When Football Worships Defense: What Expected Goals Data Says

On the night of July 10, 2026, from the stands of the Luzhniki Stadium in Moscow, I sat close enough to hear Didier Deschamps shout. France beat Belgium 1-0 through a Samuel Umtiti header from a set piece. The stands hailed Deschamps as a tactical genius. I pulled out my phone, messaged a data analyst in Brussels, and three hours later published a piece built on a number that annoyed many people: France's expected goals (xG) was only about 0.8, while Belgium's was 2.1. The winning side defended; the losing side created more chances. Football hands the trophy to whoever scores, but it hands legend status to whoever knows how not to lose. That was the starting point for almost everything I have written over the following seven years. Modern football carries a silent value system taught from academy level: safety is wisdom, risk is naivety. When a team keeps a clean sheet, people call it character. When a team loses while attacking, people call it inexperience. These two labels are repeated often enough to become truth, even though they rest on nothing but the final result. I have spent hours in training sessions in Osaka and Saitama, logging every duel, and what I saw was coaches teaching players to avoid risk before teaching them to create. The result is a generation of good players who hide mistakes rather than make a difference. That is the foundation for understanding why the back-three trend has returned to the top, and why I do not believe it is progress. It is rarely a bold attacking choice; it is a shield. When a back four gets pierced a few times, the instinct of a manager under scrutiny is to add a centre-back, shrink the space, cut the goals conceded, and ease the pressure on his own seat. The goals-conceded data improves. The chances-created data worsens. But the second metric rarely makes the bulletin. The whole world praises good defending; I only see a team hiding behind fear. Read the 2026 France-Belgium match through three layers of data and it becomes clear. The first layer is the scoreline: France won, in perfect control. The second is xG: Belgium generated clearly higher-quality chances. The third is PPDA, a pressing metric, which shows France deliberately dropped deep and ceded the ball. That was a chosen approach, not an accident. Deschamps bet on a low-variance plan: few goals conceded, wait for one set-piece moment. It worked, and because it worked, it is assumed to be right. But a single sample does not make a philosophy, just as one clean sheet does not prove that defending is an art. I tried to test the opposite in a place where no match was being played. In 2026, when the J.League was suspended by the pandemic, most reporters went home to write predictions. I stayed in Osaka, following Cerezo Osaka through four months without football, logging how the club switched to online training with GPS data from smart vests. The 2026 season was wiped out, but I still held my bet on a second-tier side. When the ball rolled again, the way they reorganised their defensive structure without abandoning attack forced me to rewrite my own hypothesis: the issue was never the number of centre-backs, but the purpose behind the number. Deeper still, media and money reinforce the habit. A good defensive side sells stability to sponsors, to shareholders, to season-ticket holders. An inconsistent attacking side sells emotion but not certainty. So the bulletins tend to reward the side that errs least rather than the side that creates most. When a club announces an injury, it also picks a timing that suits its image, and the fans are left in the fog. That is why statistical breakdowns rarely reach the front page, even though they describe the match more honestly than the scoreboard. This is where I could be wrong, and I want to be clear about it. Disciplined defending demands real skill: holding distances, reading passing lanes, absorbing pressure in silence. Calling it cowardice is an oversimplification. If I argue against myself, I must concede that in a knockout tournament, where one mistake ends a four-year cycle, optimising risk is logical. Denmark at Euro 2026 is a fascinating counter-example: after Christian Eriksen's collapse, they went deep as a collective playing for one another. Opponents deliberately dropped deep out of fear of facing a team running on emotion, and that very caution handed Denmark control. The lesson is not whether defending is right or wrong. It is that we usually read one layer of data and turn it into morality. When I was thrown out of the press conference in Tokyo, my question stayed on the table. In 2026, after Japan lost to Syria in a qualifier, I stood up and asked the head coach directly whether he knew he was eroding Japan's attacking heritage. He walked out midway. I was warned, but the recording spread, and a week later a former international emailed me about divisions inside the squad. The question was not silenced; it just changed seats. Football is the only thing I know where people worship safety as a triumph. But a right question outlives a press conference. So what will change? I do not believe in grand tactical revolutions. I believe in managers who dare to keep a back four when criticised, who dare to accept a high xG while conceding once, and who dare to take responsibility for the process and not only the result. If you are looking for one signal to track this season, do not look at goals conceded. Instead, look at the chances your team creates in an even game, and ask yourself: is this team playing not to lose, or playing to win? People ask why, at 68, I still write as if the apocalypse were near. I just smile. Because every season, that question still has no answer.