Where the Stats Lie: Switzerland 3-4 Germany

By Wandrille P. , March 27, 2026

Tags: germany , where-the-stats-lie

Switzerland 3–4 Germany · Friendly International · Basel · March 27, 2026

Germany outshot Switzerland 22 to 7. They won the xG battle 2.50 to 0.57. They had 10 corners to Switzerland’s zero. By every volumetric measure, this was a match Germany dominated so completely that a comfortable win was the statistically expected outcome — not a 61st-minute comeback from 2-1 down, not a 79th-minute equalizer threatening to deny them, not a final scoreline that required one player to score twice and assist twice to drag his team over the line.

The stats were right about who controlled the game. They were entirely wrong about how difficult it would be.

What the numbers said

Germany’s xG of 2.50 against Switzerland’s 0.57 represents a 4.4-to-1 ratio in expected output. Switzerland’s keeper Gregor Kobel made 5 saves — Oliver Baumann, his German counterpart, made zero. On the xGOT model (expected goals on target), Germany registered 3.06 against Switzerland’s 1.05. In short: every model predicted a comfortable German win by two or three goals.

The actual match produced seven goals, two Swiss leads, a Fabian Rieder crossbar rattle that could have made it 3-1, and a Joel Monteiro equalizer in the 79th minute that briefly suggested the models had been embarrassed entirely.

Where the model broke down

The disconnect between xG and scoreline has a structural explanation: Switzerland scored three goals from seven shots, converting at a rate of 43%. Germany scored four from twenty-two, converting at 18%. That gap — Swiss clinical efficiency against German profligacy — is the core reason the match felt like a contest long after the statistics had settled in Germany’s favor.

But the efficiency gap alone doesn’t explain the two Swiss leads. That requires looking at how Switzerland scored, not just how often.

Their first goal, Dan Ndoye’s 17th-minute near-post finish, came from a fast break — a low-xG, high-execution scenario. Ndoye received the ball from Granit Xhaka wide on the left, cut inside, and placed a left-footed shot across Baumann before Germany’s backline had recovered its shape. The xG on that chance was negligible. The finish was perfect.

The second Swiss goal, Breel Embolo’s 41st-minute header, was the result of something models don’t capture at all: a defensive individual error. Jonathan Tah — who had equalized for Germany with a header minutes earlier — was beaten in the air by Embolo at the back post from Silvan Widmer’s cross. Tah is 6’4”. Embolo got ahead of him anyway. No expected goals model accounts for a lapse in aerial concentration from a player who had just proven his aerial dominance from the other end of the pitch. Switzerland’s second goal was a human error masquerading as a tactical one.

And then there was Fabian Rieder’s crossbar at 43’. A shot from outside the box, hit on the break after a fast counter, that struck the frame when Baumann was beaten. Had it gone in — and it was 3 centimeters from doing so — the match narrative changes completely. Germany would have gone into half-time 3-1 down with Switzerland barely threatening. That single moment, which appears nowhere in the xG model, contained the entire drama of the evening.

The player who broke the model: Florian Wirtz

Match stats: 90 min · 2 goals · 2 assists · 3 shots (3 on target) · 77 passes · 69% accuracy · 5 key passes · 7.6 tackles won from 9 duels · Rating: 10

The xG model cannot produce a 10/10. But it can explain why Wirtz got one, and it does so by revealing the absurdity of his evening: he scored two goals that had no right to go in and assisted two that required pinpoint execution from teammates who rarely deliver it at that level.

His first goal, in the 61st minute, was described by FotMob as a shot “from a difficult angle and long range on the left to the top right corner.” The key phrase is difficult angle and long range — both are xG penalties. The model assigned the chance a low probability. Wirtz curled it over Kobel and in off the underside of the bar. It was not a lucky shot. It was a technically extraordinary one.

His second, in the 85th, was cut from the same cloth — a right-footed finish to the top right corner from the centre of the box, assisted by Pascal Groß. The quality of the finish, not the quality of the chance, is what made it a goal.

What the stats don’t show is that Wirtz was also Germany’s metronome in possession — 77 passes at 69% accuracy, 5 key passes, the best passing volume of any attacking player on the pitch. He pressed, he recovered (4 tackles, 1 interception), and he was the only player in either squad who combined creative output, defensive contribution, and clinical finishing in the same 90 minutes.

There is a secondary story buried in his rating: Wirtz has managed only 4 goals and 2 assists in 27 Premier League appearances for Liverpool this season. In his last two Germany games, he has contributed 2 goals and 4 assists. The international stage is where he is still, unmistakably, the best version of himself.

The player the stats exposed: Leon Goretzka

Match stats: 80 min · 0 goals · 0 assists · 4 shots (0 on target) · 21 passes · Rating: 5.6

The model would have predicted a positive contribution from Goretzka: high shot volume, central position, minutes at the heart of Germany’s midfield. What it couldn’t predict was the quality of those shots. Four attempts, zero on target. His 21 passes in 80 minutes — compared to Stiller’s 72 in the same half-field role — reflect a player who was not a priority destination in Germany’s build-up.

The 5.6 rating is a fair verdict. He was substituted at 80 minutes for Anton Stach, a telling decision from Nagelsmann in a match Germany needed to close out. His inefficiency in front of goal was the clearest statistical exposure of the evening: four shots generated zero xGOT, meaning the model agreed with the eye test. He wasn’t unlucky. He was simply off.

What to take away

This match is a useful reminder that xG models are better at predicting season-long patterns than individual games — especially when individual brilliance and individual error are concentrated at both ends of the pitch in the same 90 minutes. Switzerland scored three goals from chances totaling less than 0.6 xG combined. Germany needed one player to score twice from low-probability positions to overcome it.

The model said Germany would win comfortably. It was right about the result and wrong about almost everything in between.

Written by Wandrille P — football analyst specializing in data-driven match analysis and creator of Ultrivia.