Why Lunin didn't take Bernardo Silva's penalty and change the shootout: RDT's goal that convinced him to hold on

PreviousLiveChronicle

When Andriy Lunin saw Bernardo Silva approaching to take the second penalty of the shootout on Wednesday, he remembered a sunny February day at two-thirty in the afternoon in Vallecas. So, he had instructions from Luis Llopis, Real Madrid's goalkeeping coach, to hold in the center. The analysis of Ancelotti's people predicted that this is where the Rayo striker was going to aim. However, the Ukrainian stretched to his left and RDT did indeed score in the middle. Lunin was then somewhat regretful about that, although the coaching staff knows that it is not easy for a goalkeeper to hold on without diving: they prefer to take the wrong side than not move.

MNCM. City

1

Ederson Moraes, Rúben Dias, Manuel Akanji (John Stones, min. 111), Walker, Gvardiol, Bernardo Silva, Foden, Rodrigo, De Bruyne (Kovacic, min. 111), Grealish (Jeremy Doku, min. 72) and Erling Braut Haaland (Julián Álvarez, min. 90)

RMA Real Madrid

1

Andriy Lunin, Ferland Mendy, Nacho, Dani Carvajal (Eder Militao, min. 110), Rüdiger, Federico Valverde, Jude Bellingham, Kroos (Modric, min. 79), Rodrygo (Brahim Diaz, min. 83), Camavinga and Vinicius Junior (Lucas Vázquez, min. 102)

Goals

0-1 min. 12: Rodrygo. 1-1 min. 75: De Bruyne.

Referee

Daniele Orsato

Yellow cards Dani Carvajal (min. 37), Grealish (min. 59), Gvardiol (min. 61), Rodrigo (min. 89) and Ferland Mendy (min. 94)

And there was Bernardo, a kind of second chance, in the most decisive moment of the round, perhaps the most important of the Ukrainian's career. City had scored the first and Ederson had stopped Modric's goal. The team that opens the tiebreakers has a certain advantage: it wins between 55% and 58% of the time. In part, because of the pressure that going from behind adds to the opponent. Even more so, if you advance with a disadvantage. A success by the Portuguese, a temporary 2-0, would have meant a gigantic weight. On the night in Manchester, Lunin stood on the goal line with the same instructions as in Vallecas: hold in the center. This time he followed the plan, grabbed the ball and began to turn the round.

The path had been laid out for days. During extra time, there was a section of the Madrid bench convinced that if they reached penalties, the tie was theirs. After the victory, Lunin, who thanked Llopis for his help, only revealed part of the plan: “You had to take a risk with one. In one you had to stay in the center. Thank goodness it turned out well,” he said. But it wasn't just any one.

When extra time ended and the Madrid players gathered around Ancelotti, the Ukrainian passed by and went with Llopis and Kepa Arrizabalaga, his substitute, to review data and graphs. They were convinced that Bernardo was going to shoot to the center, according to sources familiar with the reports they handled. They had detected that it was his preference in the most pressured situations. Kepa even had a personal memory of the 2019 Carabao Cup final that he had played with Chelsea. The Portuguese had City's fourth, at a point of maximum stress: the London team had scored two and missed another two, and Guardiola's team had two hits and one miss. The Spanish goalkeeper dived to his left and Silva scored in the middle.

In the 2019 Carabao Cup final, Bernardo Silva scored a penalty against Kepa.

Last night, Kepa advised Lunin to stay down the middle and it worked. 🤍

pic.twitter.com/WQIlQK4san

— TC (@totalcristiano) April 18, 2024

The analysis of the Madrid coaching staff coincides with that of the London School of Economics professor Ignacio Palacios-Huerta, the greatest student of penalties through the use of game theory. In the first part of the season, the economist prepared a report for a Premier team that he advises in which he recommended the same thing: that the goalkeeper stay in the middle if Bernardo had a penalty. This Thursday morning, intrigued, he rescued the document to review his forecast: “It's fascinating,” he says. “He had taken only three penalties! One on the left, one in the center and one on the right. And it seems like the algorithm should treat them as similar. There is very little data and they are all very same. Well no! The algorithm tells you that there is a 51% probability that he will shoot in the center and the rest is divided between right (15%) and left (34%).” A very significant difference with such a small sample.

Palacios-Huerta has been building a sophisticated algorithm for years that makes recommendations based on the processing of data from some 20,000 pitches passed through the Nash equilibrium filter of game theory. This equilibrium predicts the strategy that each contestant will choose taking into account what the rival will do. The algorithm uses information from the player complemented with that of similar players. In the case of Bernardo Silva, left-handers who would have taken their first three penalties each to one side. “But also what do they do when there is pressure,” he explains. “A friendly is not the same as a Cup final. Here there were very few penalties, but a lot of pressure difference.” The same thing that Llopis and Kepa had concluded to convince Lunin not to do as with RDT in Vallecas.

Although after the game Rodri pointed out at random: “We had the feeling that penalties were obviously better for them. It's their turn again. “As normal.” Real Madrid has not lost a penalty shootout since 2012, in the Champions League semi-final against Bayern, the rival in two weeks' time also in the semi-final.

Palacios-Huerta always rejects the approach of chance: “It is incredible that people are superstitious and do not pay attention to how improvable a play can be that can be crucial in a totally crucial match.” Guardiola agreed this Wednesday: “Johan Cruyff told me that luck does not exist, and I quite agree. In the end we didn't manage to score a goal. Enough".