The Technology Behind the World Cup's Advanced Analytics

During Sunday’s 2-2 World Cup draw, the American forward Clint Dempsey, who scored one of two goals against Portugal, ran a total of 9,545 meters, with a top speed of 28.33 km/h over the course of 26 sprints. That wasn’t the top speed of the match, however, a title held by US defender Fabian Johnson, who reached an impressive 32.98 km/h.

Over the course of the match, detailed statistics, including dozens of other data points—the Americans ran a total of 110,299 meters, compared to the Portuguese side's 106,520 meters, for example—are collected for every player on both teams, and displayed for television audiences worldwide, as well as posted to the FIFA website.

The system FIFA employs to grab the data, called Matrics, was built and deployed by an Italian firm called Deltatre at each stadium in the World Cup, and involves the use of several technologies and manual inputs from a large crew to deliver the real-time stats.

“The real value is that it’s live,” Tomas Robertsson, Deltatre’s North American commercial director, told me over the phone. “The extensive data set in real time provides on site heat maps and attacking zones, as well as distance run, passes completed, and many other statistics.”

The 2014 introduction of goal-line technology was a great leap forward for the international tournament; a technology that’s supposed to mitigate damn clear injustices such as Frank Lampard’s goal that the refs didn’t see. But it’s just one piece of high-tech gear that’s deployed at every area in the World Cup.​

Robertsson explained that the system works like this: Three HD cameras in various locations at each arena use image recognition to recognize the 22 players, three referees, and the soccer ball. The system tracks the XYZ coordinates of each of the objects, and then relays the information to a multi-screen, digital workstation where 74 people pour over the data on-site, aided by another 20 back in Italy.

The reason Deltatre uses cameras to optically track everything is that soccer players have resisted adding tracking technology into their equipment—such as their shoes—despite it being possible for some time.

On top of the cameras, the company has written algorithms that calculate passing stats, ball possession, and other statistics, totaling 350 in all. When the tracking information is relayed to a terminal, a human operator watching a slightly delayed version of the match validates each action before its sent live to the web or TV.

The image recognition technology isn’t entirely automatic. At the beginning of each match operators have to tell the machine which team is which color—and the color of the refs’ shirts—as well as manually input each player on the pitch. While the match is getting played out, the tracking system knows which team has touched the ball because a person with a video game controller hits a button for the corresponding team.

Unlike many other sports, Robertsson said that soccer isn’t a stats heavy game, but the information gleaned from the system can still be pretty useful in determining player performance. For example, it’s much easier to determine player fatigue simply by looking at the numbers.

Off the field, most of the teams at the World Cup use some kind of system to analyze each match to a similar level of detail, but after the fact. The collected data is analyzed by coaches and trainers to help ensure that teams are working at their peak potential.