74% Possession, 0.8 Expected Goals, and What Germany's 0-2 Defeat to South Korea Taught Us About Data

Kazan, June 27, 2026. I was 14, sitting in front of an old TV at home in Saigon, notebook and pencil in hand. Germany vs South Korea ended 0-2. The stats showed: Germany 74% possession. I stared at that number, then at the scoreboard, and felt deceived. A team that held the ball 74% at the World Cup just lost 0-2? That night I wrote three pages, haunted by one question: why did traditional numbers betray us? Germany fired 26 shots. South Korea fired one shot that mattered. The ball only remembers one. To understand this match, you need the context of Group F at the 2026 World Cup. Germany entered the final group game with one point after two matches: a 0-1 loss to Mexico and a 2-1 win over Sweden, courtesy of Toni Kroos's stoppage-time free kick. They had to win. South Korea had lost both previous games to Sweden and Mexico, with virtually no chance of advancing. But having nothing to lose liberated them. Coach Shin Tae-yong set up a defensive side always ready to transition. Germany, meanwhile, still trusted their possession game. They pushed both center-backs near the halfway line, unintentionally turning the space behind their defense into a promised land for Korean counterattacks. The surface stats favored Germany: 26 shots to 6, more corners, 586 successful passes to 197. Any observer would conclude Germany dominated. But the quality data told another story. Germany's expected goals were just 0.8; South Korea's were 1.6. In other words, a team that attempted less than a quarter of Germany's shots produced double the expected goals. Simply put, Germany shot often but shot poorly, mostly from outside the box or in tight spaces. South Korea's shots came from quick counterattacks, striking before the German defense could recover. Germany's problem was not only shot quality but also how they moved the ball. Mesut Özil and the midfielders kept shifting the ball sideways in front of South Korea's box. Those passes raised possession stats but never broke down a disciplined defense. South Korea, by contrast, needed only one or two passes to reach the German goal. They did not need to hold the ball long; they needed to strike at the right moment. Kim Young-gwon's opener in the 90+3 minute came from a scramble after a corner. VAR confirmed the goal. Many called it luck. But look at the whole match: it was a reward for patience. South Korea absorbed pressure for 90 minutes without losing their structure. When the moment arrived, they still had energy left. Germany, drained by meaningless passing, collapsed into panic. Son Heung-min sealed the 2-0 in the 90+6 minute with a perfect counterattack — the clearest embodiment of a team that controlled the game by giving up control of the ball. There is a contrarian angle here. Analysts often call a team lucky when they win with fewer chances. But examine the mechanism. Germany pushed high because they had to win. South Korea dropped deep because they had nothing to lose. Both behaved according to the logic of the match. South Korea's higher xG was not a random statistical blip. It was the inevitable result of a tactic designed to turn ceded possession into an advantage: concede the ball, absorb pressure, open space behind the opponent, and create higher-quality chances. When a model has an explainable mechanism, calling it luck is a refusal to understand the game. Conversely, judging by the 74% possession figure and concluding Germany deserved to win confuses activity with quality. The stadium was not empty that night. But the principle I keep using in my work still holds: an empty stadium does not destroy football; it simply reveals variables we used to ignore. In Kazan, the overlooked variable was the psychology of a frightened champion. Now, writing this, I am a sports data analyst. I still revisit this match — not to prove South Korea great or Germany doomed, but to remind myself that every number needs context. This week I received an analysis report whose data fields were entirely empty. The analysts did not invent conclusions; they wrote plainly: insufficient information, cannot assess. At first glance, that was a failure. I think it was one of the most honest acts I have seen in the industry. When there is no data, do not fabricate a story. Staying silent before numbers that lack context is not the analyst's failure; it is the choice of someone who once looked at a 0-2 scoreline in Kazan, saw 74% possession, and understood that football always finds a way to humble those who think they have figured it out. I look at xG, then I look at the score, and I learned not to trust either — but I never stop asking questions. Because the question, not the hasty answer, is what keeps us honest with the game.