Expected Goals: The story of how data conquered football and changed the game forever

by Rory Smith

Football has always measured success by what you win, but only in the last twenty years have clubs started to think about how you win. Data has now suffused almost every aspect of how football is played, coached, scouted and consumed. But it’s not the algorithms or new metrics that have made this change, it’s the people behind them. This is the story of modern football’s great data revolution and the group of curious, entrepreneurial personalities who zealously believed in its potential to transform the game. Central to this cast is Chris Anderson, an academic with no experience in football, who saw data as an opportunity to fundamentally change a sport that did not think it could be changed. His to infiltrate the strange, insular world of professional football by establishing a club whose entire DNA could be built around data. Expected Goals charts his remarkable journey into the heart of the modern game and reveals how clubs across the world, from Liverpool to Leipzig and Brentford to Bayern Munich, began to see how data could help them unearth new players, define radical tactics and plot their path to glory.

Genres: Football, Sports, Nonfiction, Soccer, Science, Audiobook, Technology

304 pages, Paperback

Published September 5, 2023

Reviews

[Walter U. · Rating 4 out of 5]

As an Argentinian, I have been a football fan (it's actually spelled Fútbol, but whatevs...) my whole life. It is our first love: Fútbol, then our mothers (and believe me, they are well aware of their place).

As a data scientist, math, numbers, algorithms, and the hidden patterns woven in data, have pretty much guided my understanding of the world around me and given me a fulfilling, comfortable career.

So this book, sitting squarely at the intersection of my personal and professional interests, was pretty much guaranteed to get my full attention. And that it did.

Unlike other reviewers, I will resist the urge to call it "MoNeYbAll fOr sOcCeR! lolz". This is perhaps the worst thing about the book; the inevitable comparisons. Ok, so here we go...

Unlike Michael Lewis' classic, it lacks an overarching narrative story centered on a rag-tag group of heroes centered on upsetting the established order of things. It is also much less rewarding in terms of its conclusion. But more on that later.

Among others, Rory Smith uses the journey of two Ivy League professors into the insular world of England's top-flight league as stand-ins for Beane and Co. and the alienating aspect of their data-driven approach but it just doesn't have the same effect.

He weaves many threads, connecting disparate lines to show the emergence of the same ideas (and metrics) in different parts of the world, by different teams operating under different pressures: some teams used metrics to hold players accountable, others to figure out which players to sign for the upcoming season, others still for more mundane tasks such as to find the best way to execute a throw-in or corner kick.

At the same time, it is a tale that is about much more than just the game itself or the fortunes and successes of a select group of teams. It is about the birth of an entire metrics industry. It is a bit of a fun riot of a read if you like soccer and even if you just maybe even heard about analytics that one time at the pub, but don't expect a direct analog to Moneyball.

I can't give it full marks because I found it too shallow in its treatment of the metrics being used by the teams to achieve their goals. I didn't expect any math of formulas per se, but maybe more elaborate definitions followed by concrete examples would have sufficed.

Also, while we know that Beane's As achieved incredible success as underdogs in a league that was designed to allow money to pretty much buy titles, the same cannot be said for most of the teams examined by Smith. Yes, some of them achieved immense success, but they were also the same teams that had access to astronomical amounts of money. How much money do you ask? Oil money, buy-the-metrics-company-so-that-no-one-knows-our-secrets money. But that is also part of the Moneyball way of doing things, according to Beane himself; that record-breaking signing makes sense if the data says the value is there, bargain or not.

Overall, it is a must-read if you like Fútbol, football, soccer, futebol, etc...

[Isaac Armstrong · Rating 3 out of 5]

Starts off strong with insights into the first teams that started to introduce data and the theories etc they were applying. Then there’s some interesting anecdotes along the way, but for a lot of the book he tails off seemingly mid explanation with some completely new side story that doesn’t really add much - just feels abit like he was trying to make a strong 150 pager into almost 300. Though still worth a read if you’re interested in this kind of thing.

[Ferhat Culfaz · Rating 4 out of 5]

A nice summary of the advances in data and analytics in football. Kind of like a Michael Lewis book than any technical details.

[Pete · Rating 4 out of 5]

Expected Goals : The Story of How Data Conquered Football and Changed the Game Forever (2022) by Rory Smith is another book that looks at the rise of data and analytics in football. Smith writes about football for the New York Times.

Football has been a hard sport for analytics to really make an impact on. In many popular sports like cricket, basketball and baseball more was quantified in score sheets that enabled analysis earlier. A football score sheet has very little information. The early unsuccessful efforts of Charles Reep are described.

Football management was dominated by ex-players who didn’t have the skills for data analysis and who tended to distrust ideas coming from people who had not played football at a high level. To be fair, they also didn’t have the data until the 1990s at least.

In the 1990s as games became ubiquitously videoed with multiple cameras firms started mark encode all the actions in the game. Expected goals describes how Ram Mylvaganam started ProZone, one of the first companies to use video for game encodings. Once the data were there then analysts started to use the data to work find players who were likely to be stars and how the game could be better played. Initially the main thing that was sought out was which players were not putting in enough effort in terms of sprints. Smith makes the point that in the 1990s food and fitness had been the focus of much improvement in football.

Two of those analysts, the academics Chris Anderson and David Sally fill much of the book. The two wrote a book ‘The Numbers Game’ about how statistics could be used in football and then moved to England to use their insights.

Other bloggers and analysts around the world world were also picked up by teams who wanted to improve their efficacy.

There is a chapter on how Liverpool became one of the Premier Leagues big successes using data and analytics. Klopp had an open mind and had some familiarity with data use from the Bundesliga, but Liverpool took it to a new level.

Expected Goals is a very good read about the growth of analytics in football. The book also points out how hard it still is to get actionable results in football.

[Colin Thomson · Rating 4 out of 5]

Rory Smith is one of my favourite regulars on the totally football show. I also happen to work in Data. So of course, the merging of Data, Football, and a favourite journalist, was something that really appealed to me.
I really enjoyed the focus on humans. It was something of a relief that I didn't have to switch on parts of my brain that I turn off at 5pm. The mistakes were by far the most interesting part, the times the data was there but was read totally wrongly by very intelligent people. It's not only reassuring but evidences the journey that data in football is on.
That I found myself so invested in the lives of so many of football's background figures, people not meant to be protagonists, is testament to the writing and research of the author. There are so many interesting stories in football beyond Messi and co., this is absolutely one of them.

[Yousif Al Zeera · Rating 4 out of 5]

Expected Goals, or 'xG' as it is widely known. One of the football metrics that became mainstream as of late. But the book is not solely about xG. It is about the journey of the data revolution in football since the early proponents of data before the mid of the twentieth century till now, especially the accelerated push in the last 1-2 decades. Data, once shied away, became an integral part of world football. Decisions in the transfer market, tactics, formations and more, all revolves around data. If not, then you might see yourself unable to compete anymore at the big stage.

[Brian · Rating 4 out of 5]

Smith is a good writer and story-teller and this is a very engaging read. Even as someone who follows data analytics in soccer/football, I learned a lot about this history of how the analytics got there in the first place.

[Devin Ruff · Rating 4 out of 5]

Okay book

[Griffin · Rating 4 out of 5]

Rory Smith is such a talented writer. This is ostensibly a very niche subject, but he manages to make it interesting, simple, and, most impressively, human.

[Erik · Rating 4 out of 5]

Endelig! En bok jeg likte skikkelig godt, etter en rekke 3-stjerners leseopplevelser. Suveren lesning for alle som er nysgjerrige på hva som skjer bak kulissene i internasjonal fotball.

[Eivind Haugen · Rating 4 out of 5]

Fun, interesting and educational.

[Mbogo J · Rating 3 out of 5]

Ardent fans of a particular sport are wont to have their own opinions on what makes a team or player tick. I would hazard data ranks very low on their analysis. I number in their company and in football [if you call it soccer you should stop reading now] I think subjective factors such as a player's form, the coach's tactics or even the locker room attitude matter far more than seemingly objective metrics like ball possession, attempts on goal and their ilk. I have left a few souls exasperated when they tried to use data to try to convince me that CR7 is the goat. Messi is, simply because he is. It is a truth that is self evident and needs no data to back it up... I read this book not to change my mind but rather to see if there is anything new to make a case for data use in football decision making.

This book is a bit of a mixed bag. It has an easy flowing narrative and describes the various behind the scenes things that happen while running a football club and this might be enough to read the book were it not the fact that it falls short of the goal espoused in its title. The narrative never makes it clear how data was used in decision making, it just claims that an individual used data to make some decision. How exactly? The authors should have made effort to speculate at least how various data points were used and how insight was acquired. While its understandable that teams might guard their models, part of the book's research should have included speculative models. The authors failed in this task to the extent that they never even defined the expected goals model from which the book derives its title. If you want to know how the sandwich is made, you'll have to do your own separate research. Unfortunately for me, making sweeping declarations with no proof does not cut it, without evidentiary rules you can claim anything. There is a story that usually does rounds whenever Pep inevitably bottles his chances of winning the Champions League with Man City; it claims that African shamans cursed him never to win the cup on account of how he treated Yaya Toure, when there is no need to provide proof, this story bears the same weight as claiming that someone made a signing using "data" with no further elaboration on what data or even the insights gleaned. It also did not help that "data" was conveniently being used to justify players already known to be stars. You can claim anything retroactively. What was data's predictive ability?...

The book eventually degenerated into a series of stories on some personalities you've never heard of and how they used or tried to use data to improve football. All the stories follow a familiar path of some person usually in a stable white collar job but with a gut instinct that data should be somehow useful in football and with little to no knowledge of how football operates they set about accomplishing their goal including quitting their stable jobs...

My initial biases on data use in football were never changed. Not one bit, I still believe data has more of an auxiliary role rather than a primary one in running a football club. No one wins championships with data alone. The story of data use in football would have benefited from the input of a data practitioner rather than only being told by journalists. Countdown starts to when we will start hearing stories of how A.I revolutionized sports.

[Sunny · Rating 5 out of 5]

6 stars. Perhaps one of the best and most underrated books i have read on football. Implanted the 5.5 concept in my mind which i’ve passed onto the boys and its fundamentally changed their game. Here are the best Bits:

He called them his 'fantastic four. He knew that, if his team kept 16 clean sheets over the course of a 38-game season, it would not be relegated. He knew that if his team scored first. it would have a 70 per cent chance of winning a game. He knew that if his players covered more distance at a speed above 5.5 metres per second than their opponents, their chances of winning would go through the roof.

If you looked at that game, Brazil had more shots, more passes, more corners. But Germany won 7-1. It told you that those statistics were not telling the right story from the game.
Impect's approach is different. The company's foundational metric - the piece of information it is looking for from a game - is known by the slightly uncomfortable anglicism of packing. It is, at heart, a measure of how many opposition players are bypassed by any single action on the pitch.

football's in-built conservatism, its cherishing of the old ways, its reverence for tradition, the scepticism and suspicion it reserves for anything new-fangled or vaguely intellectual, or, the greatest sin of all, American.

Reep found that with every extra pass in a move, the chances of retaining the ball fell significantly. He discovered that football was a game not of possession but of turnovers, and that those turnovers were disproportionately valuable: long before Jürgen Klopp and Ralf Rangnick and even Marcelo Bielsa made it the bedrock of their playing philosophies, he could claim that the most sure-fire way of recording a shot on goal was to win the ball back in or near the opposition penalty area, and he had the data to prove it.

To understand the data, in his mind, was to understand the pattern. The same principle applied to football.

Marc Andreessen, the Silicon Valley investor and founder of Netscape - one of the dominant browsers of the early internet - has a dictum that, in his business,
"being early is the same as being wrong. The timing of technology is as vital to its success as the substance of it

football was a lucrative industry, but it was mired in inefficien-cies, held back by outdated thinking and tangled up in moribund traditions. Clubs did things because that is what they had always done, and even owners new to the game seemed to allow themselves to be shaped by that herd mentality. He wanted to do something completely different: cutting edge and untested and, to some extent, heretical.

Among the many and varied ways in Which football misinterpreted the lessons of Michael Lewis' book, and Beane's life's work, was the assumption that any club adopting a data-led approach would, inevitably, not spend significant sums of money. Why would they, when the analytics they had access to would give them the ability to sign hugely talented players for a fraction of the cost? 'Money-ball' was about exploiting inefficiencies in the market; paying premium fees was the very definition of an inefficiency.

He had long been interested in physical data: he knew, for example, that there was a correlation between how many high- intensity sprints a team made and how likely they were to win;

that the greatest challenge to the use of data in football was not - despite received wisdom running to the contrary - that the game was too fluid to be quantified, but that it was occupied by a group of traditionalists and conservatives who would see the empirical truths offered by the numbers as a threat to their power.

Since buying Brentford in 2012, and Midtjylland two years later, Benham has always been clear, though, that data is not the only tool at his teams' disposal. Ankersen, for a long time his trusted consigliere at both clubs, speaks frequently about what he calls football's 'end of history illusion, the idea that the game as we see it now. They broadcast around the world every three days in ultra high definition, is the highest and ultimate form of the game. We fall into it, he C believes, because when we review footage of old games - even this case, 'old' goes back no further than the 1990s - the players seem so much slower, so much less athletic, and the systems and formations feel so rudimentary. We look at how far we have come and we assume that there can be nowhere else to go. Ankersen, certainly, has little time for that idea. We think the players can't get thinner or fitter or run more or work harder, so this must be it' he said. 'But there are always edges? Midtjylland and Brentford see it as their task to find those edges. They might, as Ankersen said, be in terms of nutrition, or injury prevention anc recovery, or even in sleep, disciplines in which footballs interest t still young. They might be structural: it is telling, for example, that in 2016 Brentford took the unorthodox decision to scrap its academy

They realised that they lost possession from throw-ins with alarming regularity, so they hired a throw-in coach to help them improve both their technique and their strategy. "There is a lot that has been disre-garded, Ankersen said, whole areas of the game that nobody has ever really tried to improve. When we spoke, late in 2020, he even had his eye on kick-offs. Central to all of that, of course, has been data. Midtjylland - and Brentford - run on data; everything is checked and assessed and verified according to the data: the players they sign, the way they play, the decisions they make. Half-time team-talks are influenced by text messages sent to the coaching staff, detailing how the team is performing according to a prescribed series of performance indicators. In Denmark, they are now even trying to apply the same approach to one area that would seem immune to it: the psychology Of players. Midtjylland have partnered with a university to study whether there are any common characteristics among players who have succeeded at the club, to ask what sort of traits they should be searching for when recruiting.

There has, for example, been a notable demise in the popularity of the long-range shot. In the 2003/04 season, the Premier League's players took more than 5,000 shots from outside the penalty area. By 2020/21, that had collapsed to 3,333.

It seems likely that is testament to the increased awareness among teams - even those who are not quite so devoted to data as Brighton and Brentford and Liverpool - of metrics like Expected Goals; it is reasonable to assume that, five years or so since it first appeared on Match of the Day, most coaches have worked out that taking a shot from long distance is often less valuable than retaining possession and waiting for a better opportunity to present itself. But it is also possible that it has been driven by a desire to ape the style popularised by the dominant club teams of the era: Pep Guardiola's masterpieces at Barcelona, Bayern Munich and Manchester City. Their predilection to keep the ball circulating until the perfect chance arose was not - or at least not solely - because of an innate understanding of probability, but because of a deliberate philosophical choice made by their manager.

At RB Leipzig, Rangnick installed a custom-made clock at the training ground. During small-sided training games, it would be started, ticking loud enough for the players to hear. It gave them eight seconds to win the ball off their opponents, and ten seconds after that to have a shot on goal. If they failed, they had to give the ball back. The principle, for Rangnick, was simple. The greatest moment of danger for any team is in the moment when they switch from a defensive mode to an offensive one. His team, therefore, had the best chance of creating a goal-scoring opportunity if they won the ball back quickly after losing possession, and then wasted no time in moving it as close to goal as possible. He did not just believe that. He knew it, because that is what the data told him.

The clubs that parse the data best will, if the early years of the digital era have illustrated anything, make more good decisions or, at the very least, fewer bad ones. That will,
in the future, be the difference between success and failure.

[Liam Power · Rating 2 out of 5]

Fan of the writer but felt a bit let down here. A lot of chopping between threads with slightly awkward pacing. There's also a few subbing errors and repeat sentences which is a shame.

[Neil Bradford · Rating 2 out of 5]

Not quite the Moneyball for football/soccer it wants to be.

[J · Rating 1 out of 5]

Had high hopes for this book, but basically it’s just treading water for 250 pages.

[Dipra Lahiri · Rating 3 out of 5]

A choppy narrative, flitting from story to story, made the reading discombobulated. Analytics in football has come a long way, and today, is central to strategy and player selection.

[James · Rating 4 out of 5]

This successfully did what it set out to, which is tell the story of how data analysts increased in importance in football without going into details about the metrics themselves. Smith relied heavily on Chris Anderson, credited author of The Numbers Game, to provide details of how football owners operated during his doomed efforts to become a footballing Billy Beane, but he had also spoken to many people who developed or used analytics tools.

The structure of the book was similar to The Blind Side in that it followed the individual story of Chris Anderson across the book, while covering the story of analytics in general. I think Smith tried to portray Anderson sympathetically, but to me he came across as a complete chancer. Going into football stadiums posing as an American tourist probably supplied some great anecdotes, but to what end? Without any proof his methods had worked on a smaller scale, he wanted to be put in charge of a football club purely with a theory that there were inefficiencies in how clubs were run. Smith wrote that he didn't just want to play Football Manager, but it's not clear what he actually thought he could do, and he was disillusioned when he was actually in charge of Coventry because he had to balance the budget and arrange commercial deals. Having read two books that cover this aspect of his career, I'm still left perplexed as to what he wants to do that would be in any way plausible for someone with little relevant experience.

The interest from this individual story was in the characters he met and the extent to which football ownership is quite opaque, as it's often unclear who actually has some power and inflence. There were insights new to me (although I have mostly avoided the few agent/deal themed football books) in these chapters but in general I preferred the broader history, from Opta generating stats for broadcasters early on to a race from rival companies to collect the most data. Due to the levels of secrecy Smith rarely got to explain what data clubs used to improve performance, but several figures from Comolli to Monchi were interviewed and Smith hadn't only relied on quotes previously given to the media.

Since granular details aren't available success stories are generally illustrated by modest transfer fees for players that turn out to be brilliant, or mention of set piece prowess. One interviewee used a prior recommendation of a younger De Bruyne to illustrate the value of his data, a discovery less impressive when you realise he had already been bought by Chelsea and was only on loan at Bremen at the time. Without the precise details it was therefore impressive that Smith had put together a coherent but interesting narrative thread.

The writing itself was fast paced but with the odd 'couple of dozen' Americanisms thrown in as a nod to his main employers, along with introductions I associate most with US journalistic profiles*. The acknowledgments suggest this started from research he'd done for newspaper pieces, but it didn't feel as though there were disparate chapters, and quotes were used sparingly, a merciful decision when one of the interviewees was Duncan Alexander.

There are two main things stopping this being a five star book in my eyes. One is the lack of critical evaluation of Anderson, who isn't asked why he should be in charge of a club and isn't just a committed fantacist. The second is raised by Smith himself, which is that he doesn't have any details as to what effect the analysts had on their teams - we are left to conclude that Liverpool's success was helped by data, and it seems plausible, but he couldn't demonstrate it.

*The sort of intro that starts "With his Yoshi figurine on the shelf and his handlebar moustache, X Yman didn't strike you as your typical football coach, and only the odd grey at his temples hinted that he was indeed a grown man. But as we shared thoughts about yesterday's Champions League tie to break the ice, it was clear he had an incredibly sharp tactical mind."

[Scott Cumming · Rating 4 out of 5]

“The Data”

For Aberdeen fans, like myself, the above phrase will always be tethered to our chairman, Dave Cormack’s antagonistic and somewhat barmy appearance on BBC Scotland’s Sportsound as he attempted to defend his decision to stick with Stephen Glass as manager before giving him the heave-ho a few months later. This book indirectly makes sense of that decision and a few other things that have happened at the club as they have sought to integrate data analysis into their decision making and transfer policy.

Glass’s Aberdeen side were profligate in front of goal and leaky at the back, but I imagine the data league table made probably put a good sheen on things as they were a team creating chances and not conceding many albeit they were high percentage chances inevitably put away by the opposition as we found ourselves on a string of 1-0 defeats or rather uncreditable draws. Such league tables are explained in the context of manager’s keeping their jobs or as an explanation for why a manager should keep his job i.e. Jurgen Klopp’s final season at Dortmund was not so disastrous when looking at the numbers.

Unsurprisingly this book is somewhat akin to being the prologue to football’s Moneyball moment. Having been published in September 2022, there is time to give full credence to the success of Brighton and Brentford, who are mentioned mostly in passing within the story of Data coming to football. The birth of the internet saw data move into football as suddenly there was the possibility of analysing every incident on the football pitch with companies evolving the process evermore over a short space of time. What we haven’t seen yet is the Eureka moment of the undervalued stat being discovered as with Billy Beane’s Oakland A’s. There is, however, a chance we won’t actually discover it until long after it has been discovered as one key point mentioned in the book is that with American Sports it is the leagues who own the data, therefore it can be analysed for everybody.

Alongside this fact is that baseball had records going back many years before Beane and his merry band of analysts hit upon the OBP (On Base Percentage) importance. Football may be reaching that point, but it could also be decades away as the game has rapidly evolved from the 90s to now due in no small part to money and the advances in sports science we have seen.

One area it is easy to see the evolution is through the types of signings we are seeing coming into the Scottish game now with a slew of foreign players being signed presumably off the back of data trends which link in with the type of player being sought who would be available for a fraction of the cost of a Scottish or English counterpart. Aberdeen have been forward in advising that data has contributed to the signings of the likes of Duk, Miovski, Ramadani and more latterly Rubezic and Sokler.

It’s seeing this on a local level that shows how far things have progressed in a short span. The secrecy of clubs means some of the stories within the book were somewhat revelatory in terms of how much data was incorporated in the likes of Spurs, Arsenal and Liverpool until it wasn’t. Spurs partnership with Decision Technology ended in 2018 and it is fair to say they’ve not hit the same heights since then and Liverpool have somewhat flattened out as Klopp’s influence has grown the data seems to have taken a back seat there too. Meanwhile, Brighton and Brentford seem to go from strength to strength and one of the growing compliments surrounding the clubs is their ability to deal with departures with another player ready made to slot into the vacated role within their side. This must in some way be attributed to data and comparing things across undervalued markets as well as comparisons with those they have on their books already.

It’s a fascinating subject to be sure, but the football man vs data boffin debate continues to rage even now in what should be more enlightened times.

[Kevin Burke · Rating 2 out of 5]

This book felt like it could have been more. One of the problems, of course, is that clubs are fiercely protective of what data analysis they're doing, and that makes it hard for the book to delve into too much detail. So we visit people in the Philippines and Laos and Egypt who spend an increasing amount of time - up to 16 hours - tagging games, identifying outcomes from players under pressure and where chances are created from, but ehat does the end data ultimately look like?
We don't really get hugely under the skin of this.

Too many of the threads in the story seem to fade into dead ends - there's some data there, but it ends up not being used much at a club. Chris Anderson takes over as CEO of Coventry City with a view to trying to prove the benefit of data and turning the club into a fully data-driven one - they're top of the league by Christmas, but Anderson's data tells him they're being lucky and regression to mean is inevitable (and ultimately happens). But he never really has time to roll out his data plan before leaving the role inside 18 months. Clubs are hiring throw-in - does it help any though?

Very late on, we start seeing some results. Brentford, Brighton, Liverpool and Midtyjlland are cited as clubs who took data fairly serious, identified specific players because of it, and reaped the rewards. Players themselves are using data to analyse clubs to see which suitor would be the best fit for them. On the pitch, shots from distance are on the way down, as are crosses. Passes are on the way up - keeping possession is more important than a speculative shot. Corners are being targeted for marginal gains - though again, this could have been elaborated on. How are they being targeted? A 1.8% scoring rate from corners is quoted, but we don't see where the marginal gains are coming from. Are clubs improving the scoring rate? Are they identifying the most dangerous way to deliver a set piece? We don't learn. Expected Goals, the titular stat and arguably the best known at this stage, get a very slight covering. Midtyjlland have a Table of Justice - the xG table, which suggested regression to mean would see bad form turn itself around in due course. That's about it.

There's a feel that this could have been a much shorter read; the sections tend to go into a bit too much character development at the expense of the actual point. There's an irritating habit of starting a new section with a Dan Brown-like mysterious opening sentence. "In hindsight, it was not what you would call an auspicious start", "The first time Ashwin Raman read the message, he assumed it was a joke", or "As the clock ticked towards 5pm, the Masters of the Universe started to file into the wood-panelled boardroom." There's a bit of repetition too; we're told more than once what Moneyball was.

Overall, it feels a bit of a missed opportunity.

[Sheng · Rating 4 out of 5]

The increasing permeation of data science into football has been growing much more rapidly in the last decade and Expected Goals: The Story of How Data Conquered Football and Changed the Game Forever documents this data revolution. Covering the various personnel involved from data scientists to managers, the influence of statistical approaches from other sports and the challenges in finding acceptance in the world of football, Expected Goals gives an inside look into the world of football data.

Expected Goals was a really interesting read and unlike a few other football books I've read, author Rory Smith has taken great care to do a deep dive into the topic instead of just a surface level coverage. He covers a lot of the challenges initially faced by data science in the sport before its prevalent uptake (this book made me realise that I can't even remember the exact moment when expected goals was introduced before it became the commonplace statistic that it is now). While Moneyball in baseball popularised the use of statistics, there was a lot of doubt in using statistics in the chaotic and unpredictable game of football. The use of statistics initially favoured the long ball and consequently gave terrible results. Along with a whole list of challenges such as lack of statistical data, the inertia in accepting new approaches and even the misinterpretation of data when it is readily available, the road to the high regard that data science is held in now in football was an extremely rocky one. It was really quite intriguing to see the various data companies that entered the sport and the passionate data fanatics which drove it into something that became taken seriously. The final few chapters also covered more familiar names in sport that embraced the data, such as ex-Liverpool manager Jurgen Klopp.

My main issue with the book is that sometimes it focuses too much on the people participating in the data revolution to the point where the data aspect seems quite irrelevant. This was most noticeable when covering the story of Chris Anderson and his gamble in uprooting his family from the States to the UK in the hopes of finding a football club owner that would let him run the club using his data driven strategy. It was alright at first but then it started consuming multiple chapters where he would talk about his personal challenges and every step or rejection of his journey that was frankly less about the data and more about his unrealistic expectations in the first place. Those parts of the book felt like they could have been trimmed.

Overall, Expected Goals was a great inside and in-depth look at the current data revolution in football. 4/5

[Clay Collins · Rating 3 out of 5]

This one is a challenge to review, because there is a lot of good factoids here. The author gives a valiant effort to tell a human-centric story of the rise of data in football, and it really tries to overcome its main obstacle. But we sort of have to talk about the big obstacle.

The vast majority of advances in football analytics have been conducted from the club side of things (or by hobbyists who later worked for clubs). As a result, many advanced techniques and findings are still proprietary. This prevents the author from going into detail about what these contributions are. The book becomes a writing challenge: how do you tell a story of individual achievements without actually detailing what those achievements are?

The author gives a good effort. Many of these stories are interesting, maybe even inspiring. But I struggle to see who this book is for. The book is missing many of the technical achievements that make up modern data analysis of football. Even expected goals- the namesake of the book- is discussed only sparingly. But many of the Rolodex of characters introduced throughout the book disappear at midway points. Some of the book recedes into summaries of “this person brought an analytical touch to this club- exactly what or how, we don’t quite know because this person signed an NDA, but it was important.” Only one person, Chris Anderson, is a constant.

As a result, the book ends up being a slapdash collection of mini-chapters, profile pieces, and a few truly fascinating articles. The Quixotic tale of Anderson- from Ivy League academic to navigating the murky world of club ownership (the author spares no expense to tell you how murky the world of football club ownership is) to eventually running a club of his own, is inspiring, and could’ve made for a great book on its own. There are a few other fascinating tidbits in this book that the author is light on details on. The rivalry of different companies to provide new types of data and analysis to clubs while battling within-club units and each other to innovate could’ve made for a great long form article. But there’s little discussion on the “exploitative” nature of the analyst industry. Analysts are often working longer hours than many coaches but for only a fraction of the pay. Much of the data collection is relied upon by workers in developing countries with strict deadlines to provide clean data to clients. It’s striking the realization that almost every individual profiled in this book, many of which incredibly influential in bringing data to the sport, no longer works in the football or analytics industry- the author never poses why that is.

[David Margetts · Rating 4 out of 5]

Having spent my career in sales and sales management, I have always valued 'The numbers', the data to understand performance, productivity and development opportunities. The numbers tell you about the 'possible' causes of performance by identifying effort, proficiency, skill ratios and therefore the possible the strengths and weaknesses of your people. Discussion, observation, practice, and customer feedback provide the confirmation of what the numbers might be telling you. Coaching and leadership help to produce and implement the solutions and desired results. This is just simple basic stuff that even the most junior of managers is aware of in business (or should be) and yet in football it has taken decades to even acknowledge data may help. Even as an amateur coach for a junior team, my request for data was derided by my colleagues. Moneyball was the breakthrough in baseball across the pond and quickly adopted, but football is another story. Preconceived ideas, secrecy, myths and legends, closed mindedness and basically a lack of professionalism in the game has meant that what should have been obvious and a necessary investment has taken decades. Thank goodness for minnows such as FC Mittyland, Brighton and Brentford to finally prove that in using data effectively, they can democratise the game and compete with the 'best'. Funnily enough the coaching team that teased and ridiculed me for wanting to track basic data at games were mostly Liverpool fans, unaware a revolution was taking place under their very noses. This book is an excellent insight into not only the power and value of data and information, but an indictment of the incredibly poor, short sighted and incompetent people managing a multi billion dollar business and losing most it, due to poor decisions in recruitment, strategy, tactics and investment. Fortunately things are beginning to change, and as such there will be winners and losers, but hopefully more Brentford's and Brighton's who can upset the 'big boys' (sadly Southampton, my team, in spite of making early progress in this area, through a succession of poor managers, seemingly 'lost the plot')

[Cam · Rating 1 out of 5]

Terrible.

This book loudly proclaims that it’s a book about people and that it doesn’t have one single equation. Maybe if done currently this decision would be a wise choice. But instead there are way too many people. The ones that feature briefly then don’t contribute further. Ones that receive a bunch of build up only to never quite make an impact. Or my favourite, the ones that are so forgettable that they get mixed up with other people. Or maybe they come back but it’s not clear the person has already featured since the writing is supper muddled and there are always weird overlapping time lines. Oh and with each character there is a weird fetish on credentials. And like a bizarre personal story created for the person which I’m not convinced is based in reality.

Shying away from equations/stats/maths could have been a wise decision to make the book a clearer and easier read. Could have been. Except forgoing the precision of maths for the poetry of prose leads to absurd reading experience where the poor reader needs to wait until Pg 187 to get to an intuitive definition of expected goals. Next time help poor math-phobic reader and put some the books entire point earlier on?

Other things that annoyed me:

One - Starts building a narrative (in typical expected goals fashion very slowly and with a lot of side quests) to something vaguely representing a point. Then bam, the writer stops suddenly and dives into another point. I think the objective is to weave together related points. But the confusing timelines and detail mean this isn’t really effective.

Two - Write your own book. Moneyball must have been mentioned over 100 times. Money all moment. Money ball club. Blah blah blah. We get it.

Three - the excruciating long time you have to wait to learn the whole output of all of this. For example, Pg 240 finally gets to some concrete details on how data flows through to coaching. “Cruising the ball is not particularly efficient; in training, the clubs coaches started to declare the corners of the pitch out of bounds for players”. Why the wait?

Four - Meaningless history on every false start.

Five - Too many adjectives and flowery language

[Paul Jones · Rating 2 out of 5]

For me this lacked substance. The writing style is fluent enough but I came away thinking that the content could have been covered in 87, not 287 pages.

The main protagonists are Chris Anderson and David Sally - it felt like half the book was teeing them up as the Moneyball pioneers of football and the plot would come together in a satisfying climax. But the moment never came. Anderson got as far as being CEO at Coventry for about a year where the day to day realities of managing a cash strapped lower league club meant that he couldn't implement his ideas. Ironically this was one of the more interesting sections of the book, yet it wasn't the point of the book. I thought their adventure was a little weak to hang a large part of the plot on.

The book also took us to several clubs looking at their respective data journeys. Again the conclusion each time appeared indeterminate. Much of the content focussed on big clubs who you would expect to hoover up most of the trophies anyway, making it unclear what the added value of their evolving analytical prowess was.

Brentford get a passing mention and it would have been more insightful to develop their story and how they have surpassed the achievements to of much better resourced clubs with intelligent use of data, making them the Oakland A's of the book.

Perhaps this book was a missed opportunity. Perhaps given the secrecy clubs attach to such operations it was always going to struggle to tell the whole story. At least I know why crossing the ball has gone out of fashion.

[Frederik · Rating 2 out of 5]

I really had high hopes for this book as I wanted to learn more about how data is used in the game, and how it improves performance. Instead, this is a history book about some of the people who tried to introduce data to the industry, and how they had a hard time doing so because of the conventional wisdom that only previous professional players could really have a say about the game. Obviously, the resistance to data and their interpretors such as mathematicians, physicists and others, has decreased seeing that more clubs hire a broad scope of people without any prior experience with football to analyze and interpret matches. However, how clubs work with all the "newly" available data is so well kept secrets that it wasn't really a theme in the book. It is and will increasingly become a competitive edge for clubs (e.g., Brentford, Midtjylland, RB Salzburg, Liverpool, Arsenal) so I will probably never get the insight I find most interesting without being on the inside of a club.
Other than that, Smith is a good writer and story-teller, and it is a very engaging read. I just found the content pretty meaningless without diving into the details of how data is changing team compositions, style of play, formations, etc. in specific clubs or leagues.

[Rob Sedgwick · Rating 4 out of 5]

This is very much a Moneyball: The Art of Winning an Unfair Game for football and it's a good read. The problem though is there is no central club and character in Oakland and Billy Beane respectively to really make it as effective a book as MoneyBall. The story switches from club to club, and I lost track of how many English teams were actually profiled for using data and which individuals did what. It definitely lacks a continuous narrative, entertaining and interesting as it is. There's no doubt that data does play an increasing role in football, but even the book itself seems to contradict its subtitle that data has "conquered" the game. Even in the clubs at the heart of this story, it's far from clear that that was ever the case.

It's also a book entirely about the history of data in football. You'll find little in here if you actually want to know what the data can tell you other than a few snippets and references to other works in the endnotes.

[Dewi Powell · Rating 4 out of 5]

I was half expecting an academic read full of calculations I wasn’t going to get close to understanding accompanied by an overload of theory, but it was much more than that.

From the collection of key metrics right through to data-driven decision making, the concepts are impressively explained and well-presented. That much of the history is in such recent times is possibly why the book meanders into relevant, but secondary, sub-stories and themes. Ultimately, they don’t distract too much and do compliment the overall focus.

I didn’t, however, feel much of a connection with the book’s most consistent characters, the professors. Not the connection that the author is intending, anyway. They’re fascinating people but for all of their work, they seem to witness their ideas being realised by others rather than managing to make industry-changing moves themselves.

There are a lot of interesting anecdotes, especially the transfers. I look forward to seeing how the book dates and there is probably scope for a sequel in years to come, especially if the author could get access inside the nuts and bolts of decisions as clubs make them.

[Tim Lucas · Rating 2 out of 5]

Eh. I enjoyed some of it. A lot was more about the people than the analysis. It said at the beginning that this would be the case, so I can't complain, it just wasn't quite what I was looking for.

But some of the style and content annoyed me a bit. The main "story" in the book (Anderson) in the end turned out not to be very important. This person apparently spent a long time trying to buy a club and run it with data at the core. But he never succeeded in buying a club, and seemingly didn't make that big a difference to the use of data in football.

The idea that the the problem wasn't the data or the models, but the fact that the managers weren't interested, was mentioned about 20 times.

Often you'd have a long description of how someone added some component of data analysis to some club. Then it would be mentioned that a few other clubs had done it a few years before.

Towards the end it mentioned that Brighton, Brentford and some Danish club are by far the most data driven. But the role of data in these clubs is described in a couple of paragraphs. It's just very light on "what is data used for in the present".

So overal not my favourite.

[Tim Harrington · Rating 3 out of 5]

This felt very much like a book trying to prove that data has indeed 'conquered football and changed the game forever', yet not really managing to convince with a clearcut case study that proves it beyond doubt. A decent enough read, but I do not think the 'football Moneyball' story has arrived just yet, indeed the book itself seems to conclude that whilst football has opened itself up to data in a big way over the past decade or so, there does not yet seem to be an obvious example of its resounding success. Perhaps that is because the author was not able to get access to a Brentford or a Brighton, who hold their cards very close to their chest - but without the example of a club using data extensively (difficult to use Liverpool due to the influence of Klopp in my view), I am not sure the book's title quite fits the narrative.

Perhaps this book might look very different updated in a decade or so, but equally perhaps many clubs will still be scrambling to find a use for their huge data sets...