With each click and drag of a mouse, young soccer fanatics are creating the building blocks of the advanced stats that are changing how the sport is played, watched and analyzed.
Opta and Prozone are among the companies that have taken soccer stats far beyond goals and saves, into the realm of pass completion percentage, defensive touches, percentage of aerial balls won, tackle percentage and goals scored above expectation. Cameras alone can’t process all these stats. So companies employ people — mostly young, mostly male, most logging matches in their spare time as a second job — to watch matches and document every event.
Their work has helped develop stats that capture the value of players who don’t score many goals, but who set them up with pinpoint passing and hustle. Teams use advanced stats to decide which players to buy and put on the pitch. And fans, whether they like it or not, read and hear more numbers than ever before about this sport that for so long bucked the sports-analytics trend.
On a Sunday last month, Opta let me watch as the loggers at its South London headquarters tracked the last 10 matches of England’s Premier League season. I stood among rows of young men at computer monitors as they scrutinized games, sometimes rewinding on one monitor to check a tough call while keeping track of the live feed on another. I tried to stay out of the way while their supervisor leapt away from watching his favorite team’s match to confirm every goal was attributed correctly. And I watched as Opta’s media team processed the raw numbers — 1,600 to 2,000 events per game — into TV-ready factoids, which they heard commentators repeat to TV audiences moments later.
In soccer stats, as in so many other numbers-gathering endeavors, big data sets are built piece by piece by human collectors with human imperfections, moods and preferences. Throughout the year, 350 part-time analysts working in London and a half-dozen other Opta branches in Europe and North and South America record every pass, header and goal while watching live or recorded video of more than 14,000 matches around the world. The London operation I watched will be logging each of the World Cup’s 64 matches.
Opta says software, standards and oversight can help it harness the best of human judgment while curbing any potential downsides. It sees the people behind its stats as a selling point. I wasn’t the first to be invited to watch. Many prospective customers visit during matches, said Aidan Cooney, chief executive of Opta. “Frankly, that sells the business.”
The business is providing stats to professional clubs, to national teams, to leagues — as the official data provider for the top divisions in England, Spain and Germany — and to the media.
A Tebow jersey and a Yankees cap
My day at Opta was an unusually busy one: Every Premier League club was playing its last match of the season. The finale wasn’t as exciting as 2012’s: Manchester City was all but assured of edging Liverpool for the title, and most Champions League and Europa League slots had been sewn up. The biggest suspense was whether Tottenham would finish in sixth or seventh in the league.
That was the case, anyway, for Paul Pettitt, 31, who is the assistant manager of data collection and a Tottenham Hotspur supporter. He spent the two hours between kickoff and final whistle alternately tracking Tottenham’s match against Aston Villa — when Tottenham took an early lead, he said he wanted a 25-goal win to contend for fifth place on goal differential — and jumping out of his chair to check on calls in other games, such as whether an early Swansea goal was a deflection. All logged events scrolled down a screen at his station, and when an important one came up, he conferred with the analyst who entered it.
This is when soccer’s rare stoppages of play are so valuable for analysts. A lengthy goal celebration allows loggers to rewind and rewatch goals and other major events, often while Pettitt looks on.
But most of the work is logging routine passes. Opta’s analysts log each one by dragging and clicking a mouse at the spot where the pass was received, then keying in the player who received it. Their monitors have an image of a soccer pitch in the background with video of the live match superimposed on top.
Confusingly, to my eyes, the broadcast image hardly ever corresponded to the image on the field. So loggers had to drag the mouse to a spot that had nothing to do with the ball’s location in the video rectangle. None of the loggers I watched got stuck on this point: After all, this was the 38th and last match of the season.
Each of the 10 matches had a pair of analysts assigned to it, plus a checker. Each analyst had his own monitor and tracked only one team’s touches. Sometimes the analysts conferred over calls — “Is it a tackle?” was a question in the fourth minute of the Liverpool match. (It wasn’t.)
Until eight years ago, Opta didn’t even produce the live numbers that are now such a staple of TV broadcasts. Pettitt started at Opta in 2001, fortunately just as the company was phasing out pen-and-paper logging. He wasn’t lucky enough to miss the VCR era. “My elbow started aching after a while” from all the rewinding, he recalled.
The more unusual a team’s formation, the harder it is to log its matches. A well-organized side like Barcelona can be easy to log, Khalid Hussain, U.K. training manager for Opta, said. Today he particularly enjoys challenging matches.
At his peak, Hussain was logging 10 to 15 matches a week during each Premier League season. His primary assignment was Arsenal, and he also worked four nights a week covering matches around the world. He once logged six matches in a day. “Then I went home at the end, in a pretty bad state,” he said.
All this meticulous work changed how Hussain, now 33, watches soccer. He became “very passionate” about Arsenal, to the point where he’d enjoy watching a Gunners match against Stoke more than Real Madrid versus Barcelona, a minority opinion in global soccer. When he clicked a name at one end of the pitch and then entered the same name at the other end seconds later, he came to appreciate the players who covered a lot of territory more than the flashy dribblers.
And he learned that his previous pet stat of possession time doesn’t mean much. “Working here burst that bubble,” Hussain said. “It doesn’t matter how much ball you’ve got. You’ve still got to do something with it.”
Hussain is mainly a supervisor now, though he pitches in as an analyst when needed. On this day, he logged Cagliari for its 1-0 loss to Chievo. Like other pinch-hitters who aren’t familiar with their assigned clubs’ players and formations, Hussain watched DVDs of recent Cagliari matches to prepare.
The loggers Hussain supervises generally are between 18 and 24 years old and male. (“We’ve got two girls in Leeds, and one girl in Germany,” he said.) They love sports. They enter an office fantasy NFL league. They go home and play video games. They day I watched, none wore soccer apparel but I spotted a Tim Tebow jersey and a Yankees cap.
It helps to be nuts about soccer, to appreciate “a job where they get to come in and watch football,” as Pettitt put it.
There is occasionally cheering in the analysts’ box. “As much as you can try to control them, if Liverpool score a goal while Man City are down a goal, you might hear a yelp from our Liverpool fan, and probably some censored words as well,” Pettitt said.
Candidates are tested for their understanding of soccer and their hand-eye coordination when using the Opta logging software. They have to type quickly with their left hands, without looking at the keyboard. Certified soccer coaches sometimes don’t have the required hand-eye coordination; the avid PlayStation players often do. “We give them five-hour tests, and pick out the ones who are best,” Hussain said.
At that stage, successful applicants remain far from match-ready. It will be at least a month before they’ll produce usable data, even under the easiest conditions of logging a recorded match. “For training, they do the same game over and over for two or three days,” Hussain said.
Cooney, the Opta chief executive, has tried his hand at logging, “much to everyone’s amusement,” he said. “It’s impossible, absolutely impossible for someone of my motor skill set,” he added. “If you don’t play PlayStation, basically, you’re finished.”
Opta employs full-time analysts to review every event of the matches it logs, a process that can take three to five hours. Its live analysts get 99 percent of player identifications correct, Pettitt said.
The match-trackers are rated on their performance, and the best get spare games. It creates a competition, and “keeps them on their toes,” Hussain said. He’s confident that today he’s one of the best loggers in London. He also gets to travel to train loggers at offices around Europe.
The dubious goals panel
Among Opta’s competitors is Prozone Sports, which tracks players on the pitch using cameras and player-recognition systems. Stewart Mairs, the U.S. operations manager for Prozone, said the company’s optical tracking system — like SportsVU’s for the NBA — gives it a leg up over Opta. The system produces millions of data points per game.
Prozone, like Opta, needs human loggers, too. Prozone’s cameras sometimes can’t tell players apart when they cluster, and don’t distinguish crucial game events. So it employs coders, usually interns or students who are interested in soccer, Mairs said. Like at Opta, they are supervised and trained by more experienced managers, and, for big matches, supplemented by more experienced coders.
Cooney said Opta is offering something different from camera tracking. “People want analytics,” he said. “That requires holistic data sets, which only we can deliver.”
Keeping standards consistent across offices is vital for Opta. An assist needs to mean the same thing in London, New York and Montevideo. Soccer stats already have plenty of doubters, and it doesn’t help that different companies track different numbers. Also, individual companies sometimes change what they track, as Opta does nearly every year after an annual review. (Possibly coming soon: more detail on fouls.)
So it’s all the more important that a company’s data can be trusted across space and time. “What we’ve had as a clearance” — a defender clearing a ball out of the goal area — “has always been the same, and will not change,” Pettitt said.
In addition to post-match reviews, Opta monitors stats across leagues, to make sure they don’t vary too much — and if they do, that it’s because of style of play and not analyst inconsistency.
Opta also updates its stats according to decisions of a Premier League group called the dubious goals panel, which weighs whether a player should be awarded a goal when, say, the shot deflected off a defender.
Close calls mean the live data is provisional. It’s good enough for television broadcasters, who pepper Opta’s media team with questions via instant message during the matches. I wandered over to watch the media group in action during play. They sat next to a wall with six television screens, usually more than enough but four short of the required number on this day. So laptops filled the gap.
During play, the media team moved quickly. Liverpool’s Martin Skrtel scored an own goal in the 20th minute. Duncan Alexander, 36, head of U.K. content and customer services for Opta, told his colleague to “run it” — in other words, to check that Skrtel had just set the league record for most own goals in a season, with four. The stat was confirmed, sent to the broadcasting company Sky, and announced by studio host Jeff Stelling right after the commercial break.
Later, Stelling mentioned that Fulham had used 38 players this season, a new record. I asked if that was from Opta. Alexander nodded.
These sorts of stats are nice to have, but won’t change the way managers set their lineups or choose tactics. However, the work of Opta and its ilk have brought soccer, very slowly, into the wider statistical revolution in sports. Alexander and Pettitt pointed to the increasing prominence of assists. A decade ago, “some people would refuse to give assists credence,” Alexander said.
Opta’s soccer-stats professionals acknowledge their numbers aren’t for everyone. “There will always be fans who, to use a phrase we hear occasionally, say the only stat they care about is the one in the top left-hand corner” — the score, Alexander said. “We’re not zealots. We don’t bang the drum saying, you have to view football the way we do.”