—because they weren’t in the Premier League in 2012; they were in the third division—are Brighton and Brentford. And you know, the reason they use data analysis is their owners are sports bettors, and they insist on things running that way. And really, you need smart people. You need to make sure the models that you’re developing to understand the qualities of players are good models and they’re predictive and so on. But the implementation is the difference maker. And the three clubs—Liverpool, Brentford, and Brighton—that really embraced it, it’s because the guy that owns the club says, “You have to do things this way.”
In the U.S., we’ve seen how analytics has not only changed the way that leagues value players but also the way that teams play. For instance, post-Moneyball, there was a big emphasis placed on on-base percentage in baseball. Basketball has really become a game of shooting three-pointers and trying to get dunks. So people are coaching around that very aggressively. In the book, you focus a lot on the value of playing an attacking style in football, and where on the pitch shots are more valuable, but do you see that type of analysis translating into on-field strategy yet to any large degree?
In smaller ways. But the differences are much, much less pronounced than in the American sports that you mentioned. And partly that’s because of the difficulty of analyzing the game compared with other sports. Baseball is a series of one-on-one encounters that you can measure very precisely the impact of. And also you can say, “Was it a good pitch? Was it a good hit? Did the player get lucky?” Because you can look at the trajectory of the pitch and so on. With [soccer], that stuff’s hard to do. You can say something today about the options that a player had. Did he pick the right option, and did he execute well? Is it a decision-making problem that the player has or an execution problem that the player has? I don’t think much has happened on the pitch as far as tactics are concerned yet. I think it’s the next step for football analytics. That’s currently still the domain of the coach, and it’s right that it’s the domain of the coach, because it’s a very difficult problem to solve.
When Michael Lewis wrote The Blind Side, he examined the economic effects of American football becoming mostly a passing game. Therefore the quarterback was the most valuable player, and the value of the left tackle, who protects the quarterback, had skyrocketed as a result. Has analytics in football changed the economic value of position in a similar way? Who’s the most valuable player on the field? And has that changed?
Yes, it has changed, but the rank ordering is still the same. So forwards are the best players and therefore the most valuable players. And it goes back down the pitch. The further back you go on the pitch, and the further to the sides of the pitch, they kind of become less valuable. And if you talk to many professional players, you know, I asked some central defenders and some fullbacks at Liverpool, where did you play as a youth player? Everyone played centre forward because they were the best player on their team until, you know, age sixteen or seventeen, when they went to a big academy. And when you move up a level and you’re no longer the best player, you move backward, or you move to the edge of the pitch, depending on your sort of qualities.
Forwards used to be overvalued, and I think there has been a correction now. But they are still the most valuable players. And the reason is that attacking is a much more individual skill than defending. That poses a problem for data analysis, because when a defensive unit malfunctions, we want to always say, Well, how are we going to credit or penalize the defenders for the chance that happened or the chance was prevented? Teamwork is a big part of it in defence. So it’s hard to say. Whereas if you’ve got Kylian Mbappé or Mo Salah just creating a chance for themselves, you can say this guy created 100 per cent of that chance.
The places where it’s changing are where the sort of final step is quite often not the most important step. In the early days of [soccer] analytics, there just wasn’t that much information available. Shots was the most important piece of data you saw. So players who took a lot of shots were very valuable. And there’s been a correction with that as well, because there’s been a recognition that a player that can only take chances and not create them for themselves is less valuable. Salah is an example of a player who can create chances for himself. He can obviously take chances, but he can create them for himself, and he can create chances for others as well. One-dimensional centre forwards who are just finishers are slowly becoming understood as less valuable, because when you put them in a team that can’t supply them with chances, their impact will be a lot lower. Whereas if you put Salah on a team that can’t supply chances, that’s fine. He won’t be as good on a worse team, but he can create his own chances, and he can create chances for others.
So who’s the most valuable player in the Premier League right now? Erling Haaland of Manchester City seems to score a hat trick every other game. He would be the obvious answer for a lot of people, but he is more like the one-dimensional scoring machine, right?
Haaland is a throwback to “all he does is finish.” But he’s so good at it that he’s in a different category than the other finishers in the Premier League. That type of exceptional talent hasn’t been seen for years, but it’s a very one-dimensional talent. For most players, being that one-dimensional is a problem, but he is so good at finishing. He hardly touches the ball. He just scores. You don’t need to create many chances. If you give him three chances, he’ll likely score a goal. That’s not true of very good centre forwards, but it is true of Haaland.
I think the players who can do both are more valuable. So Kevin De Bruyne—because of his age, he’s not currently the most valuable. But if you look at the impact that he makes, he can do everything. Salah in a different position is a similar sort of player for that reason. I’m not an Arsenal fan, but I love Bukayo Saka for a similar reason. I talk about triple threats. And I see little bits of Salah and Sadio Mané in Saka. You don’t know what he’s going to do. He can chop in and shoot, he can cross, he can run past you. It’s very difficult to defend that sort of player, that sort of multi-dimensionality. Saka’s very young as well. So if you look at what he’s going to bring to Arsenal over the next five years in terms of long-term value, I’d have him as one of the most valuable in the Premier League.
Saka’s agent likes your analysis of his future value. Speaking of Man City, they have massive financial backing from the owners in Abu Dhabi—to a controversial degree. (We are waiting for the results of an independent hearing about 115 alleged breaches of Premier League financial regulations by the club.) They buy great players. They have maybe the greatest manager in the world in Pep Guardiola. But from an analytics point of view, what gives Man City an advantage? Is it just that they have great players at every position?
The defining factor of whether you are going to be a successful team or not is: Do you have great players? Other things can spoil that. If you’ve got a manager who will only play defensive football. you’re unlikely to win the league even if you’ve got the best players. But if you ask: What do you want from a team? Do you want the best coach or the best players or a new style of tactics, or do you want the best medical department so your players are fit? I want the players. And if you give me ten points to spend on the different aspects of the club, I’m going to spend nine of those points on the players, one point on the coach. Maybe eight and a half on players and one point and a half point on everything else that the club does.
That being said, they’ve maximised their success because Pep Guardiola’s approach maximises the chance of winning in each game. It’s a very unusual approach, and I think it’s dependent on having world-class players in each position. And not just world-class players but world-class players of a particular type that can keep hold of possession. Another thing that makes Haaland fascinating is he’s not the sort of player who can hold possession and just keep circulating the ball around. Haaland is so extremely good that I think they’re happy to live with only playing with nine outfield players, and Haaland’s just there to score the goals. But their tactical tweak is they’ve had extremely long chains of possession, much longer than any other team in the league. The idea is we’ll just keep the ball away from you. So it’s a very good defensive strategy to keep the ball all the time, but you need extremely talented players. Keeping hold of possession is a hard thing to do if you’re in the opponent’s half.
I don’t know if they do it consciously or Pep’s just intuitively hit upon the solution, but it’s the right solution, which is they just drag the opposition around. And again, you need extremely talented players to pass and run in tight spaces. They’ll drag the opposition around until there’s a mistake or a little bit of space. And it’s like, okay, after twenty passes, now we’re going to make our chance. So they get extremely high-quality shots.
Chelsea over the past couple years has spent well over $1 billion in the transfer market. Just the sheer number of highly rated players they’ve acquired has been extraordinary. As somebody who’s done this for a living, evaluating players and preparing for transfer windows, what do you make of the owners’ strategy? Is it more coherent that it looks from the outside?
It’s tough to say that it’s coherent, but I can see some advantages and also some disadvantages. And I would guess that they just weighed the advantages more heavily and the disadvantages less heavily than I’d weigh. So on the plus side, the first thing they did was some financial engineering by long long-term contracts. And then the second thing they did was to realise that academy players are pure profit in the transfer market, and Chelsea had a lot of really good academy players. So that was a huge amount of money that could be spent on upgrading the squad. I think you should also look at net spend. So Chelsea's growth spend is incredible, and we’ve never seen anything like it. They have recouped a lot of that money through sales. The net spend is still very, very high but just not quite as extreme as the gross spend looks.
The second part of the plus side is Chelsea have got a very large number of some of the most talented young players in the world. And if those players follow a typical aging curve, in three years’ time they’ll have a large number of the best players in the world. They’ll be twenty-seven instead of twenty-two or twenty-three or wherever they are today. Now, on the downside, you have to manage a squad. You can only play eleven players in a game. And I think that’s the thing that I would put most weight on. That disadvantage outweighs the advantages for me. So players only become world class if they play and if they play relevant minutes. You have to play against the best players if you’re going to become the best player. And so they’ve got lots of players who just aren’t going to play. Or they might be out on loan, but then you’ve got much less control over the minutes they get.
On the other side of the coin is your old team, Liverpool. They spent very little in the summer transfer window, which got a lot of fans upset. The director of football and the manager have said it’s hard to find players that will improve the team. But I have two questions about that. First, you want to make sure you get the right player, but doesn’t that sort of box you out at a certain point from bringing in new players that will refresh the team over time and developing? And second, when you’re projecting the strength of your team, how do you value quality depth? Because it seems like there’s a fine line between saying, “We don’t want to bring in too many players” and undercutting yourself by not having enough depth.
It’s a very fair point to make. If you think about how to maximise your chances, if we know that our total wage bill is going to be lower than Manchester City’s and if we know that our net spend in the transfer market is going to be a fraction of Manchester City’s—and when I was at Liverpool, that was the case—if you ask where Manchester City’s spent money, they spent it on world-class depth in every position. And they’ve done really well to get to that position. I can’t afford world-class depth in every position. So do I want full depth but slightly less than world-class players? Or do I want world-class players and compromise on depth?
I’ve got a better chance of winning the league if I have a world-class set of starters and I compromise on depth. Players who never play, or players where there’s only a 5 percent chance that they’ll play more than five games a season, let’s say, those are the players that we were happy to compromise on in terms of depth. So the third-choice goalkeeper; fifth-choice centre back; maybe even fourth-choice centre back, who often doesn’t play five games in a season.
The other thing we did, which was less of a compromise but it just made the transfer market more difficult and it made the analysis more difficult, was any player that we buy we want that player to be multifunctional. And that means we can have similar depth, but the same player could be depth for fullback and central midfield, or centre back and fullback. James Milner was a free transfer. And as good of a player as James Milner was, his real value is he could play almost any position on the pitch.
From an analytics point of view, how do you factor in injuries and the skill, for lack of a better term, of staying healthy and available when you're evaluating players?
Yeah, injury prediction is very hard, and it’s because the data is not available. It’s personal medical information. If you want to sign a player, you can read news reports about what his injury history is, look at how many games he’s missed, try to find out why he missed those games, but you’re never going to get good-quality data on what his training exposure was. Does he just break down if training is too hard or he hasn’t had the right rhythm of training in leading up to games? The only thing that reliably predicts injury is past-injury history. Thiago was a player with a past-injury history, and we went into that with our eyes open. But Naby Keita was an example of a player that didn’t have a past-injury history.
It’s fundamentally a difficult thing to analyse. We didn’t have enough data to analyse it. We could analyse our own players, but it was the same set of players that was injured all the time. So say we have fifty injuries per season but half of those injuries come from four players. How that played into the squad depth analysis was my simulation of: How many games is this player going to start per season? I would randomly injure players—not in real life but inside the computer. Rather than saying I’m going to predict what Naby Keita’s injury rate is going to be in two years’ time, we’d just run scenarios instead. For a player that we might think has got a risk of being more injured in the future, we can just increase that injury risk and see what happens to the number of starts that he makes. In that case, you might find that an academy midfielder is having to start ten games per season. And that gives you a signal to say, Well, are we happy with that risk?
It’s a similar analysis to what an insurance company would do. Run different scenarios and you ask the decision maker: Are you happy with this risk? And it’s always difficult, because, I don’t know, managers want as big a score as possible, as a general rule. but having a really big squad means you’ve probably not got really, really world-class players in that squad, because you’ve had to spread the money more thinly across the squad.
Back to your old club for a moment. Let me ask you a question on behalf of Liverpool fans. The club doesn’t spend on the same level as Man City, Manchester United, or Chelsea in the transfer market. That frustrates the supporters at times. But as someone who worked for the current owners, do you believe that Liverpool has enough financial support to continue competing for the Premier League title and the Champions League title?
I’ve got a very biased opinion. So when I joined in 2012 and I knew what their approach was, I’d seen what they’d done with the Boston Red Sox. And you know, I believe in the ownership group, and I stayed there for eleven years, so that kind of tells you what I think about the ownership group. They’ve always been happy to invest if the investment has got an impact. I think the difficulty with the transfer window is it’s the thing that generates interest, because everyone likes a shiny new toy. And what is much less publicized is the money that is spent on the team in other ways, specifically in player wages.
If you look at transfer fees—net spend, gross spend, total fee—spent to bring the current set of players to the club, there’s a correlation between that and how well the team does. But it’s a much smaller correlation than the correlation between player wage bill and how the team does, to the point where, on a one-season basis, if you put all that financial information in, the answer is you only need to know about wage bill if you want to know how good the team’s going to be. You don’t need to know about transfer fees. Over longer time periods, transfer fees do start to look important.
Liverpool’s wage bill has been second or third highest in the Premier League over the past few years. I’d like to see Liverpool spend a lot more in the transfer market, but I prefer to see them spend £350,000 a week on Mo Salah. We could say, Okay, Mo, thanks. Off to Saudi you go. That £350,000 per week can be put into the transfer market instead. Liverpool will spend that money better than other clubs will spend that money, but it’s less of a surefire success than just continuing to pay Mo Salah for his proven production. And so, you know, the money is spent, but it’s just spent in a place that isn't sexy for fans.
So as a Liverpool fan, you’d like to see them buy more players?
My friends and my ex-colleagues will kill me for saying it, but yeah, as a fan, I want more players. But that’s because I don’t have to worry about financial efficiency or getting the biggest bang for our bet anymore. I can just be like any other fan and say, “Come on, I want more world-class squad depth!” And I don’t care if you’re going to fall afoul of [profitability and sustainability rules] or whatever.
Okay, so what does the data say? Who’s going to win the Premier League this year? Is Man City going to win it again? Is Arsenal going to take it this year? Your prediction.
Man City. It’s not a done deal. I think the bookmakers might have priced in some non-football reasons for Manchester City not winning the league. If it’s down to football, it’d be similar case to last season. There’s a chance Liverpool or Arsenal could win certainly, but City we have as still the strong favourites. The thing you hope for is just they get bored or complacent. They’ve got the best squad and the best manager. That’s what it comes down to.
You Might Also Like
