The science of penalty shootouts: How teams are using data to win

This is article was published on May 31, 2017

The most important few minutes of Gianluigi Buffon's career were the penalties that decided the Italy-France World Cup final in 2006. Going into the shootout, the Italian keeper didn't feel confident. "I was not in tune with what was happening. It seemed to me that France could have taken 2000 penalties and scored all of them."

Buffon wasn't the kind of keeper who prepared for shootouts by studying opposition kickers. Instead he relied on intuition. But then his Juventus teammate David Trezeguet stepped up to shoot for France. Trezeguet and Buffon knew each other intimately. They sometimes practiced penalties after training sessions, writes British author Ben Lyttleton in his book on penalties, "Twelve Yards". Facing a keeper who knew his repertoire, Trezeguet decided he had to take a difficult kick: high in the left-hand corner. Had he hit it a few centimeters lower, football history might have been different, but his shot crashed out off the crossbar. Buffon had got lucky, and went on to win the biggest prize of his career.

Finals are often decided on shootouts. As Real Madrid showed in last year's final, shootouts are increasingly being decided by science.

The scientific era of the penalty shootout probably began with the Chelsea-Manchester United Champions League final in Moscow in 2008. Chelsea's coach Avram Grant received help from a Spanish economist named Ignacio Palacios Huerta, who had spent years building a database of thousands of penalties. Nobody can be sure what the world's best kickers and keepers will do on any given penalty, but Ignacio (he and I have since become friends) had enough information to make better guesses than anyone else. His database had revealed a crucial fact about United's keeper Edwin van der Sar: on penalties, the Dutchman often dived to his right.

Chelsea used Ignacio's research in the crudest possible way. After the match finished 1-1, all their first six penalty-takers in the shootout kicked to Van der Sar's left. The simple strategy very nearly worked: Van der Sar mostly went the wrong way, didn't save a single shot and Chelsea would have won if John Terry hadn't slipped when taking his penalty. He sent it to Van der Sar's left, the keeper duly went right, but the kick just missed.

Then Nicolas Anelka stepped up to take Chelsea's seventh penalty. On United's bench, manager Alex Ferguson was growing frustrated with his keeper. Ferguson later recalled, "I was thinking -- dive to your left. Edwin kept diving to the right."

But after six kicks, Van der Sar, or someone else at United, had figured out Chelsea's strategy. The Dutchman realized that every shot was going to his left.

As Anelka prepared to kick, the gangling keeper, standing on the goal line, extended his arms to either side of him. Then, in what must have been a chilling moment for Anelka, Van der Sar pointed with his left hand to the left corner. "That's where you're all putting it, isn't it?" he seemed to be saying.

Now Anelka had a terrible dilemma. Van der Sar had read his mind. What was Anelka to do? He decided to avoid the left corner, where he had presumably planned to put the ball, and instead kicked to Van der Sar's right. That might have been fine, except that he hit the ball at mid-height -- exactly the level that Ignacio had warned against, the height of shot that Van der Sar was best at stopping. He duly saved. Watching the kick on TV, Ignacio was "very upset." Anelka's decision to ignore Ignacio's advice probably cost Chelsea the Champions League.

I was in the stands in Moscow that night, and I completely missed what was going on. I wasn't alone. Millions of people watched the shootout live, yet nobody even seems to have spotted Van der Sar's pointing hand, although it's blatantly obvious when you watch on YouTube. That's typical: often in shootouts, the real action goes unnoticed.

Two years later, Holland reached the World Cup final against Spain in Johannesburg. I grew up in Holland, I support the team, and I arranged with an official I knew in the Dutch camp that Ignacio would write a penalty report on the Spaniards. Ignacio comes from the Basque region, an autonomous community, and was quite happy to conspire against his own country. He worked on the report day and night for 48 hours, and on the morning of the final, Holland's goalkeeping coach Ruud Hesp emailed to say, "It's a report that we can use perfectly."

With 10 minutes to go in extra time, the score was still 0-0 and it looked as if the final would go to a shootout. On the stands I was too nervous to watch the match. Instead I opened my laptop and began rereading the PDF of Ignacio's report. It contained some fascinating predictions as to what Spain's keeper Iker Casillas and the kickers Fernando Torres, Xavi and Andres Iniesta would do. But perhaps luckily for the state of my heart, just then Iniesta scored the winner, and I began writing my match report.

The use of stats in shootouts soon became routine. In 2012, Chelsea played Bayern Munich in the Champions League final. By then the Londoners had enough in-house penalty knowledge not to need to call someone like Ignacio. Their keeper Petr Cech prepped for the match by watching a two-hour DVD containing every Bayern penalty of the past five years. He also received an elaborate briefing from his club's data team. On the night, Cech chose the correct corner for all six of Bayern's penalties (one during the game, and five in the shootout). Chelsea won their first ever Champions League.

But the spread of best practice remains uneven. Last year's Champions League final in Milan between Real Madrid and their city rivals Atletico went to a shootout. Atletico won the toss, but then made a kindergarten error: they chose to shoot second. Ignacio's research shows that the team that kicks first in the shootout wins 60 percent of the time. That's probably because of the psychological pressure on the team kicking second: it keeps having to score to stay in the game.

Not everyone knows this initial advantage exists. TV commentators rarely even mention the toss. Bookmakers don't shift their odds immediately after the toss is done. One of the captains who blundered after winning a toss was Buffon himself: he may have decided the outcome of Euro 2008 when he let the Spaniards shoot first in their quarterfinal against Italy. Spain won, and went on to win the tournament.

Why did Atletico choose to shoot second against Real? Because earlier in the tournament, they had won a shootout against PSV Eindhoven after shooting second. Ignacio emailed me after the final: "Incredible to see this at this level"

Atleti had ignorantly gambled the European title on a single event (victory over PSV) whereas Ignacio by then had 11,000 events (penalties) in his database.

When Real Madrid's captain Sergio Ramos told his teammates that night that they would get to shoot first even though they had lost the toss, they were incredulous. Real had the penalty smarts that Atletico lacked.

Just as in the 2008 final in Moscow, almost nobody watching the shootout spotted how the match was won, but a couple of days later the Dutch soccer analyst Pieter Zwart posted a remarkable video on Facebook, titled, "Did Real Madrid know what Jan Oblak was going to do?".

The video shows that Atletico's keeper Oblak had a crucial tell, a giveaway: just before each penalty was taken, he stepped towards the side where he was going to dive. The step helped him get to his chosen corner faster. The problem comes when the opposition knows what he is doing -- and Real looked like they did. The players seemed to be working off a sophisticated penalty report. Four out of the five Real shooters ran up slowly, waited for Oblak to take his step, and slotted the ball gently in the other corner.

Data analysis won the European final. Football has got smarter.