By Merlion
June 6 2023

An interesting read is, “Expected Goals” (2022) by Rory Smith. The extracts discussed below show why Brighton and Brentford are so successful in their recruitment, and punch above their weight in EPL.

Michael Lewis’ book “Moneyball” has been misinterpreted on the assumption that adopting a data-led approach would inevitably not cost significant sums of money. Moneyball was about exploiting inefficiencies in the market; not paying premium fees. Matthew Benham’s Brentford and Tony Bloom’s Brighton are “Moneyball” clubs, almost exactly what the early pioneers of and advocates for the use of data in football. They recruit players from undervalued markets or with underappreciated skill sets. They spot talent where others see none. They play as style of football that has been shaped and honed by the study of vast tracts of data. Their approach has enabled them all to surpass teams with far greater revenues and much deeper pockets. They are trailblazers of what football might look like, how it might work, in the digital age.

Bloom and Benham are, generally, described as “professional gamblers”. Both of them use data analytics to give them an edge in the gambling industry. Bloom’s firm has Starlizard. Benham’s consultancy, SmartOdds, does much the same. Both have staffs that run into the hundreds assessing every single factor that might have some impact on the outcome of a game: not just form but anything from the weather to the distance the visiting team has had to travel.

Starlizard’s data makes the firm an “excellent judge of talent”. The firm has a trove of detailed information on tens of thousands of players around the world: not just rudimentary information but in-depth, complex performance analysis.

Brighton has made a habit of recruiting players from leagues where other teams are too wary to shop – competitions where data is deemed unreliable or incomplete, or where quality of competition varies so wildly that it can be hard to draw concrete conclusions – or signing talent that might have slipped by unnoticed by rivals.

That sort of data analysis in the transfer market is part of Brentford’s approach. In 2016, Brentford took the unorthodox decision to scrap its academy. Brentford had realized that, under England’s youth development rules, it was essentially nurturing players so that bigger, richer teams could come and poach them for a pittance. Instead, the club chose to establish a so-called “B Team”, largely stocked by players aged between 17 and 20: often, the ones who had been cut loose from other sides’ academy systems. Rather than being a rival destined to lose out to Chelski, Arsenal and Spurs, Brentford saw this as a chance to be something closer to a partner. In the years that followed, the B Team would provide the club with a steady supply of first-teamers, players who had developed late or improved rapidly or needed a change of environment. It had found, in other words, an inefficiency in the system, and exploited it.

Hendrik Almstadt was persuading Arsenal executives and Arsene Wenger to buy StatDNA for £2 million. The approach Almstadt took was deliberately provocative. He discussed two players Arsenal had signed in previous summers: Maroaune Chamakh, on a free transfer and Park Chu-Young for a nominal fee. Wenger had signed off both. Almstadt told him he should not have done. He had both players’ performance data at their previous clubs. It showed what Almstadt believed to be “clear red flags”. Chamakh had undeniable talent, and had scored goals consistently, but did so in streaks, and his output outside the box – how much he was involved in play, his work-rate, his ability to fold into the team’s structure – was limited. Park’s metrics were not impressive enough, even for a young player, to believe that he would soon be ready to play first-team football for a club of Arsenal’s scale and scope.

Neither had cost a vast amount of money, but both represented another player, a more fitting player, not signed. Both occupied a space in the squad. Both ate up a little of the wage bill. Both had been poor choices, and both had been entirely avoidable. Or rather, both would have been avoidable had Arsenal used data as a matter of course in their recruitment process.

Data most immediate use, in a football context, was to outline what not to do. Almstadt said, “Stupid player transactions kill clubs. It helps you not to do stupid stuff.”

The problem lies with Arsene Wenger who never spent money easily. As Arsenal tried to offset Emirates Stadium building cost, Wenger had developed an especially parsimonious streak. Wenger saw his job as maintaining Arsenal’s presence in the Champion League, while working on a strict budget, and enduring and regular departure of some of his finest players. He needed to find bargains like Chamakh and Park, to take risks, and had little or no margin for error. He agreed to buy to buy StatDNA for £2 million.

Why did Arsenal’s fail to take advantage of the head start the club had been granted by the arrival of StatDNA? Arsenal was ahead of the curve with StatDNA. Unfortunately, by the time StatDNA arrived, the club was already deep into the aimless, listless malaise of late-era Wenger, waiting to find out what the future might be. Decisions were made in a spirit of uneasy consensus, while the parsimony of the Kroenke family, once they had complete control, meant that Arsenal could not be sure of signing players the club really want.

 It was not an ideological resistance to data nor suspicion to those StatDNA nerds telling lifelong scouts who they should be signing. It was, instead, something far more mundane: self-preservation.

Traditionally, coaches and scouts as a group in possession of “protected knowledge”. They alone knew the formula for either identifying or creating talented players; they knew it because of the precise and personal experiences and backgrounds. It could not be shared or acquired. That knowledge gave them their value. It gave them their jobs. And why Almstadt was accused of “trying to take someone’s job”. To Edu’s and Arteta’s credit, they were ruthless enough to dismantle this Arsenal out-of-date scouting network and re-start again.

By 2015, StatDNA was known as Arsenal Data analytics. Data did not transform Arsenal into champions. As counter-intuitive as it is, Almstadt believes that StatDNA did its job by ensuring the wrong players did not arrive, as much as making sure the right player did. “If you look at what happened afterwards with £75 million paid for Nicolas Pepe, that was not happening when StatDNA was at the centre of everything”, he said. (A solid explanation as to how Pepe came to be signed is not really provided).

In 2017 Sven Mislintat had been appointed to oversee the club’s recruitment department. He was an analytics devotee and had developed his own platform, Matchmetrics. He was a victim of a schism with Raul Sanllehi and lasted less than a year. His influence remained; the analytics used by the club’s scouting department were different from those offered by Arsenal Data Analytics. Mikel Arteta’s solution to that problem was not to unify the two teams, but to throw in a third: he brought in Lee Mooney, a former colleague at Manchester City, as a consultant.

At Arsenal, it seems that every signing seemed to be a “compromise” not just economically but politically: the club would frequently move for a player that nobody objected to, rather than the one that a specified department advocated.

Arteta will need to complete the 2022/23 data analysis, taking into account all sorts of data: the number of chances a team was creating, the quality of those chances, the amount and the nature of opportunities it was conceding, and whether the club’s model showed that the club was the best team in EPL. That did not just mean that the team had been a little unlucky, a little profligate. Plus, to study whether there are any common characteristics amongst players who had succeeded in 2022/23 season, and determine what sort of traits they should be searching for when recruiting, will Mikel Arteta be as successful as Bloom and Benham in recruiting the correct players that will strengthen his 2023/24 squad to challenge for the title again.