xGenius, page 11
Analytically minded clubs have sought to restructure their organisations in an attempt to reduce bias, noise and human emotion from the equation. These teams have diversified the spread of power. The heavy burden previously weighing entirely on the manager’s shoulders is now being shared among experts, executives, Directors of Football, transfer committees, technical performance analysts and advanced mathematical models. Human intuition is gaining greater support from data. Traditionally, the role of the manager was to be the single pillar supporting the entire infrastructure of a building. Clubs have recently begun to build other pillars to help share the responsibility and carry the weight. Before, the whole roof would cave in if the managerial pillar collapsed. Now if this pillar disintegrates, the Director of Football pillar, Head of Recruitment pillar, and a host of other pillars maintain the solidity of the structure and a new managerial pillar can slot seamlessly in. The structure remains intact and an expensive, time-consuming rebuilding process is avoided. Any club not ripping up their current model and emulating Liverpool, Brighton and Brentford is being left behind. It’s easy to forget the resistance these clubs were met with when the structural changes were first brought in. Fenway Sports Group and Matthew Benham were ostracised by the media, their radical organisational shakeups branded as insanity.
Indeed, genius usually looks a lot like insanity at first.
Chapter Summary
The traditional role of the manager is all-encompassing. They have historically been responsible for tactics, recruitment, training, and a host of other tasks. Smart clubs are splitting the role between a cohort of specialists and using wisdom of the crowds to eradicate individual biases from decision-making.
Transfer committees and Directors of Football are becoming increasingly prevalent as clubs appreciate the need to take power out of the hands of the manager.
Forward-thinking teams are becoming better at recruiting managers, as well as measuring their performance once they’re installed in their role. These clubs will ultimately win more football matches.
7
The Recruitment Department
How to Unearth Hidden Gems
‘Player recruitment is the most important application of analytics by a factor of 10’
Dr Ian Graham, former Liverpool FC Director of Research
Deepak Ravindran and his wife Emilie Vanpoperinghe are in the wonky vegetable business.
It all started when they came across a misshapen tomato while on holiday in Portugal. The vegetable looked ugly but tasted delicious, which got the couple wondering why fruit and vegetables in the UK always looked appealing but never tasted as good. A short amount of research revealed that the food production industry creates huge amounts of perfectly edible produce, only for supermarkets to turn 30-40 per cent of it away because it doesn’t meet ‘consumer standards’. One-third of food produced for human consumption is lost or wasted globally, the equivalent of 1.3 billion tonnes per year. The UK alone sends 90,000 tonnes of produce to landfill, but this food is never recorded as supermarket waste because it’s rejected before it hits the shelves. Ravindran reckoned a lot of consumers would be happy to buy fruit and vegetables that didn’t meet supermarket standards. He staged a trial, buying misshapen produce from local suppliers and delivering them to houses around his home in London. The trial was successful enough that Ravindran and Vanpoperinghe pooled their savings and launched their own company: OddBox.
Football clubs can learn a thing or two from the repurposing of forked carrots, conjoined apples, and scarred aubergines. Just like supermarkets turn away tasty produce, football clubs too easily dismiss perfectly good footballers. The market for players is massively inefficient and can be exploited by those who don’t mind buying the odd misshapen tomato. There’s a great deal wrong with how people judge the ability of footballers: some traits are drastically overvalued, while others are hugely undervalued. Smart clubs are using analytics to identify oddly shaped fruit which they can buy for a pittance but still tastes just as flavoursome, if not more so, than beautifully formed alternatives.
The size of transfer fees being paid for footballers in the modern day has occasionally prompted a sense of unease, particularly from those outside the sport. Many believe that the relentless growth of transfer sums is unsustainable. However, analysis shows Premier League clubs, who account for roughly a quarter of global transfer expenditure, have consistently spent around 15 per cent of revenue on transfers over the last few decades. As budget lines have swelled, so have transfer fees. These figures indicate the market isn’t as ‘crazy’ as some people believe – the expenditure on playing staff is merely mirroring the finances flooding into the game from TV revenue, increased commercial deals and billionaire owners. The numbers also reveal how much the purchase or sale of a player can impact a team’s financial state. Clearly getting these decisions right is crucial to the success of a club, particularly those operating on restricted budgets.
The four key revenue drivers for teams are broadcasting rights, match-day ticketing, commercial activities, and the transfer market. The only way to materially grow the first two in this list is to achieve promotion and the third is closely aligned with the size of the fanbase, which is also largely dependent on the performance of the team on the pitch both historically and in the present day. The fourth of these, player trading, is the only one fully in the dominion of the club. Every team has the ability to become smarter and make better recruitment decisions. In a sport where 75 to 85 per cent of the explanation of results can be put down to the money a team spends on their first-team squad, clubs who can learn to control the transfer market gain a massive advantage. Analytics has a massive part to play in unveiling the true ability of footballers and helping teams bring in vital revenue through acquiring, developing and selling players.
But player trading is difficult. Dr Ian Graham, former Director of Research at Liverpool and an integral part of their recent success, estimates that half of all transfers fail. Graham outlined the six ways a new recruit might flop:
The player isn’t as good as you thought.
The player doesn’t fit your style.
The player is played out of position.
The manager doesn’t like the player.
The player has injuries or personal problems.
A player already in the squad ends up being better.
There are obviously ways to mitigate the risk of these factors, but even if you’re 90 per cent sure of success in each of these areas, that means a transfer only has a 53 per cent chance of succeeding (0.9 x 0.9 x 0.9 x 0.9 x 0.9 x 0.9). The difficulty in getting transfers right, combined with the importance of transfers on the success and financial standing of a team, means making smart recruitment decisions is essential. ‘Player recruitment and retention is the most important application of analytics by a factor of 10,’ said Graham. ‘It’s where the action is at.’
Data can help accurately benchmark the ability of the teams you wish to buy players from, making you more comfortable that their talent can make the step up. It can prevent us from overpaying for unimportant attributes or qualities. It can help us better understand age profiles and identify depreciating assets. Even simple data is often overlooked when deciding whether to buy or sell a player. For example, ‘minutes played’ is a metric hardly ever mentioned when discussing the quality of a footballer. How often is a player injured? Availability is both one of the most important and one of the most forgotten stats. We might salivate over a player generating 0.90(xG/90), but this output is next to useless if he is only able to play 90 minutes each season. And buying talent is only half the story. Spotting when to sell a player and extracting value for your deflating assets is equally as important as finding bargains.
Brighton & Hove Albion are perhaps the most well-known masters of the transfer market. Their market-leading recruitment methods based on the chance-creation data collected by Starlizard have allowed them to consistently identify hidden gems. A host of players, many of whom are outlined in Figure 7.1, were undervalued prospects who were picked up by Tony Bloom’s company’s model and signed on the cheap. The club’s bold approach to selling players has then allowed them to make large profits, before reinvesting that money in a new array of undervalued talent. The eight players in the table were bought for a combined £77.5 million and sold for nearly five times that figure. The £280.5 million profit made on these players represents a windfall roughly 70 per cent greater than a team tends to make from winning the Champions League. This graphic only goes some way to outlining the Seagulls’ transfer-market brilliance. A host of talented players are still currently playing for the team at the time of writing, including the likes of Evan Ferguson, Pervis Estupiñán, Kaoru Mitoma, Julio Enciso, Simon Adingra and Facundo Buonanotte. These players are all now valued several times higher than the fees the club paid for them.
Figure 7.1: Brighton & Hove Albion’s Transfer Market Activity
In the summer of 2018, Tony Bloom was looking to acquire another team. His requirements were that it should be in Europe but easily accessible from London, relatively inexpensive, have the potential for growth, and possess a welcoming fanbase. They landed on a small Belgian club called Royale Union Saint-Gilloise. Its stadium sits on the corner of a picturesque park and houses just 9,000 fans. The club resided in the second tier and was dwarfed by close neighbours Anderlecht, the Belgian giants who boasted 34 league titles, nine cups and five European trophies. Royale Union Saint-Gilloise had spent the previous half-century bouncing around the lower leagues and had a smaller budget than many amateur teams. Bloom spotted a club which was surviving, and hoped his methods could make it thrive.
Bloom’s xG data showed that the 5-0 match had actually been very close.
Recruitment was always going to be the key to their success. The club suddenly had access to Starlizard’s pool of data and analytical services, the same way Brighton did. Success didn’t come immediately to Union. They made mistakes along the way, but the adoption and refining of the recruitment methods outlined in this chapter saw the club’s fortunes change on its head. The club were promoted back to the first division in 2021. At the beginning of their first season back in the top flight, Union were handed odds of 500/1 to win the league. In the previous season Anderlecht had beaten them 5-0 in the cup. ‘The effect that it had was it put a load of teams off our players and off our team,’ said Alex Muzio, Bloom’s business partner and chairman of Royale Union Saint-Gilloise. ‘They thought we were bad.’ This turned out to be a blessing in disguise as it meant Union could keep their squad intact. In reality, Bloom’s xG data showed that the 5-0 match had actually been very close. Anderlecht had scored with virtually every shot they’d taken. Starlizard’s algorithms reckoned Union were much better than the pundits, bookmakers and fans were giving them credit for. They were right. Royale Union Saint-Gilloise ended up five points clear of second-place Club Brugge and 13 points clear of third-place Anderlecht in the regular season. The transfer policy espoused by Bloom has succeeded at various clubs in different levels and different leagues. Their blueprint is one every club should be copying.
Even clubs at the other end of the financial spectrum to Brighton and Royale Union Saint-Gilloise require constant transformation of their playing staff. For a while, retaining the Premier League title proved impossible – no club managed to do so in the nine seasons between 2009 and 2018. After that spell, Manchester City won it five times in six seasons. Yes, the depth of their owners’ pockets played a large part, but so did their attitude to refurbishing the squad after every season – even (and perhaps especially) after successful ones. Pep Guardiola recognised that if you stand still for too long, you will be caught up. He’s allowed talented players like Gabriel Jesus, Raheem Sterling, Ferran Torres and Leroy Sané to leave the club, bringing in hungry new talent to replace them. He’s also been willing to let go of older players like Sergio Agüero, Vincent Kompany and İlkay Gündoğan sooner than other clubs might have done, not allowing himself to be nostalgic and keep players simply because of their past achievements. What matters is their quality in the present moment. The previous side to retain the Premier League title before Manchester City were Manchester United under Sir Alex Ferguson in 2008/09, a team who displayed a similarly aggressive approach to player trading and a similar desire to constantly revamp the playing staff. Ferguson was always on the look-out for signs of decay and took no prisoners. Players including Mark Hughes, David Beckham, Paul Ince, Jaap Stam, Ruud van Nistelrooy and Roy Keane were all released with surprising haste. Premier League winners cannot rest on their laurels; the teams chasing them certainly won’t.
The transfer market, just like any other market, is about buying and selling. Football fan forums, social media feeds and gossip columns tend to focus on the former of these activities. Supporters get far more excited by potential arrivals than potential outgoings, as shown by the modern trend of ‘announcement videos’ that have spawned in recent seasons. ‘In the know’ journalists like Fabrizio Romano have found huge success by being the first to reveal who clubs are looking to sign. There is something of a journalistic arms race to be the first one to announce new arrivals to hordes of baying supporters. All the hype centres around the procurement of footballers. But selling players at the right time and the right price is equally as important as purchasing new ones. Let’s first turn our attention to how intelligent clubs dispense of their talent.
Goalscorer Inflation
Michu had reached god-like status. Arriving at Swansea City for just £2 million from cash-strapped La Liga side Rayo Vallecano in 2012, the striker was previously unheard of in the English game. Michu announced himself on his Premier League debut against QPR by scoring two goals and assisting another. He ended up scoring 22 goals in the 2012/13 season, including one in Swansea’s victorious League Cup final against Bradford City. His performances even led to a first cap for the Spanish national side. Following his break-through campaign, Michu was linked with a move to some of the giants of the sport. Arsenal reportedly submitted a £25 million bid, while Liverpool also enquired about his services. But Swansea refused to accept any offer less than £30 million, a price no club was willing to match.
Michu’s second season in South Wales started less well. A handful of mediocre performances were followed by an ankle injury. When he returned to the field he couldn’t replicate the form he had shown in 2012/13. He was loaned out to Napoli, where he flopped. The Siren’s Song, one of Napoli’s most popular blogs, described Michu as ‘just plain bad’. The Spaniard is now the unfortunate poster boy for a group of players known as ‘one-season wonders’.
Michu fell victim to what is known as ‘regression to the mean,’ another statistical phenomenon which originated with Francis Galton. When he wasn’t eliciting the wisdom of the crowds to measure the weight of oxen, Galton was investigating the way personal characteristics change between generations. Using their parents’ height, can we predict an adult offspring’s height? Galton had the classic Victorian scientist’s obsessive interest in collecting data. He compared 465 sons’ height to their fathers’ height and made an interesting discovery. Tall fathers tended to have smaller but still taller-than-average sons, while short fathers tended to have taller but still small-than-average sons. Galton called this ‘regression to mediocrity,’ but it’s now known as ‘regression to the mean’.
If Michu’s 2012/13 campaign was a 6-feet-6-inch father, his 2013/14 campaign was always likely to be a shorter son. The Spaniard had enjoyed an extraordinary season and was always likely to regress to the mean. In fact, even his performances at Swansea during his miracle campaign hinted at an eventual decline; Michu wasn’t a dazzling performer, his work-rate off the ball was poor and he amassed few assists. But the footballing world became engrossed by him because of the sheer number of goals he was scoring. Goals are the ultimate currency in football, but their randomness can lead to players becoming grossly and irrationally overhyped. Arsenal and Liverpool dodged a bullet by not agreeing to sign Michu for the £30 million Swansea were demanding.
Players who score lots of goals one season aren’t necessarily likely to repeat the feat the next. Figure 7.2 shows the average regression for each ‘bucket’ of goals scored in a Premier League season by a player between 2004 and 2023. For example, players who scored 15 goals in a campaign went on to score an average of 7.1 goals the next season, a regression of -7.9 goals. The sample includes all 112 players who scored 15 goals in a season and went on to play at least 900 minutes the next campaign over this 19-year time period. Of these 112 players, only 17 went on to improve on their scoring tally the following season (15 per cent). The average regression for a player who scored 15 goals or more was -6.2 and overall the cohort of 112 players scored 669 fewer goals the season after their 15+ campaign.8
Figure 7.2: Regression in Goalscoring, Premier League 2004–2023
The majority of the 17 instances when an improvement was made season-on-season were when players approached the peaks of their powers. The best progression was made by a 22-year-old Cristiano Ronaldo, who went from 17 goals in 2006/07 (he was 22 at the end of this season) to 31 in 2007/08, while a 26-year-old Luis Suárez went from 23 in 2012/13 to 31 in 2013/14, and a 25-year-old Sergio Agüero went from 17 in 2013/14 to 26 in 2014/15. Harry Kane improved his goals tally for three consecutive seasons from 2014/15 to 2017/18 as he became one of the world’s most prolific marksmen. Kane also improved season-on-season from 2018/19 to 2019/20 and then again to 2020/21. He’s ‘progressed’ rather than ‘regressed’ six times over the sample period, meaning Kane alone accounts for more than one third of the player seasons in this sample to have improved year-on-year. If you strip Harry Kane from the study, only 11 of the remaining 104 campaigns (11 per cent) actually show improvement versus the prior season.
