Empty Stands, Empty Stat Sheets: Lessons from 64 Bundesliga Matches Without Fans

On 16 May 2026, Signal Iduna Park did not hold a single soul. Borussia Dortmund crushed Schalke 04 four-nil in the first Ruhr derby after the pandemic. I sat in my rented room in Nha Trang, rewound the tape three times and typed every metric into a spreadsheet. What made me stop was not the scoreline. It was the silence. Before the pandemic, the home win rate in the 2026-20 Bundesliga hovered around 42.7%. After the stands were closed, across the 64 matches I collected, that figure fell to 31.3%. The average xG of the home side dropped by 0.19 per match. Dortmund's PPDA in the away role improved by 0.8. Four numbers, and one question: where does home advantage actually live? I started this work in 2026, when I was nineteen, a statistics student in Nha Trang. In round 8 of the V-League that season, Hanoi FC held 61% possession and took 15 shots, yet their xG reached only 0.8. Ho Chi Minh City FC managed exactly 3 shots, with an xG of 0.6, and the match ended 1-1. That day I understood the first rule of this trade: possession does not produce evidence. To touch the truth you must add running distance, duelling positions and chance quality. Each match took me four hours of manual notation. Nobody paid me. But it was my first standardised process, and it has stayed with me ever since. In 2026 I scaled the model up to the World Cup. Before the tournament I published a warning about the German national team: their average PPDA had risen from 8.1 in 2026 to 11.6 in qualifying; high-speed running distance had fallen by nearly 18%, worst of all in midfield with Toni Kroos and Sami Khedira. My conclusion then: Germany would be eliminated in the group stage. The forums called me a number freak. I took that as a compliment. Germany finished bottom of Group F. At the end of 2026 I was tasked with building the prediction model for the Qatar World Cup, standardising 68 teams into 12 metric clusters. Morocco averaged only 28% possession yet forced opponents to shed 0.35 xG per match; goalkeeper Yassine Bounou posted a PSxG overperformance of +2.4. Argentina were the only side to keep PPDA below 8.0 in every match. I was challenged for removing Brazil from the candidate list, but both teams I selected reached the final. Then May 2026 arrived. When the Bundesliga reopened in silence, I realised I was holding an enormous natural experiment. Not one match. Sixty-four matches. The same league, the same laws, almost the same players — differing in exactly one variable: noise. My method was simple. I split the 2026-20 season into two phases: before and after the stands closed. For each match I recorded the result, xG, xGA, PPDA for both teams, yellow cards, fouls and the number of referee decisions inside the penalty area. I then compared the two phases' averages, checked standard deviations, and excluded matches with excessive squad churn. The result did not live in a single number. It lived in the whole cluster. The home win rate fell 11.4 percentage points. Home xG dropped 0.19. Yellow cards for away teams declined slightly — the detail I watched most closely. Referees under crowd pressure tend to lean toward the home side; with the stands empty, that lean vanished. The PPDA of strong away teams such as Dortmund improved by 0.8, meaning they pressed earlier and with more confidence once nobody was booing behind them. Three variables, three different mechanisms, all pointing one way: most of home advantage is a social signal, not a tactical one. But that is a probabilistic conclusion, not a law. The greatest temptation in this trade is jumping straight from correlation to causation. Empty stands, fewer home wins, therefore home advantage is an illusion. It sounds very tidy. Yet at least four alternative hypotheses explain the same dataset, and I am obliged to list them before concluding. The schedule was compressed. Teams had to play every three days, rotating their line-ups, with fitness declining unevenly. The substitution rule rose to five, completely changing how coaches manage a match. After two months off, teams lost match sharpness at different rates depending on their individual training conditions. And splitting the season shrank my denominator — 64 matches is enough to see a trend, not enough to lock a conclusion. Had I published only the 31.3% figure and called it proof, I would have sold my own data cheap. There is another lesson here, and it is the one I actually want to make. While collecting data, I ran into matches whose stat sheets were nearly blank: no detailed xG, no positional tracking data, no touch maps. A novice analyst fills that gap with feeling — this team is dominating, that defence is loose. I do not. I mark the cell as insufficient information, leave it blank, and note the reason. A blank cell does not mean a team played clean. A blank stat sheet does not mean nothing happened. It only means I have not seen it. And not having seen it is entirely different from having checked. This is the mistake I encounter most often in Vietnamese football arguments. A team goes three matches without a booking, and people say they play fair. It could be that the referee let things go, or that the data was never recorded. A striker fails to score for five rounds, and people say he has declined. It could be that his xG still sits at 0.45 per match and the opposing goalkeeper is on an abnormal run of saves. The match ends, but the data is still there. An empty stadium does not need spectators; it needs an analyst willing to look. And when the stat sheet is blank, the right thing is not to invent a story to fill it, but to say plainly: I have nothing to say here yet. I wrote blogs from a rented room in Nha Trang; now probability takes me everywhere. The only constant is the rule: whatever the numbers say, that is what I write — even when the answer is silence. Next round, try one small thing: pick a metric column you usually skip, and ask yourself whether it is empty because of the team, or empty because of you.