xGenius, page 8
For four years, from 2008 to 2012, Graham consulted at Tottenham. Or at least, he ‘consulted’ in the loosest meaning of the term. The managers who ran the club over that period had little interest in his consultation. The struggle of football analytics has been one of getting engagement from those who run clubs. Countless stories have been told of analysts finally breaking into the inner sanctum of a team, only to be shunted to a corner and ignored by the head coach. Modern-day managers can no longer be seen to not have an analytics department, but they’re under no obligation to pay attention to them once they’re there. Fenway Sports Group wanted to be different. They wanted to build a team with analytics at its core. They wanted Ian Graham.
Modern-day managers can no longer be seen to not have an analytics department
When FSG bought Liverpool, they hired Graham to build a replica of the Boston Red Sox’s research department. The reaction from the staff already inside Liverpool Football Club was less than favourable. Graham and his team were touted as ‘laptop guys’ who ‘didn’t know the game’ according to Barry Hunter, who was installed in Liverpool’s scouting department. Graham was unfazed. He became completely absorbed in his quest to defy conventional football wisdom and bring success to his childhood team.
Another key cog in the Liverpool backroom staff was Michael Edwards, a man who keeps a notoriously low profile, never speaking to the press and rarely making public appearances, but who has nonetheless been instrumental in Liverpool’s recent success. Edwards showed a hint of what was to come when, while playing for Peterborough U18s, he took a keen interest in computing. His teammates teased him for throwing himself into an IT module that the rest of the youth squad considered a waste of time. After leaving the club, Edwards studied business management and informatics at Sheffield University before landing a job as an analyst for Prozone, a budding football data company. The business operated out of a warehouse in Leeds, where its small contingent of part-time processing staff would film matches and then meticulously break them down into datasets that could be used to objectively analyse the performance of teams. At the time, the work done there was pioneering.
The company was looking to place their analysts at professional clubs around the country. Edwards got assigned to Portsmouth, where he was responsible for analysing the first team’s performances, presenting the tactics of upcoming opponents, and evaluating transfer targets. But the year was 2003 and football wasn’t ready for the data and technology revolution that Prozone were looking to launch. Edwards once received a phone call from Portsmouth manager Harry Redknapp early in his placement at Fratton Park. The gaffer had rung him to complain that he had put a CD-ROM containing player data into the CD player in his car and couldn’t fathom why it wasn’t playing anything. For younger readers, that’s the equivalent of trying to put a USB stick into an iPhone.
The players were equally as inexperienced when it came to analytics, but Edwards quickly struck up a rapport with them. He wasn’t afraid to tell even the more illustrious members of the Portsmouth team, which housed the likes of David James, Sean Davis and Peter Crouch, that, statistically speaking, they’d been very disappointing that weekend. If Edwards had an opinion, he’d let you know it. The players started coming to him, asking about any weaknesses they could exploit in upcoming opponents. Such questions were usually reserved for the manager, but Edwards had gained the respect of the team. His input wasn’t just worthwhile, it was actively sought out. Edwards even ran a Champions League prediction game with the players. Each week he’d collect score predictions and whoever finished bottom of the table would have to drive a Robin Reliant to training, and add an outlandish modification to the vehicle in time for the next Champions League gameweek.
Harry Redknapp left Portsmouth for Tottenham in October 2008 and recruited Michael Edwards to join him a year later. Edwards’ stint in the Spurs analytics department lasted two years, before he was poached by Liverpool and installed by John Henry as the Head of Performance and Analysis. He served an important role in setting up the club’s initially much-maligned transfer committee. Over the years he worked his way up the ladder, eventually serving as the club’s Sporting Director. His role involved trying to cater for the medium- to long-term interests of the club. His position was fairly revolutionary in English football at that stage. The fact that Edwards could focus on longer-term strategies meant Klopp could throw himself into the day-to-day coaching of the squad.
From the outset of his tenure, Klopp embraced the analytical work the club was doing behind the scenes. He met semi-regularly with Edwards, Ian Graham, and other members of the analytics department to discuss player recruitment and tactical matters. Klopp became skilled at translating the data and research of the performance analysis department into easily digestible team talks and strategy meetings with the squad. Before each game, Graham and the three analysts who worked under him compiled a packet of information. By the time these insights reached the ears of the players on the training pitches or the changing rooms, the PhD-level equations were long gone. Liverpool’s playing staff were only faintly aware that the tactical advice was rooted in doctorate-level mathematics. ‘We know someone has spent hours behind closed doors figuring it out,’ said midfielder Alex Oxlade-Chamberlain. ‘But the manager doesn’t hit us with statistics and analytics. He just tells us what to do.’
Klopp’s ability to blend traditional coaching methods with pioneering data analysis quickly led to a change in the club’s fortunes. Liverpool finished in the top four in each of his first six full seasons at the club, winning their first ever Premier League title in 2020. They reached the Champions League final in three of the five years between 2018 and 2022, lifting the trophy in 2019. Between 10 March, 2019 and 24 February, 2020, Liverpool played 36 league matches. They won 35, drew one and lost none. The Reds accumulated 106 points from a possible 108 over this period. Luck obviously plays a part in forming such an unbelievable record – the best record over a 36-game period in league history, no less – but the performance of the team over this timeframe was a world away from where the club had been before FSG took the reins. Fenway Sports Group had returned Liverpool to the pinnacle of world football.
And they’d done it on a much tighter budget than their competitors.
Ian Graham’s role, alongside preparing data packages before matches, was to assist in the recruitment of players. Here, too, the new structure of Liverpool gave them an advantage. Graham built his own database to track the progress of more than 100,000 players from around the world. The power of transfer veto had been taken out of the hands of the manager and placed into the hands of a transfer committee, on which sat the manager, members of the scouting team and members of the analysis department. Instead of relying on one man’s knowledge, as is football tradition, Liverpool began making decisions based on the shared wisdom of the collective. Much of this wisdom, of course, is founded on the advanced statistical analysis of Graham and his team.
In 2017, Graham’s model identified Mohamed Salah as a transfer target worth looking at. Salah’s record in England left a lot to be desired. He scored two goals in 13 games over two seasons at Chelsea, spending lots of time loaned out at other clubs. Graham’s data suggested Salah would pair up well with Roberto Firmino, already on the payroll at Liverpool and creating more xG with his passes than almost any other player in his position. Jürgen Klopp preferred Julian Brandt, a youngster who was making waves in the Bundesliga, to Salah. Brandt seemed the obvious choice, but Graham and his model won out in the end. Liverpool paid Roma $42 million for Salah, who went on to score 32 league goals in his first campaign at the club – a record in a 38-game season in the Premier League era. The Egyptian became a symbol of Liverpool’s revival under Fenway Sports Group.
Perhaps Ian Graham’s most important acquisition came a few years earlier. One of his first assignments under the FSG was to research a forward who was playing at Inter Milan, Philippe Coutinho. Graham’s model approved of the young left-winger, who Liverpool ended up signing for $16 million. Coutinho contributed greatly to the club on the pitch over the next few years, but perhaps his greatest contribution came in the windfall he facilitated when Barcelona bought him in 2018. Liverpool sold Coutinho for $170 million, a whopping $154 million more than they’d bought him for five years earlier. The money was reinvested in several new signings, once again recommended from Graham’s model. Among them were Alisson Becker, Virgil van Dijk and Fabinho, the core of the team which drove Liverpool to Champions League and Premier League success.
Jürgen Klopp clearly played a huge part in Liverpool’s ascension. But his appointment only tells half the story. The team of veiled figures who were installed by John Henry guided the strategy and direction of the team. Behind the scenes, these figures lent on xG data and several other advanced metrics to influence decision-making. This hidden team of number crunchers were happy to let Klopp take the public acclaim. The manager acted as the outward face of the operation, allowing Liverpool’s team of analysts to fly under the radar. But these puppet masters were equally as, if not more, responsible for Liverpool’s return to prominence as Klopp or any of his players. FSG’s restructuring of the Liverpool organisation is what set them apart from the competition and allowed them to compete with the likes of Manchester City, Chelsea and Manchester United, despite possessing far shallower pockets.
Liverpool’s approach under FSG is what Isaiah Berlin would describe as ‘foxy.’ They deploy a transfer committee which takes multiple viewpoints into account. They utilise knowledge from spheres outside of football – their analysis team was headed by Graham, a PhD physicist, and counted in its midst a former junior chess champion, an astrophysicist, and a former CERN employee who helped verify the existence of the subatomic Higgs boson. Even their owner, John Henry, initially achieved success from an algorithm that predicted fluctuations in the soybean market. The same sort of analysis is knit into Liverpool’s DNA. The club’s ownership have constructed a world-class side while maintaining a competitive wage balance. And they’ve done it by taking power away from emotional individuals and placing it into the hands of cold, hard data.
Chapter Summary
John Henry’s adoption of statistical methods in business, and then baseball, was a reflection of the broader trend of using data analytics in sports and also its growing importance across the commercial and financial world.
Liverpool FC initially demonstrated the type of traditional mindset that has long been at the forefront of football management, namely that it is people, not data, that make decisions. However, that soon changed when Jürgen Klopp embraced the new-found analytics and grasped its effectiveness and potential.
Liverpool FC’s data analytical team was composed of people from diverse backgrounds. This integration was critical to the success of the club as it brought together expertise from various disciplines, underlining that there is no one ‘perfect’ collection of analysts; people can add their own strengths into a wider system.
Data-driven approaches require human ‘buy-in’ to succeed – neither can work without the other.
Note
7 Ian Graham’s analysis found that Klopp’s Dortmund were the second-unluckiest team over the previous 10 seasons of the Bundesliga.
6
The Manager’s Office
A New Breed of Head Coach
‘We re-designed the club based on one question: “what would a football club look like if it had no human eye and ear?”’
Rasmus Ankersen, Southampton (and former Brentford) Director of Football
The Brentford manager, Mark Warburton, was sitting in his office on a cold February morning in 2015 when the story broke.
‘Rayo Vallecano head coach turns down Brentford approach,’ read the headline in the Mail on Sunday.
Warburton’s side had enjoyed a positive victory the day before and were sitting in the play-off places in the Championship. They were pushing for promotion to the Premier League despite possessing the 21st biggest wage budget in the 24-team division.
David Weir, Brentford’s assistant manager, turned to Warburton, ‘I think that’s a game-changer, gaffer.’
Warburton picked up the phone and called his owner, Matthew Benham.
‘What’s going on Matt?’
‘I can’t lie to you, we need to have a chat.’
‘Lunch tomorrow?’
They had lunch the next day. It was confirmed that Mark Warburton, the man who had led Brentford to the Championship and was on the brink of pulling off miraculous back-to-back promotions, was to leave the club when his contract expired at the end of the season – regardless of whether the team reached the promised land or not.
When rumours of Warburton’s departure first reached the fans, they assumed he’d been poached by a bigger club. As is the way when a manager outperforms expectations at a small club, a big club will eventually come knocking. Brentford’s fearless playing style had put Warburton’s name on the map. They’d swept aside clubs 10 times their size financially, the likes of Leeds United, Fulham and Nottingham Forest. Their attacking brand of football had earned them and their manager admirers all across the country. Speculation swirled that Norwich City were interested in securing Warburton’s services, with Premier League interest also rumoured on the football grapevine.
Then another story broke. The Times published a piece alleging that Warburton would be ‘sacked’ at the end of the season, irrespective of where the already over-performing team finished in the table. (This statement was true, apart from the allegation of him being fired. In reality, Warburton’s contract wasn’t being renewed. But ‘not renewed’ doesn’t sound as sexy as ‘sacked’.) Fans and the media quickly turned on the club. The narrative was no longer that Brentford were the plucky newcomers sticking it to the big boys. They were now the laughing stock. Who do they think they are, casting aside the man who led them from League One obscurity to the Championship play-offs places? The Guardian posted an article headed, ‘Silly Party candidates bidding for a majority in football’s boardrooms’ which compared Benham to a string of recent tyrannical owners to have taken over at Football League clubs.
The Brentford players made their feelings known the evening of The Times article’s publication. When Andre Gray opened the scoring against Watford, he and his teammates ran straight to the touchline to mob Mark Warburton. The Bees went on to lose that game 2-1, then got beaten 3-0 by Charlton Athletic a few days later. Shortly after the Charlton hammering, the club finally broke its silence on the situation. Brentford released a statement confirming the departures of Warburton and his coaching assistants that summer.
What had been brewing behind the scenes at Brentford before this dramatic story broke serves as the perfect vignette for the struggle that analytics has faced in conquering football. The January transfer window had come around a month earlier. Brentford had been the surprise package that season. Many had tipped them for an immediate relegation back to League One, but they’d defied the odds and were third in the table on Christmas Day 2014. However, things were not looking so rosy according to Matthew Benham’s mathematical model. Smartodds had crunched the xG numbers and seen that Brentford were overperforming; their results had been better than their performances should have allowed. The Bees were lucky to be third in the table. In fact, the xG data showed they’d been the 11th-best-performing team and were likely to regress in the second half of the season. Benham determined the best chance of maintaining their promotion push was with new additions to the team, many of whom they’d already identified via their data-driven scouting tools. Moreover, the current squad was small and likely to tire as the brutal 46-match campaign wore on, making new recruits even more of a necessity. According to Benham’s analysis, Brentford’s chances of promotion would be boosted by more than 10 per cent with a sprinkling of new players in vital positions.
Warburton disagreed. His view was that the tightly-knit dressing room had got the club that far and new arrivals would disrupt the squad’s harmony at this crucial stage of the campaign. The dressing room was strong and solid and didn’t need changing. As Benham became increasingly determined to empower his data analysts, Warburton became increasingly resistant. It was as if both men were stuck in quicksand; the harder they tried to pull free, the more they found themselves stuck. Warburton would state that you can’t use data to measure factors like team unity, combativeness and togetherness. He would claim the team had momentum. Benham’s model told him a different story.
The two camps had set their stalls out. On one side, Warburton, the traditional ‘football man’ who knew he’d have the backing of the football world and plenty of leverage in the form of job offers if he didn’t get his way. On the other side, Benham, the analytical ‘outsider’ who dared to dream of a new model for his club, a world where the manager wasn’t all-powerful and where data was harnessed to achieve an edge. The dispute eventually reached a point where Warburton and his team banned Benham’s analysts from accessing the training ground. This is perhaps the most literal representation of the struggle that analytics has faced in sport. ‘Football men’ holding the keys to the palace and refusing entry to science. The culture war had reached its pinnacle. Warburton had locked the door to analytics, both ideologically and physically. If Benham and his ideas were to prevail, he’d have to break the door down.
And that’s what he did. He went out in search of a coach that would fall in line with his radical ideas. His xG model identified that Rayo Vallecano were performing particularly well in La Liga, so Brentford made an approach for their manager Paco Jémez. The Spanish coach turned down their advances and news of the approach filtered through to the media. When the story broke, Benham decided to go all-in on his revolutionary philosophy. He confirmed that Warburton’s contract would not be renewed at the end of the season and braced himself for the inevitable backlash from the rest of the football world.
