Inside the Data Models That Let Brentford and Brighton Outsmart Bigger Clubs

In August 2023, Chelsea paid Brighton and Hove Albion 115 million pounds for a 21-year-old Ecuadorian midfielder named Moises Caicedo, a British transfer record at the time. Brighton had signed him from Independiente del Valle in Ecuador for a small fraction of that fee. The deal was not luck. It was the output of a recruitment model built around statistical modeling that started, decades earlier, in the sports betting industry rather than in football at all.

Brighton and its Premier League neighbor in ambition, Brentford, have become the two most cited examples of clubs using data to compete with rivals that outspend them many times over. Both stories trace back to owners who made their fortunes finding inefficiencies in betting markets before turning the same tools on the transfer market.

Where Brentford's model started

Matthew Benham, Brentford's owner, studied physics at Oxford before moving into finance in the City of London. He later founded Smartodds, a statistical research firm serving professional gamblers, after backing a scoring-probability model built by researchers Stuart Coles and Mark Dixon at Lancaster University that outperformed bookmakers' own odds. Smartodds now compiles data on more than 85,000 players worldwide, a database originally built for betting clients that Brentford's football staff use to price transfer targets the way Benham once priced match odds.

Ahead of the 2015-16 season, Brentford committed to identifying undervalued players with that data, developing them, and selling on for profit rather than trying to outbid richer clubs for the finished article.

The results on the balance sheet

The numbers back up the approach. Brentford bought striker Ollie Watkins from Exeter City for around 1.8 million pounds in 2017 and sold him to Aston Villa for a club-record 28 million pounds three years later. They signed Ivan Toney from Peterborough United for roughly 5 million pounds and sold him to Saudi club Al Ahli for 40 million pounds in 2024, at least eight times the purchase price.

Operating with one of the smaller budgets in English football's top two divisions, Brentford climbed from League One to the Premier League by 2021, its first season in the top flight in 74 years.

Tony Bloom's version at Brighton

Brighton's chairman, Tony Bloom, followed a parallel path. A mathematics graduate from Manchester University and a former financial trader, Bloom founded the sports betting research firm Starlizard in 2006 before applying its analytical methods to Brighton's recruitment. The club's model, sometimes referred to internally by the codename Starlizard or Jamestown, is built to surface talent in regions traditional European scouting networks cover thinly, particularly South America and Eastern Europe.

The commitment to data went further than most rivals were willing to go. In November 2024, Brighton removed the majority of its traditional scouting staff, choosing to lean almost entirely on the analytics software to generate transfer targets rather than treat data as one input among several.

Turning academy-adjacent signings into profit

The financial return has been extraordinary relative to Brighton's size. Beyond the Caicedo sale, which delivered Brighton a net profit reported at roughly 87 million pounds once the original transfer fee, development costs and sell-on structure are accounted for, the club posted a record Premier League profit of 123 million pounds for the 2022-23 financial year after selling four players for a combined 121 million pounds. Brighton's pattern is consistent: identify a player for a modest fee, develop him with strong coaching, and sell to one of the traditional "Big Six" clubs for a large multiple.

Two more sales that show the pattern lower down the price scale

Caicedo is the headline, but the same model produced smaller, equally instructive deals. Brighton signed left back Marc Cucurella from Getafe for 15.4 million pounds in 2021 and sold him to Chelsea a year later for an initial 56 million pounds, rising to 63 million with add-ons, a club-record fee received at the time. Winger Kaoru Mitoma arrived from Kawasaki Frontale in Japan's J-League for around 2.5 million pounds in 2021, was loaned to Union Saint-Gilloise in Belgium, a club Bloom also owns, to accelerate his adaptation to European football, and returned to become a first-team regular whose market value now sits many multiples above his signing fee. The Union Saint-Gilloise loan is itself part of the model: owning a lower-cost European club gives Brighton a development pathway that data alone could not provide.

Why the model is hard to copy

The obstacle for bigger clubs is not access to data, most Premier League clubs now run analytics departments of some kind. It is that Benham and Bloom built their statistical edge over years inside the betting industry before ever applying it to football, and their clubs' entire recruitment cultures were built around trusting the model over reputation and name recognition from the start. A club with an entrenched traditional scouting hierarchy cannot simply install new software and expect the same discipline, it has to be willing to sign unheralded 19-year-olds ahead of proven names when the numbers say to, and to sell its best player at his statistical peak even when supporters want him kept.

That discipline, more than any single algorithm, is what let two mid-sized English clubs consistently beat richer rivals to undervalued talent, and it is why both clubs keep reinvesting sell-on profit straight back into the same scouting infrastructure rather than treating a big sale as a one-off windfall to bank.