Getty
Show Hide image

So much for "the table never lies" – data unravels football's biggest lie of all

London side Brentford FC are using data to rethink the usual football club model.

It’s a miserable day for practice, the rain spitting down on the manicured training pitches of Brentford Football Club. Inside a tiny office marked Director of Football, Rasmus Ankersen is waiting for his phone to ring. The winter transfer window closes in 11 hours and there are deals to finalise.

Ankersen, a 33-year-old Dane with a trim beard and hair pulled into a small ponytail, seems relaxed. Perhaps he knows that the £12m transfer of the striker Scott Hogan to Aston Villa is as good as done. Or maybe his comfort comes from Brentford’s performance this season. The small west London club sits safely in the top half of the second tier of English football – at least according to management’s own version of the league table, which is based on “deserved” rather than actual results. Officially, on 31 January, when we meet, the team is 15th of 24.

“There’s a concept in football that the table never lies,” says Ankersen, whose own playing career was ended by a knee injury in his teens. “Well, that’s the biggest lie in football. Your league position is not the best metric to evaluate success.”

Brentford are an outlier in English football. Since the professional gambler Matthew Benham bought a majority share in 2012, they have relied on the scientific application of statistics – the “moneyball” technique pioneered in baseball – when assessing performance.

The early results were positive. In 2014, Brentford were promoted from League One to the Championship and the next season finished fifth. That same year, Benham’s other team, FC Midtjylland, which is run on similar principles, won the Danish Superliga for the first time.

Yet in 2016 Brentford slipped to ninth. Despite the disappointing season so far, Ankersen insists the strategy is the right one for “a small club with a small budget”.

Underpinning Brentford’s approach is the understanding that luck often plays a big part in football. “It is a low-scoring sport, so random events can have a big impact,” Ankersen says. “The ball can take a deflection, the referee can make a mistake. The best team wins less often than in other sports.”

In a match, or even over a season, a team can score fewer or more than its performance merits. A famous example is Newcastle in 2012, says Ankersen, who besides his football job is an entrepreneur and author. In his recent book, Hunger in Paradise, he notes that after Newcastle finished fifth in the Premier League, their manager, Alan Pardew, was rewarded with an eight-year extension of his contract.

If the club’s owners had looked more closely at the data, they would have realised the team was not nearly as good as it seemed. Newcastle’s goal difference – goals scored minus goals conceded – was only +5, compared to +25 and +19 for the teams immediately above and below them. Statistically, a club with Newcastle’s goal difference should have earned ten points fewer than it did.

Moreover, its shot differential (how many shots on goal a team makes compared to its opponents) was negative and the sixth worst in the league. That its players converted such a high percentage of their shots into goals was remarkable – and unsustainable.

The next season, Newcastle finished 16th in the Premier League. The team was not worse: its performance had regressed to the mean. “Success can turn luck into genius,” Ankersen says. “You have to treat success with the same degree of scepticism as failure.”

Brentford’s key performance metric is “expected goals” for and against the team, based on the quality and quantity of chances created during a match. This may give a result that differs from the actual score, and is used to build the alternative league table that the management says is a more reliable predictor of results.

Besides data, Brentford are rethinking the usual football club model in other ways. Most league clubs run academies to identify local players aged nine to 16. But Ankersen says that this system favours the richer clubs, which can pick off the best players coached by smaller teams.

Last summer, Brentford shut their academy. Instead, they now operate a “B team” for players aged 17 to 20. They aim to recruit footballers “hungry for a second chance” after being rejected by other clubs, and EU players who see the Championship as a stepping stone to the Premier League.

It’s a fascinating experiment, and whether Brentford will achieve their goal of reaching the Premier League in the near future is uncertain. But on the day we met, Ankersen’s conviction that his team’s fortunes would turn was not misplaced. That evening, Brentford beat Aston Villa 3-0, and moved up to 13th place in the table. Closer to the mean.

Xan Rice is Features Editor at the New Statesman.

This article first appeared in the 16 February 2017 issue of the New Statesman, The New Times

Photo: Getty
Show Hide image

The Universal Credit nightmare shows there’s nothing more dangerous than a good idea

The hardest thing to build into any benefits IT project is common sense.

The trouble with Universal Credit is that everyone thinks it’s a good idea. Labour has long backed the concept of rolling multiple benefits into one payment but studiously refused to implement it when in power. Why? Because it takes all the mess and complication that claimants have to navigate and transfers that to the government. It’s like Whitehall volunteering to find your next house, sort out the survey and fix the best mortgage for you. It sounds brilliant – and that should make you suspicious.

“I think it’s quite a good idea, having it all in one go,” says Jo Whitaker when I speak to her at home in Moulton, North Yorkshire. Unfortunately, the reality fell short. Diagnosed with breast cancer in late 2016, Whitaker had to give up her cleaning business as she underwent chemotherapy. She was told – oh, happy day! – that her local jobcentre was one of those testing Universal Credit ahead of its countrywide roll-out.

There was a catch. In order for her to claim Universal Credit, her existing child and working tax credits had to be stopped for six weeks, while her eligibility for the single monthly payment was assessed. She created an online “journal” to record her income and provide supporting evidence and was told that she could apply for an advance, which would have to be paid back later, to cover the time she spent waiting.

She received her payments in November and December, then ran into a problem. Whitaker, a mother of three, owns a house jointly with her ex-husband, but it was on the market and had no tenants. (She was renting elsewhere.) This seems to have given the jobcentre computer conniptions: did Whitaker have an asset that meant her housing benefit should be reduced, or not?

She received a demand in her “journal” a few days before Christmas: show us that you’re paying rent, or we’ll stop your benefits. “I was on my fifth round of chemo and I wasn’t well at all,” she says. “After Christmas, I couldn’t get hold of anyone to give me a straight answer. This went on for about a month.” The January payment didn’t come. Whitaker spent hours on the phone – her mother, listening to our call, chimes in to amplify this point – and she eventually received a letter admitting that it was a mistake to withhold her benefit. “I can remember being on the phone, crying my eyes out,” she says. “Chemo, it does your brain in. It was the last thing I needed. It was an absolute nightmare.”

Yet Jo Whitaker’s story is not a particularly extreme one. She is, she says, lucky to have a great support network, and she never felt truly helpless. Her business experience helped her budget and cope with rectifying the jobcentre’s error. I’ll also admit that when I heard she had a house, I thought: hang on, why is she claiming benefits when she has an asset? As she talked, the situation became clear. But this is the kind of detail that computer systems struggle to deal with: the hardest thing to build into any IT project is common sense.

Many aren’t as resilient as Whitaker. New figures from the Department for Work and Pensions show that around a quarter of new claimants wait more than six weeks for their first payment. And because Universal Credit is paid to tenants, rather than directly to landlords, it has significantly increased the number of people falling behind on their rent.

There’s a cruel double bind here. Most people claim benefits precisely because they are in difficult personal circumstances. They have lost their job, got sick, or broken up with a partner and had to move house. Those same circumstances make dealing with bureaucracy more challenging. When the computer says no, it doesn’t just take away one of half a dozen benefits; it can disrupt the only assistance people are getting.

The quiet unhappiness of Jo Whitaker’s story should worry the government. In 2015, the possibility of cuts to tax credits caused enough concern on the doorstep and in constituency surgeries that even Tory MPs quailed. George Osborne’s resulting fudge was to kick back the cuts, promising that “savings” would be found anyway as more people moved to Universal Credit.

The idea that this can be accomplished without people feeling noticeably poorer is optimistic. That it can be accomplished using the existing IT system is even more so. Universal Credit should be a pragmatic project, but it has always been politicised: first by Iain Duncan Smith’s evangelical insistence that he would “make work pay” (even though 60 per cent of UK households in poverty have at least one member who works) and then by his flouncing anger that the project was being used as a cover for “salami-slicing” the welfare budget. IDS must have been the last man in Britain to work out that Osborne wasn’t just pretending to be into austerity; he really loved it.

In 2013, the National Audit Office found that the Universal Credit programme was struggling with a “tight timescale, unfamiliar project management approach and lack of a detailed plan”. The Labour MP Margaret Hodge, then the chair of the public accounts committee, concluded that most of the £425m spent so far would have to be written off. The programme was “reset”.

That, in effect, is what Citizens Advice wants to happen again. The organisation is calling for a pause on the roll-out, which is scheduled to accelerate next month. “[It] is a disaster waiting to happen,” says its chief executive, Gillian Guy. “People face severe consequences, like visits from bailiffs and eviction, when they can’t pay their bills.”

Like Jo Whitaker, she believes that the “principles behind Universal Credit are sound”. But that won’t be a consolation to anyone left cold, hungry or homeless over Christmas. In politics, there’s nothing more dangerous than something that everyone thinks is a good idea. 

Helen Lewis is deputy editor of the New Statesman. She has presented BBC Radio 4’s Week in Westminster and is a regular panellist on BBC1’s Sunday Politics.

This article first appeared in the 21 September 2017 issue of the New Statesman, The revenge of the left