This article was in the FT yesterday but you have to pay for their content so rather than linking it, I've reproduced it in more or less it's entirety (there were a few graphics that I may try to add in a bit but the meat of the article should all be here). If there are any formatting issues, I'll try to fix them asap.

Every tiny aspect of a football match can now be recorded and scrutinised. FT Weekend Magazine commissioned artist Giles Revell to create a series of images of the recent Champions League Final between Barcelona and Manchester United, using exclusive data extracted from the game by the analysis company Prozone

I recently visited Manchester City’s tranquil training ground in the village of Carrington. It was a glorious sunny morning, and outside the gates hired hands were washing footballers’ SUVs and sports cars. The defender Kolo Touré coasted past in a giant black contraption straight out of The Godfather. Carrington is used to cars like that: Manchester United train in the village too.

“Abu Dhabi Travellers Welcome”, said the message on the façade of City’s sky-blue training centre. Abu Dhabi’s ruling family owns Manchester City, and one thing it has done since buying the club is hire a large team of data analysts. Inside the building I found Gavin Fleig, City’s head of performance analysis, a polite sandy-haired man in a neat black City sweater. Hardly anyone outside Carrington has heard of him, and yet Fleig is a prime mover in English football’s data revolution. Largely unseen by public and media, data on players have begun driving clubs’ decisions – particularly decisions about which players to buy and sell. At many clubs, obscure statisticians in back-rooms will help shape this summer’s transfer market.

Fleig gave me the sort of professional presentation you’d expect from a “quant” in an investment bank. Lately, to his excitement, City had acquired stats on every player in the Premier League. Imagine, said Fleig, that you were thinking of signing an attacking midfielder. You wanted someone with a pass completion rate of 80 per cent, who had played a good number of games. Fleig typed the two criteria into his laptop. Portraits of the handful of men in the Premier League who met them flashed up on a screen. A couple were obvious: Arsenal’s Cesc Fàbregas and Liverpool’s Steven Gerrard. You didn’t need data to know they were good. But beside them was a more surprising face: Newcastle’s Kevin Nolan. The numbers wouldn’t immediately spur you to sign him. But they might prompt you to take a closer look.

In recent years, after many false starts, the number-crunchers at big English clubs have begun to unearth the player stats that truly matter. For instance, said Fleig, “The top four teams consistently have a higher percentage of pass completion in the final third of the pitch. Since the recruitment of Carlos Tévez, David Silva, Adam Johnson and Yaya Touré to our football team, in the last six months alone, our ability to keep the ball in the final third has grown by 7.7 per cent.”

That stat had not necessarily driven their recruitment, Fleig cautioned. Indeed, there are probably clubs that lean far more on stats than Manchester City do. I recently toured several actors in football’s data revolution, and was struck by how far it had progressed. “We’ve somewhere around 32 million data points over 12,000, 13,000 games now,” Mike Forde, Chelsea’s performance director, told me one morning in February in the empty stands of Stamford Bridge. Football is becoming clever.

Probably ever since the personal computer arrived, a few pioneers in football have tried to use data to judge players. Among the first was Arsenal’s future manager, Arsène Wenger, an economics graduate and keen mathematician. In the late 1980s, as manager of Monaco, Wenger used a computer program called Top Score, developed by a friend. A less likely pioneer was the late, great vodka-sodden Ukrainian manager Valeri Lobanovski. When I visited Kiev in 1992, Lobanovski’s pet scientist, Professor Anatoly Zelentsov, had me play the computer games that Dynamo Kiev had developed to test players. When Lobanovski said things like, “A team that commits errors in no more than 15 to 18 per cent of its actions is unbeatable,” he wasn’t guessing. Zelentsov’s team had run the numbers.

But the broader breakthrough came in 1996, after the Opta Index company began collecting “match data” from the English Premier League, explains the German author Christoph Biermann in Die Fussball-Matrix, the pioneering book on football and data. For the first time, clubs knew how many kilometres each player ran per match, and how many tackles and passes he made. Other data companies entered the market. Some football managers began to look at the stats. In August 2001 Manchester United’s manager Alex Ferguson suddenly sold his defender Jaap Stam to Lazio Roma. The move surprised everyone. Some thought Ferguson was punishing the Dutchman for a silly autobiography he had just published. In truth, although Ferguson didn’t say this publicly, the sale was prompted partly by match data. Studying the numbers, Ferguson had spotted that Stam was tackling less often than before. He presumed the defender, then 29, was declining. So he sold him.

As Ferguson later admitted, this was a mistake. Like many football men in the early days of match data, the manager had studied the wrong numbers. Stam wasn’t in decline at all: he would go on to have several excellent years in Italy. Still, the sale was a milestone in football history: a transfer driven largely by stats.

At Arsenal, Wenger embraced the new match data. He has said that the morning after a game he’s like a junkie who needs his fix: he reaches for the spreadsheets. In about 2002 he began substituting his forward Dennis Bergkamp late in matches. Bergkamp would go to Wenger to complain. “Then he’d produce the stats,” Bergkamp later recalled. “‘Look Dennis, after 70 minutes you began running less. And your speed declined.’ Wenger is a football professor.”

Few would suspect it of West Ham’s new manager “Big Sam” Allardyce, and yet his somewhat neolithic appearance also conceals a professorial mind. As a player, Allardyce spent a year with Tampa Bay, Florida, where he grew fascinated with the way American sports used science and data. In 1999 he became manager of little Bolton. Unable to afford the best players, he hired good statisticians instead. They unearthed one particular stat that enchanted Allardyce. “The average game, the ball changes hands 400 times,” recites Chelsea’s Forde, who got his start in football under Allardyce. “Big Sam” would drum it into his players. To him, it summed up the importance of switching instantly to defensive positions the moment the ball was lost.

More concretely, stats led Allardyce to a source of cheap goals: corners, throw-ins and free kicks. Fleig, another Allardyce alumnus, recalled that Bolton would score 45 to 50 per cent of their goals from such “set-pieces”, compared with a league average of about a third. Fleig said, “We would be looking at, ‘If a defender cleared the ball from a long throw, where would the ball land? Well, this is the area it most commonly lands. Right, well that’s where we’ll put our man.’”