Big data hits the football world: winning or losing depends on data or intuition?

With a dramatic end to the 2013/14 season in the Premier League, if there is anything more than the last round of winning the title, it must be the fast decline of the defending champion Reds and the emergence of the unknown Everton. Everton not only beat Manchester United, but even the extravagantly of the transfer market in London, Tottenham Hotspur, has become a hot topic for fans.

Everton's onslaught represents a new trend in football: the Data revolution.

European football is going through such a revolution. Professional fans may find clues from the news stack: The big clubs are becoming smarter and more efficient. Previously, the statistics on football were only corner, free kicks, red cards and shooting times, and now people can see how many meters a player has run in a single field, the trajectory and speed, and so on. For example, Opta, a sports data company, recorded as many as 1500 data per football match.

Today, all 20 clubs in the Premier League have dedicated technical analysts to deal with the new data, with the city being hired to hire 11 of them. In 2012, Liverpool even created a new post called "research director" to recruit Dr. Ian Graham of Theoretical physics. The researchers will take part in the pre-match and after-game summaries, which will help management identify and select transfer targets, and even offer nutritional advice to young players. The revolution was so shocking that British media exclaimed: "Frankenstein has invaded the world of competitive sports in Europe!" ”

Data doesn't lie.

In the traditional and conservative sport of football, even if the introduction of door-line technology (to determine the effectiveness of the goal) or to create a new door to the referee will cause a stir, the data revolution sounds incredible.

However, in the United States on the other side of the ocean, systematic data analysis has invaded major sports federations as early as ten years ago. The most famous of the American League Baseball Oakland sports team general manager Billy Bien's crazy experiment, http://www.aliyun.com/zixun/aggregation/19058.html ">2003 years, he boldly abandoned the traditional" strike rate " The choice of talent, instead of digging up the "top rate" high potential stars, and the players to recruit to the team.

Bean's "revolution" drew a barrage of ridicule. Because there is an old saying in the sports world, "data tells lies." A few simple examples: The basketball player does not equal the high score to the team, because he may be in pursuit of scoring and crazy shot, resulting in the team can not smooth operation of the attack, rugby four is a low percentage of the cut is not necessarily good, because he may be in order to avoid being copied to delay the shot time, and ultimately by the opposing defense team to kill Some tennis players, such as the Swiss king Roger Federer, have a higher percentage of involuntary errors than other players, but that is because he plays more aggressively and has more winning points than anyone else. Therefore, the strength and potential of athletes is difficult to judge by a simple figure, especially in collective projects.

However, the athletic team's subsequent performance has been the best comeback. In the new season that followed, the team made 20 straight wins, a total of 103 wins in the regular season, with the Yankees and listed as the largest league wins the most number of teams (the foreign gene less than one game and a good record); you know, the sports team's total salary is only more than 51 million, not yet the sum of the top three stars in the New York Yankees.

Even though it has been criticized for his too much emphasis on the barriers to the selection of talent, his side did poorly in the playoffs, but he inspired the other teams in the big leagues to embrace better big data analysis instead of relying on a handful of data and scouts ' intuition ' to pick people out. '

Bean and his sporting legend were written by bestselling author Michael Lewis, the best-selling book titled "Penalty Gold" (also translated as "Orb Theory") was then made into a movie, not only in the sporting world, but also on Wall Street hedge-fund managers talking about "penalty gold".

In fact, the lie is not the data, but the lack of data. In today's big data age, clubs, media and professional fans are unlikely to be blinders by individual data. Because they see an NBA player scored 36 points in a single game, they will also see him complete the 40% offensive, and only 0.87 points for each attack, while his players scored 1.12 points per turn. They will also see that the other player, although the shooting rate is only 35%, but 12 times out of the hands of 7 are in 24 seconds of the last 3 seconds to receive the ball, so his low hit rate is not his own fault, but for his teammates back the "scapegoat."

The sports manager understood this earlier than the media and the fans. In 2004, Arsenal's French manager Arsene Wenger learned from the data that there was an unknown child in France named Mathieu Flamini, who was able to run 14000 meters in a race. That's a really amazing number, but it's not enough because Wenger is unsure if he's in the right direction. So Wenger went to the scene to watch his game and decided to sign the little player when he came back.

But in the big data age, you don't have to go to the scene to see if the player is in the right direction. In 1998, Garrick Barr, who served as a video editor for the NBA's Phoenix Suns for 11 years, set up a company that records every single game of the NBA, tracking every attacking turn by video, and then categorizing each player with an "attack trend report." Two years later, he worked with former Microsoft engineer Niels Larre to set up the Synergy Sports technology company, combining the finer data the NBA coaches wanted with the video corresponding to each data. For example, Synergy recorded every attack that Dirk Nowitzki had entered the NBA in 1998, and through the data provided by the data company, we would know the success rate of Nowitzki's breakthrough to the right or to the left, the low back-scoring error ratio, the 24-second final hit, and so on.

By the year 2011, 26 teams in the NBA's 30 teams have used Synergy's data systems, and media companies such as ESPN and the NBA's official website have used the company's data to study the game, and they have even launched a popular version for fans, with less than 100 dollars, Ordinary fans can also see dozens of "advanced data" combined with video playback.

Everton attack

In Europe, football's gold are also aware of the magic of the penalty shoot-out, at least because they have seen the Hollywood film starring Brad Pitt. But they think football is a very dynamic sport, very different from baseball. In the Premier League and even in other European leagues, most managers have been good professional players and they think they have a deeper understanding of the sport. Halley Redknap is one of them. When he was a coach at Southampton, when the data analyst tried to show him the weakness of the game, Redknapp turned to taunt: "Otherwise, next week, we will use your computer and their computer to decide the outcome, OK?" ”

Harry Redknapp does not know that his sarcasm is actually very close to the point of the problem: football should not be just a contest between 22 players, but more importantly, it is the two managers using their own players to fight wits.

Everton may be the Premier League sports team. FT columnist Kuper Cooper has pointed out that over the past 10 years Everton have been the highest-input and most stable team in the Premier League. Under Moyes, they scored the eighth place in every season in 2007-2013. Despite the little spending each year, the city has a big club such as Liverpool, in the transfer market is also difficult to attract big-name stars, but they rely on the training team to cultivate the talent, can maintain stable results, but also in the transfer market to make money. Wayne Rooney, Rodwell and Ross Backley are the fruits of their youth training.

After Moyes job-hopping to Manchester United, Everton invited Martinesley of Wigan Athletic to be their new manager. Martinez's obsession with data analysis is well known, he put a 60-inch touchscreen TV in his home, combined with a software prozone that can record and analyze the player's performance, watch the team's video every day, especially the losing game, and watch it more than 10 times, to dig failure. His coaching mentality is also highly creative, with most teams favoring the traditional 4-4-2 formation, while Martinez's Wigan Athletic team will switch back and forth between 4-3-3, 3-4-3 and 4-2-3-1. In short, he is particularly suited to innovative clubs such as the Everton club.

Under Martinez's guidance, Everton will record each player's action in training: "We monitor the status of each player with GPS and heart rate gauges." From a physical standpoint, the most significant data is the number of sprints, sprint distances, and the number of high-intensity movements each player puts in. We monitor this whole season to see if a player is in a state of fatigue and how much time he needs to rest. ”

Martinez knew that, if not treated, the data would always be a bunch of numbers and not be a revelation. Thus, the question of who should be the person who analysed these figures was a new one. James Smith, director of technical Scouting, who was taken to Manchester United by Moyes, said European football is still deficient in data analysis compared to the North American professional League: "The people who deal with this data in the Premier League are usually sports science majors, and I am one of them," he said, "but in the United States, You might see a Harvard Law school or a computer science guy from MIT doing this kind of work. We have a problem with the English football club, we spend too much money on the transfer fee, the player's salary and the brokerage fee, so we can't bring in enough quality technical analysis staff. ”

Of course, there are some excellent data-processing gurus in the Premier League, but since most teams are conservative, these numbers are often the prozone of a data service provider. To change the situation, Manchester City had an interesting experiment in August 2012, opening a project called "MCFC Analytics", boldly exposing a large number of Opta's 2011/12-year data-related documents and calling on anyone who "loves football and data" to analyze them.

"There are a lot of people who like football and they like to do their own analysis, but they need data and it takes money to collect data," said Gavin Fregg, director of performance analysis at Manchester City. "Their approach, known as" Open source data Analysis ", attracted more than 1500 users to log in 36 hours before opening the database. However, the project was halted a year after its operation, and its success was difficult to assess.

Data or intuition?

Marcus Dusato, a professor of mathematics at Oxford University, is a die-hard Arsenal fan who believes that data analysis has the potential to be an X factor in changing the existing club's strength pattern. "Football is more like a chess game than people think," he said. "What happens in a team is not random, it's a certain rule." With the data, we can find these patterns and predict what will happen in the future. In essence, this is no different from what hedge fund managers are doing. ”

Dusato believes we should see the football team as a network, and the 11 players on the pitch are nodes, "like a mini internet." Successful teams like Barcelona have the ability to keep all the nodes straight, and in theory any other team should be able to do that. Thus, data analysis should be able to crack Barcelona's unique trick and become the other ball will follow the guide book.

"The world of football is quite conservative today," Dusato said, "for example, teams like Arsenal and Liverpool, if they can accept the opinions of some outsiders, will certainly get a real boost." After that, he said he would be more than happy to work for him if he wanted a mathematician to help them with his analysis.

But even Martinez admits that data analysis has its limitations. He and his two top scouts, rivers and Brown, thought it was absurd to think that "you can buy a player by data." "You need to see that player, you need to fall in love with that player," Martinez said, "and you'll see how he warms up, how he communicates with referees, how he communicates with his teammates after he misses an opportunity, how he celebrates the score, and how his teammates react after he scores. The data will only help you eliminate some obvious mistakes, but the final decision must depend on the person's judgment. That's what scouts and coaches have in their gut. ”

At the same time, data analysis may have some negative effects. Blaine Plesdich, Director of analysis and research at Bolton, revealed that since their goalkeeper began to study the data on each other's penalty kicks, his penalty-kick efficiency has dropped, with only 9% success rates in the past two seasons. "We gave him a way of thinking about data analysis, but it took the human part, and his instincts as a player were weakened," Plesdich said, although he immediately added that "data analysis also brings a lot of positive effects." ”

Football is a passionate sport, and if it becomes a pure computational work without a soul, it loses its charm. However, in this era, if any club or manager denies the role of the data, it will be a huge disadvantage. If data experts want to make a bigger impact on football clubs, they may have to shift the focus from the manager to the club owner. For example, the Liverpool boss John F. Henry is a good lobbyist, after all, he is rich in stocks, the set of data analysis is no stranger, and he is also the boss of the Boston Red Sox, the Major League Baseball, to know that the Red Sox have finally broken the 86-year "Holy Baby Curse" after following the reform of Bean's "penalty gold" theory, Finally won the World Series in 2004. On the other hand, players may also be starting to ask for the data, according to Chelsea's data analysis director Ben, who is now accustomed to getting a report on their own data after each game and training, and they will find out where they are doing well based on the data. Where it needs to be strengthened.

The fans never know what's going on in the club, whether it's watching live on TV or on the spot. Would you thank the manager for his genius, or the precise calculations of his data analysis think-tank, if a substitute player scored the first time he touched the ball after playing? At this point, especially if you are a Manchester United fan, you may not care about the difference at all.