Germany 0-2 South Korea: The PPDA Warning Germany Ignored in Qualifying

Kazan, 90+3, June 27, 2026. Kim Young-gwon stood at the far post, the ball breaking loose from Manuel Neuer, who had abandoned his goal to join the attack. A second earlier, Germany were still camped in the opposition half with one last hope. A second later, the goal stood empty, and Son Heung-min only had to roll the ball into the net. I was sitting in the press area of Kazan Arena, opening the spreadsheet I had carried since early May, and the first cell I looked at was not the scoreline. It was the PPDA cell, the metric that measures how many passes an opponent completes per defensive action.

For Germany, that number was 11.3. I had written it into the spreadsheet back in March.

PPDA is easily misread. It does not measure running intensity, it does not measure kilometres covered, it does not measure successful duels. It measures something else: how many passes a team is willing to allow before it intervenes. Eleven passes before intervening means that team is standing and waiting. The leading pressing sides in Europe in 2026 sat between 8.5 and 9.5. Drawing on my experience of watching Germany's qualifying campaign directly, I collected data from ten matches and rebuilt every defensive sequence. The average: 11.3.

In March 2026, I wrote a prophecy. All of Germany laughed. When I published my "slow-burning bombs" list before the tournament, Germany topped it. The reaction from colleagues in Russia was not warm. One called me a monk obsessed with numbers. Another said it plainly: this is the reigning champion, they have Kroos, they have Hummels, they have Neuer, and numbers do not know how to play football. They told me I was causing trouble. I was only reading the ending a few months early.

I left the piece untouched.

Data context needs to be stated before we go further. The 2026 World Cup was played in full stadiums, on a four-day turnaround during the group stage, in a Russian June that swung wildly between host cities. Germany played three matches in three different places: Moscow, Sochi, Kazan. Every conclusion below is tied to that setting, not to a continuous league season.

And every prophecy carries an error margin. I am writing this part before presenting the results, so the reader knows I did not edit the spreadsheet after learning the score.

The chain of evidence starts with the first match. On June 17, 2026 in Moscow, Germany lost 0-1 to Mexico. Hirving Lozano's finish on 35 minutes came from a counterattack after Germany lost the ball in midfield. The notable part is not the goal conceded. It is that for 90 minutes Germany never closed down Mexico's midfield quickly enough to force a long ball. PPDA in that match: 12.1.

The second match, on June 23, 2026 in Sochi, Germany beat Sweden 2-1 through a Toni Kroos free kick on 90+5. The scoreline reads like a comeback. The data reads differently. Germany shot more, but most of those efforts came from outside the box with low conversion probability. In my model, Germany's total chance value across the first two matches sat below the threshold a champion normally produces in a group stage. Kroos's free kick was an individual moment, not an output of the system.

Here is the core point: the scoreline hides the problem, the index does not. A team can win through a moment and lose through an entire structure. In qualifying, Germany had been living on those moments: late goals, long-range strikes, set pieces. Nobody went back to check whether the system still generated clear chances.

Structurally, three things collapsed at once.

The flanks. When Germany lost the ball, both full-backs had pushed too high and nobody covered. Mexico exploited it in the first match, South Korea exploited it in the last. In Kazan, Germany had to push Neuer up to the halfway line just to gain an extra passer, the signature of a team with no other option left. The empty goal was a logical consequence. The accident was not there.

Transition speed. From losing the ball to rebuilding the attack, Germany needed more passes than their opponent to cross the pitch. That meant the opposition only had to drop into position. South Korea did roughly that in the second half in Kazan: sit deep, hold the distances, concede the ball, wait.

Shots conceded. I built a warning threshold: any team allowing an average above 12 opposition shots per match in qualifying belongs in the high-risk group. Germany did not cross it. That is why my model placed Germany at medium risk, and also why I only dared to talk about difficulty in the group stage, not about elimination. I had gone further than the data allowed.

Three matches in sequence: PPDA 12.1, then 10.8, then 10.9. The index did not deteriorate sharply between matches. It was bad from the start and stayed there. This is the hardest kind of data to read, because it produces no clean break for people to point at and name. It is a slope.

There is another explanation, and it is almost certainly partly right too. The 2026 World Cup belonged to teams that had played together for too long. Germany brought to Russia a block of players who had been through four major tournaments. Experience does not compensate for legs. But saying only that skips the harder question: why did a team with so much data about itself fail to read the signs of decline?

This is where correlation diverges from causation. People usually attribute psychological saturation or complacency when a reigning champion falls. Those explanations sound good but cannot be tested. What can be tested is structure: the midfield no longer covered the space behind it, and the attack no longer created high-quality chances. Those things are measurable, and they had worsened before the tournament began.

On the other hand, this must be admitted. If Kim Young-gwon does not score on 90+3, Germany only need to beat Sweden and draw with South Korea to advance, and my model gets called wrong. Data does not predict a header at the far post. Data predicts that a team like that will have to live on moments, and when no moment arrives, that team will end up staring at an empty goal.

From the Bundesliga to the World Cup, I look for the same thing: a truth that can repeat itself.

The signal for the next cycle is not in Kazan. It lies in which team, after a failed tournament, dares to change its structure instead of its personnel. Every crowd is wrong. The only thing that is not wrong is probability.

Where could the assumptions be wrong?

PPDA is calculated on qualifying data, where opponents are far weaker than at the finals. Pressure in qualifying is lower, so the index may be skewed downward. I also folded every shot into one general model, without separating set pieces from open play. And most importantly: I have no data from the dressing room, the thing a spreadsheet never sees. The spreadsheet is an altar, and I offer myself to every number on it, but an altar cannot answer every question.