Football scouts used to trust intuition. Now they also use spreadsheets
Sean Ingle
The Guardian, Sunday 20 January 2013 22.42 GMT

From Manchester United to Real Madrid, clubs are using more and more stats to try to get an edge in the transfer market

Jack Hixon was a Boy's Own football scout made flesh, with cheeks the colour of glacé cherries and weathered tales of big fish and the ones that got away. His greatest discovery was Alan Shearer, who he spotted as a 13-year-old in a park and took to Southampton – one of 47 players he put on the path to top-flight football in his 50-year career. I watched a game with him once, back in the late 90s, hoping to mine his secrets. But there were none, save for hard work, instinct and an eye for a player as sharp as a stud.

I thought of Jack last week, when up close with the analysts at OptaPro and Prozone; data companies that are de facto scouts in the digital age. I suspect Jack never owned a computer, even though he was still working for Newcastle United when he was 86, two years before his death in 2009. He didn't need to. It was all in his head and his contacts book.

Times have changed. Now almost everything that happens on the pitch, save for phlegm being flung from players' left nostrils, is codified across a vast number of leagues. And, subtly but inevitably, the scout and the spreadsheet – never natural soulmates – have started to squeeze up tight, like new lovers in a single bed.

As John Coulson, the head of professional football services at OptaPro – which works closely with clubs including Manchester City and Chelsea – puts it: "The biggest area we're involved with now is player recruitment. No team will sign a player based on data alone, but it's increasingly a shortcut to a shortlist."

Some clubs buy raw data and stay schtum, keeping their algorithms in house. Others use outside analysts as the first stage of the recruitment process, with a manager setting out the type of player he wants and getting a list of suggested targets in return.

As Prozone's business development director Blake Wooster, who counts Real Madrid and Manchester United among his clients, explains: "It's like when Amazon tells you other books you might like after a purchase. A coach might not have heard of a player in the Polish second division – but he might have similar attributes to the guy he's looking at in League One. We are just increasing the due diligence process."

Some in football's food chain remain sniffy about data, although such attitudes often dissipate the higher up you go. To some extent that is understandable. Some of it is as irrelevant to a team's performance as the colour of their jerseys. How far a player runs in a match, so beloved of Uefa's Champions League coverage, tells you little when used in isolation. Similarly, a 90% pass completion rate is probably less impressive if a ball is continually tapped five yards sideways. Context is everything.

But when it comes to the January transfer window – football's irrational global hypermarket – clubs are realising that using advanced-data analytics is sensible soccer-nomics. Even the biggest clubs find it impossible to assess every potential target, let alone the Championship manager hunting for bargains to keep his team up.

The analytics, for instance, suggested that Michu was hugely undervalued at Rayo Vallecano. But data remains a starting point, not a solution. As OptaPro's Simon Farrant admits: "We can't tell you how a player behaves off the pitch. But when clubs want to know whether a target could do a certain job, that's something we can answer."

The vastness of the data, much of it not in the public domain, is a nerd's paradise. Players' movements can be tracked every 10th of a second, while everything that happens on the ball – around 2,000 events in a match – is collated. A pass is not simply a pass: an analyst will plot where it went, whether it was driven or chipped, played as a through-ball or to feet, which foot it was kicked with, and more.

Such detail allows the technical scout to delve in, dissect, play detective. He might watch a game and note that a midfielder is more confident passing to his left than right. By pooling that player's data across many games – and controlling it for the state of each match and different formation and tactics – the scout can assess whether it's a blip or a trend.

Similarly, a striker's goals might not provide the complete story of how he has performed. Plotting actual goals against what was statistically expected – given the quality of chances and strength of the opposition – provides greater context and a potential bargain.

So why the ongoing suspicion about data? Some of it, surely, is down to clubs not going public about what they have learned – understandably so, given how valuable an edge might be. As a result some remain sceptical that analytics can ever apply to a sport as fluid as football. Mentioning Moneyball remains one of the best ways of silencing a pub full of fans.

But scouting has radically changed since the days when Hixon trawled exposed terraces and school fields for potential stars of the future. It's now about numbers and eyeballs, raw data and intuition, processors and people. And you sense this is just the start.


http://www.guardian.co.uk/football/blog/2013/jan/20/football-data-scouts-michu-transfer

Transfer window: The perils of recruitment in the Premier League
By Alistair Magowan BBC Sport

If ever a club offered a cautionary tale about the limitations of simply throwing money at a problem, then it is Queens Park Rangers this season.

Since the club won promotion to the Premier League 19 months ago under Neil Warnock, 23 players have been signed by Warnock and successor Mark Hughes - both since dismissed - yet the club survived only by the skin of their teeth last season and are bottom this time round, having taken 17 games to record their first league win.


“For some signings you don't do as much due diligence as you would like and take an educated gamble”

Dan Ashworth WBA technical director

New boss Harry Redknapp will now expect to put his own stamp on the team in the January transfer window, but his predecessors' experiences demonstrate that mass recruitment has its dangers and no amount of preparation can guarantee a player will be suited to the rigours of England's top flight, whatever their credentials.

QPR defender Ryan Nelsen noted as much before Redknapp was brought in.

"I'm not a big fan about talking about talent among a squad," said the New Zealander. "When the pressure is on, you can be the most talented player in the world but when the microscope goes on you, talent goes out the window."

So just how much planning can a club do to ensure the player they sign is the right one?

At a discussion chaired by sports data company Prozone at Manchester City's Etihad Stadium recently, the short answer seemed to be - plenty.

Prozone logs specific data for activities such as passing, dribbling, shots and saves in matches across 37 countries, covering 250 clubs and 80 competitions and Nottingham Forest boss Alex McLeish, who has also managed at Aston Villa, Birmingham and Rangers, told BBC Sport: "The biggest thing is that it saves you hopping on a plane to South America to watch a player when you can see videos up close with statistics to back it up."
QPR's signings since May 2011

Neil Warnock: Joey Barton, Shaun Wright-Phillips, Jay Bothroyd, Anton Ferdinand, Luke Young, Armand Traore
Mark Hughes: Federico Macheda (loan), Taye Taiwo (loan), Nedum Onuoha, Bobby Zamora, Djibril Cisse, Jose Bosingwa, Ryan Nelsen, Park Ji-sung, Robert Green, Andy Johnson, Junior Hoilett, Fabio, Samba Diakite, Esteban Granero, Julio Cesar, Sam Magri, Stephane Mbia

But the likes of McLeish, West Brom's outgoing technical director Dan Ashworth and GB Cycling performance director David Brailsford - named coach of the year at this month's BBC Sports Personality of the Year awards and knighted in the New Year Honours list - also agreed with Nelsen's comments that accurately accounting for a player's character is still uncharted territory.

So services like Prozone are not the complete answer, as McLeish admitted.

"Looking back, sometimes you rush signings that weren't the right choice for that particular club," he said. "You can have all the statistics in the world on a player but the most difficult part of the recruitment is when the individual is not of the same mentality. That's when you find it really hard."

Understandably, McLeish is loath to name his "mistakes" in the transfer market, preferring to point to the likes of Scott Dann, Roger Johnson and Barry Ferguson as successful acquisitions during his time at Birmingham.

The Scot also revealed that at Birmingham he was offered the now West Brom striker Peter Odemwingie for £7m before their local rivals bought him from Lokomotiv Moscow in August 2010. McLeish explained: "In his last three years he was being played as a wide man and we got a false impression that he had regressed a little bit."

Given that Odemwingie's goals helped keep West Brom in the Premier League the following season, it highlights the need to marry "gut feelings" with statistics, McLeish says, and underlines the uncertainty of the transfer market.

So difficult is it that West Brom technical director Ashworth, heading to the Football Association in a similar position in the summer, reckons player recruitment is the area where he spends most of his time - albeit, on recent evidence, very successfully.

The former Baggies academy director employs four full-time staff in recruitment plus part-time scouts, and has overseen signings that prove his system is working. Midfielders Youssuf Mulumbu and Claudio Yacob, plus defender Jonas Olsson have all impressed at The Hawthorns despite costing less than £1m combined.
Dan Ashworth's key signings for West Brom
Peter Odemwingie, Shane Long and James Morrison

Shane Long, Ben Foster, Yussouf Mulumbu, Jonas Olsson, Steven Reid, Gareth McAuley, Billy Jones, Claudio Yacob

But Ashworth looks back on his coaching appointments at Albion with most pride, having brought Roberto Di Matteo, Roy Hodgson and Steve Clarke to the club to work under him in a more European-style structure.

He is careful to point out that, even though his staff work tirelessly to capture the right players, it is the head coach who makes the final call.

The 41-year-old admitted: "Some signings have been a bit lucky, some you have a gut feeling about where perhaps you don't have the time, and with others the opportunity comes up where you didn't know that they would become available.

"So for some of them you don't do as much due diligence as you would like and take an educated gamble. Others will be assessed for a period of time and we will watch them several times, use reports and DVDs and have meetings about them."

Under Ashworth's guidance, West Brom have resisted the sort of wholesale changes seen at QPR.

"In any team sport there has to be a period of time where players can get used to each other and the manager has to understand how the team functions together to work out what his best team is," he explained. "Unfortunately in the Premier League, managers don't tend to get that sort of time."

Cycling chief Brailsford might appear an outsider to the football world, but says he has spoken with several football chairman and is mischievously coy about a potential future in football.

In any event, he is adamant that his dedication to sustained improvement at both British Cycling and Team Sky can be replicated in any field, including football.

Moreover, the 48-year-old believes that football has statistical information available that is superior to cycling - although he warns it could lead to mistakes.

"If you understand the demands of the event and therefore what you are looking for then there is more chance you can then trawl a massive amount of data to find what you need," Brailsford said.

"If you are not sure what you are looking for and you are just going into the data to see what you can find without an absolute clear end-point, then you are going to be way more challenged.

"The perfect scenario is where there is a great dialogue between those on the recruitment side and the analytics and sports science side so it's very much working together. No one element provides the full answer - that lies someone between all three or four parts of the information system coming together.

In front of a captivated audience of Premier League and Football League officials, Brailsford concluded: "Football is so task-focused and you are trying to get the task to be successful so quickly that everything becomes about delivering in the here and now and making that happen."

No wonder, then, that a desperation to improve can sometimes lead to the wrong blend of players - as QPR have found out.

http://www.bbc.co.uk/sport/0/football/20518689

Edited by dr. bellows: 23/1/2013 12:06:47 AM

Great articles Damo and Doc!

Some interesting material there. There are still a lot of sceptics about stats and player data.

Reminds me of that recent Clint Eastwood movie, "Trouble With the Curve". Eastwood plays an ageing scout for a major baseball team. He spends most of his time travelling round the country, watching minor league games in order to spot new talent. Someone tells him that he is wasting his time, and he could do it all from home- by looking up stats on the internet. Apparently, every minor league player in the country has his detailed stats up on the net.

Eastwood ignores this, and keeps attending matches in person, and never looks at stats on the net. Eventually, he spots a flaw in the game of a star batter that his club wants to sign (Nobody has seen him play, they have just looked at his stats.)

The player is brought to the club for a trial... but flops badly. Apparently he can't handle a certain type of pitching. There were no stats that predicted this. Old fashioned methods are vindicated (or is it all just moonshine?:lol: )