Читать книгу Soccernomics - Simon Kuper - Страница 5

DRIVING WITH A DASHBOARD: IN SEARCH OF NEW TRUTHS ABOUT FOOTBALL

A few years ago, the data department at Manchester City carried out a study of corner kicks. City hadn’t been scoring much from corners, and the analysts wanted to find out the best way to take them. They watched more than four hundred corners, from different leagues, over several seasons, and concluded: the most dangerous corner was the inswinger to the near post.

The beauty of the inswinger was that it sent the ball straight into the danger zone. Sometimes an attacker would get a head or foot to it and divert it in from point-blank range. Sometimes the keeper or a defender stopped the inswinger on the line, whereupon someone bashed it in. And occasionally the ball just swung straight in from the corner. Of course, you wouldn’t want to take every corner as an inswinger. It’s a good idea to hit the odd outswinger too, just to keep the opponents guessing. This is what’s known as a mixed strategy. But all in all, the analysts found, inswingers produced more goals than outswingers.

They took their findings to the club’s then manager, Roberto Mancini, who like almost all managers is an ex-player. He heard them out politely. Then he said, in effect: ‘I was a player for many years, and I just know that the outswinger is more effective.’ He was wrong, but we can understand why he made the mistake: outswingers tend to create beautiful goals (ball swings out, player meets it with powerful header, ball crashes into net) and beautiful goals stick in the memory. The messy goals generally produced by inswingers don’t.

At first Mancini didn’t change his thinking. But sometime around 2011, when City were again having trouble with corners, his assistant David Platt came to chat with the club’s data department. The analysts told Platt about the corners study. They heard nothing more about the matter, but soon they noticed that City had begun taking inswinging corners. In the 2011/2012 season City scored 15 goals from corners, more than any other team in the Premier League. Ten of those goals came from inswingers, including the header from Vincent Kompany against Manchester United that effectively sealed the title for City.

It’s a story that captures where football is today. On the one hand, the March of the Geeks has advanced fast since we first published Soccernomics in 2009. Football is becoming more intelligent. The analysts who now crunch ‘match data’ at almost all big European clubs (and at many smaller ones) are just one symptom of the shift.

Today’s plugged-in clubs know stats like ‘pass completion rates in the final third of the field’, miles run in each phase of the game and pace of sprints for all their players. These numbers increasingly inform decisions on which players to buy and sell.

On the other hand, as Mancini’s initial rejection of the data about corner kicks shows, there is still widespread suspicion of numbers in football. John Coulson of the data provider Opta told us, ‘There are still maybe a lot of teams that view data as a threat rather than as a tool.’ Statisticians don’t always make the best communicators. Baseball has had its Moneyball revolution, but in football, the transformation is still just in its first phase.

This new, updated, expanded edition of Soccernomics uses data to clarify our thinking on topics ranging from tackles through transfers to why England lose and why China might start winning. We have a new chapter on crooked business; one on how the biggest clubs might now finally be turning into serious businesses, and why that isn’t a good thing; and a final chapter arguing that the game has never had it so good (though the smartphone could bring everything down). We have also expanded our thoughts on some mystifying questions, such as: ‘How do clubs use data to judge, buy and sell players?’ and ‘How powerful are agents in the transfer market?’ In every chapter of the book we have found stories and analyses to update, and new thoughts to add.

We’ve watched fans and media shift to our point of view on certain issues: most people now recognize that hosting big tournaments doesn’t make you rich, and also that England shouldn’t expect to win those tournaments. (We wish we could claim responsibility for shifting global opinion, but we can’t.) On other issues, we’ve changed our mind somewhat. In 2009 we were confident that the rest of the world would soon catch up with the best Western European nations. That hasn’t happened, so we’ve had to rethink what’s going on. We’re with the economist John Maynard Keynes: when the facts change, we change our minds.

It’s a long way from Soccernomics’s beginnings in the Hilton in Istanbul one winter’s day in 2007. From the outside the hotel is squat and brutalist, but once the security men have checked your car for bombs and waved you through, the place is so soothing you never want to go home again. Having escaped the 14 million-person city, the only stress is over what to do next: a Turkish bath, a game of tennis, or yet more overeating while the sun sets over the Bosphorus? For aficionados, there is also a perfect view of the Besiktas football stadium right next door. And the staff are so friendly, they are even friendlier than ordinary Turkish people.

The two authors of this book, Stefan Szymanski (a sports economist) and Simon Kuper (a journalist), met here. Fenerbahce football club was marking its centenary by staging the ‘100th Year Sports and Science Congress’ and had flown us both in to give talks.

The two of us had never met before, but over beers in the Hilton bar we found that we thought much the same way about football. Stefan as an economist is trained to torture the data until they confess, while Simon as a journalist tends to go around interviewing people, but those are just surface differences. We both think that much in football can be explained, even predicted, by studying data – especially data found outside football. We decided to write a book together.

When we began writing, Stefan lived in London and Simon in Paris, so we spent a year firing figures, arguments and anecdotes back and forth across the Channel. As we talked more and began to think harder about football and data, we buzzed around all sorts of questions. Why was football such a terrible business? Might the game somehow deter people from killing themselves? And are fans really monogamous?

Applying data to these questions felt like a new project. Until very recently, football had escaped the Enlightenment. Football clubs are still run mostly by men who do what they do because they have always done it that way. These men used to ‘know’ that black players ‘lacked bottle’, and they therefore overpaid mediocre white players. Today they discriminate against black managers, buy the wrong players, and then let those players take corners and penalties the wrong way. (We can, incidentally, explain why Manchester United won the penalty shoot-out in the Champions League final in 2008. It’s a story involving a secret note, a Basque economist and Edwin van der Sar’s powers of detection.)

Entrepreneurs who dip into football also keep making the same mistakes. They buy clubs promising to run them ‘like a business’ and disappear a few seasons later amid the same public derision as the previous owners. ‘I screwed up,’ Tony Fernandes, chairman of Queens Park Rangers, told us. Fans and journalists aren’t blameless, either. Many media headlines rest on false premises: ‘Newcastle Land World Cup Star’, ‘England Underachieve’ or ‘World Cup Will Be Economic Bonanza’. The game is full of unexamined clichés: ‘Football is becoming boring because the big clubs always win’, ‘Football is big business’ or ‘The big money will turn fans off’. None of these shibboleths has been tested against the data.

Most male team sports have long been pervaded by the same overreliance on traditional beliefs. Baseball, too, was until quite recently an old game stuffed with old lore. Since time immemorial, players had stolen bases, hit sacrifice bunts and been judged on their batting averages. Everyone in baseball just knew that all this was right.

But that was before Bill James came along. James was from the rural state of Kansas in the middle of the US. He hadn’t done much in life beyond keeping the stats in the local children’s baseball ‘Little League’ and watching the furnaces in a pork-and-beans factory. However, in his spare time he had begun to study baseball statistics with a fresh eye and discovered that ‘a great portion of the sport’s traditional knowledge is ridiculous hokum’. James wrote that he wanted to approach the subject of baseball ‘with the same kind of intellectual rigor and discipline that is routinely applied, by scientists great and poor, to trying to unravel the mysteries of the universe, of society, of the human mind, or of the price of burlap in Des Moines’.

James told us that baseball set the trend for the global data revolution, because the game’s record-keepers had begun gathering stats in the nineteenth century – before almost any other human activity. James explained: ‘So when the computer revolution started 100 years later, we were ahead of the game. We had 100 years of really interesting data to play around with. So the analytical revolution hit in baseball before places where sensibly you would think it would hit.’

In self-published mimeographs masquerading as books, the first of which sold seventy-five copies, James began demolishing the game’s myths. He found, for instance, that an extremely telling statistic in batting was the rarely mentioned ‘on-base percentage’ – how often a player manages to get on base. James and his followers (statisticians of baseball who came to be known as sabermetricians) showed that time-honoured strategies like sacrifice bunts and base stealing didn’t make any sense.

His annual Baseball Abstracts turned into real books; eventually they reached the best-seller lists. One year the cover picture showed an ape, posed as Rodin’s Thinker, studying a baseball. As James wrote in one Abstract: ‘This is outside baseball. This is a book about what baseball looks like if you step back from it and study it intensely and minutely, but from a distance.’

Some Jamesians started to penetrate professional baseball. One of them, Billy Beane, general manager of the little Oakland A’s, is the hero of Michael Lewis’s earth-moving book Moneyball and the film starring Brad Pitt. In recent years Beane, like so many Americans, has become a football nut. He has spent a lot of time thinking about how his insights into baseball might apply to football, and in 2015 he made his first official foray into the game, as adviser to AZ Alkmaar in the Netherlands. In December 2017 he was one of a consortium of foreign investors that took over Barnsley. (We’ll say more later about Beane’s gaming of baseball’s transfer market and its lessons for football.)

For several seasons Beane’s Oakland A’s did so well using Jamesian ideas that eventually even people inside baseball began to get curious about James. In 2002 the Boston Red Sox appointed him ‘senior baseball operations adviser’. That same year the Red Sox hired one of his followers, the twenty-eight-year-old Theo Epstein, as the youngest general manager in the history of the major leagues. (Beane had said yes and then no to the job.) The ‘cursed’ club quickly won two World Series. Today large statistical departments are the norm at American baseball clubs. Now football has embarked on its own Jamesian revolution.

A NUMBERS GAME

It’s strange that football always used to be so averse to studying data, because one thing that attracts many fans to the game is precisely a love of numbers.

The man to ask about that is Alex Bellos. He wrote the magnificent Futebol: The Brazilian Way of Life, but has also written several books about maths. ‘Numbers are incredibly satisfying,’ Bellos tells us. ‘The world has no order, and math is a way of seeing it in an order. League tables have an order. And the calculations you need to do for them are so simple: it’s nothing more than your three-times table.’

Though most fans would probably deny it, a love of football is often intertwined with a love of numbers. There are the match results, the famous dates, and the special joy of sitting in a coffee shop with your phone on a Sunday morning ‘reading’ the league table. Fantasy football leagues are, at bottom, numbers games.

In this book we want to introduce new numbers and new ideas to football: numbers on suicides, on wage spending, on countries’ populations, on passes and sprints, on anything that helps to reveal new truths about the game. Though Stefan is a sports economist, this is not a book about money. The point of football clubs is not to turn a profit (which is fortunate, as few of them do), nor do we get particularly excited about any profits they happen to make. Rather, we want to use an economist’s skills (plus a little geography, psychology and sociology) to understand the game on the field, and the fans off it.

Some people may not want their emotional relationship with football sullied by our rational calculations. On the other hand, the next time their team loses a penalty shoot-out at the World Cup these same people will probably be throwing their beer bottles at the TV, when instead they could be tempering their disappointment with some reflections on the nature of binomial probability theory.

We think it’s a good time to be rewriting this book. The amount of information available is expanding exponentially. In recent years the world has entered the era of ‘big data’. The phrase describes the unprecedented mountain of information that is now collected every day. This information comes mostly from the internet (from innumerable search terms, Facebook pages and emails), and from sensors that are attached to ever more physical objects – among them, footballers during training sessions. We have much more data to help us understand events than human beings could gather using only their eyes and ears. Moreover, all this data can be stored and faithfully reproduced, without the annoying tendency that humans have to misremember or just plain forget. Computing can identify patterns in datasets that would not be visible to a person ‘reading’ the data. We believe the data revolution enhances the capacity of humans to make good decisions. Note that we say enhance, not replace. Cyborgs replacing humans is, for now, still science fiction. But humans can make better decisions if aided by data analysis.

That’s true in football too. For the first time in the game’s history, there are a lot of numbers to mine. Traditionally the only data that existed in the game were goals and league tables. (Newspapers published attendance figures, but these were unreliable.) In 1979, after Steve Daley was transferred for a then record £1.43 million, from Wolves to Manchester City (where he flopped), the Treasury considered a tax on football transfers. The problem was that it couldn’t find any reliable financial data on the topic. In the end a young civil servant had to page through the Rothmans Football Yearbook to work out more or less how much clubs had spent on transfers the previous season.

Now the game is drowning in information. Data companies such as Opta can collect millions of observations (facts) about a single game. Clubs, which used to rely on gut alone, now use the new stats to analyse games and players. Every day, data analysts collect ever more information about every player’s every move on the field, the training ground, and even in bed – they know how well he slept last night.

Academics are pitching in as well. At the end of the 1980s, when Stefan went into sports economics, only about twenty or thirty academic articles on sports had ever been published. Now countless academics work on football. Many of the new truths they have found have not yet reached most fans. Much of what we argue in the book – for instance, that a club’s wage bill is an excellent predictor of its league position – is taken from Stefan’s academic work. Other insights come from his colleagues’ work. Generally speaking, we are more confident of what we assert when it is backed up by research that we believe is credible: meaning that the methods are clear, the data is adequate, and the results carefully explained and preferably peer-reviewed. You could still disagree with the work, but it has a solid foundation. Peer-reviewed academic research is, for us, the gold standard.

However, this book is more than just academic work rewritten for laypeople. Simon has been covering football as a journalist for over thirty years. He has met and interviewed many of the people who have shaped the modern game. This kind of knowledge is not always susceptible to formal statistical tests, but it is knowledge just the same. It informs our understanding of the sport. For instance, Simon came up with our theory of football networks: that Western Europe keeps winning World Cups because the countries of this little region are constantly exchanging knowhow with each other. We have much less data for this theory than, say, for Stefan’s insight that coaches have little impact on results. However, we think the network theory is plausible – and Stefan has set his economist’s brain to developing it. We think our combination of data and experience is the best way to understand social activities.

Like Stefan, Simon has also drawn on the knowledge of his colleagues – the growing number of people who write football books. When Pete Davies published All Played Out: The Full Story of Italia ’90, there were probably only about twenty or thirty good football books in existence in English. Now – thanks partly to Davies, who has been described as John the Baptist to Nick Hornby’s Jesus – there are thousands in untold languages. Simon, in his office in Paris, has a library containing a large proportion of them. Many of these books contain truths about the game that we try to present here.

What has happened in football mirrors a trend across all sports. Michael Lewis, author of Moneyball, wrote in 2009: ‘The virus that infected professional baseball in the 1990s, the use of statistics to find new and better ways to value players and strategies, has found its way into every major sport. Not just basketball and [American gridiron] football, but also soccer and cricket and rugby and, for all I know, snooker and darts – each one now supports a subculture of smart people who view it not just as a game to be played but as a problem to be solved.’

In football, these smart men (it’s still part of the game’s own ‘ridiculous hokum’ that they have to be men) have even begun taking key roles at some of Europe’s biggest clubs. European football, professional on the field for over a century, is finally creaking into professionalism off the field, too. Given the global obsession with the game, there could come a time when some of the best and brightest young people are working in the front offices of football clubs. Already the rising generation of club executives understands that in football today, you need data to get ahead. If you study figures, you will see more and win more.

One early harbinger of the Jamesian takeover of football was the Milan Lab. Soon after it began work, AC Milan’s in-house medical outfit found that just by studying a player’s jump, it could predict with 70 per cent accuracy whether he would get injured soon. Later the Lab began testing, almost day by day, each player’s muscle weaknesses, the movement of his eyes, the rise and fall of his heart-rate, his breathing, and many other obvious and less obvious indicators. Jean-Pierre Meersseman, the Lab’s cigarette-puffing director, was given a power of veto over the club’s prospective signings. ‘The last signature on the contract before the big boss signs is mine,’ he told us in 2008. By 2013 the Lab had performed 1.2 million physical tests on Milan’s players, collected millions of pieces of data on computers, logged even the slightest injury to every player, and in the process had stumbled upon the secret of eternal youth.

Most of Milan’s starting eleven who beat Liverpool in the Champions League final of 2007 were thirty-one or older: Paolo Maldini, the captain, was thirty-eight, and Filippo Inzaghi, scorer of both of Milan’s goals, was thirty-three. (After the final whistle, Inzaghi still had enough juice to kick a ball around on the field for fun.) In large part, that trophy was won by the Milan Lab and its database. It is another version of the March of the Geeks story. In the last few years, cash-strapped AC Milan have reduced the Lab’s funding and power. However, other big clubs all over Europe now lead the data-driven quests to reduce injuries, and to predict which twelve-year-old will grow up to be the next Xavi. Meersseman says data-driven scientists seem to be better than experienced youth coaches at making those predictions. He told us: ‘In football, they say, “You know about football or you don’t.” And when you go and test the ones who “know”, it’s surprising how little they know. It’s based on the emotion of the moment.’

What started in Istanbul in 2007 as a book idea has turned into a long-term collaboration. These days our contact is transatlantic: Simon is still in Paris, but Stefan is now at the University of Michigan. Together with Ben Lyttleton we have also founded the Soccernomics consultancy. On its website, Soccernomics-agency.com, we publish an occasional blog with our thoughts on football. And we have kept rewriting and updating this book. It has sold a total of about 250,000 copies in over two dozen languages. We’ve had the chance to influence the opinions of lots of people, some of whom work in the game.

All the while, we have continued to distrust every bit of the game’s ancient lore, and tested it against the numbers. As Meersseman says, ‘You can drive a car without a dashboard, without any information, and that’s what’s happening in football. There are excellent drivers, excellent cars, but if you have your dashboard, it makes it just a little bit easier. I wonder why people don’t want more information.’ We do.