When the stands are empty: home advantage, empty data, and the limits of football analysis

Summer 2026, in a data office in Seoul, I opened a spreadsheet and thought my system had failed. K League 1 had to play in stadiums without fans because of the pandemic, and the league's average home advantage fell from 1.48 points per match to 1.12 points per match. A sample of 200 matches. Not small. But what made me stop was not the data; it was a familiar and frightening feeling: I had been ready to write a conclusion from data I had not finished verifying.

I tell this story because it lands exactly on a situation every football analyst meets: you are handed an empty dataset, and you feel the pull to fill it with intuition. How you handle that moment decides whether you are an analyst or just a storyteller with charts.

Context: home advantage is not a constant

Home advantage in football has never been a law of nature. It is a sum of at least five variables: the crowd, familiar turf, travel, pressure on the referee, and player psychology. When the stands emptied because of Covid-19, most of those variables vanished at once. That is a rare chance to isolate the crowd variable from the rest, something normally impossible.

In 2026 I was 25, an analyst at a sports data company. When the pandemic forced K League 1 to play without fans, I had to handle a paradox that broke every precedent I had learned. At first I set the result aside. I spent three weeks re-running models, cross-checking week by week and team by team, filtering out the pandemic factor, before I dared publish an internal report.

My process had three steps. First, I split the data into two periods: before and after fans returned in part. Second, I removed matches where a club changed manager mid-season, because tactical change can create noise. Third, I compared home win rate, draw rate and average goals per match to see which variable moved most. The result showed the home win rate falling most, while average goals barely changed.

What stood out was that average goals per match did not fall. If the crowd only created psychological pressure, we would expect home teams to score as before but concede more. In fact both ends worsened, showing that crowd pressure acts on the whole structure of a match, not only on the defence.

The final conclusion was not about how much home advantage fell. It was that the change came from the absent crowd, not from the manager. An off-pitch variable rewrote the result on the pitch. When data and observation conflict, the analyst must verify before concluding, and must accept that sometimes the truest answer is not enough data.

That is why I added a crowd-pressure variable to every piece I write about home form. And why, whenever I receive an empty analysis, no title, no source, no numbers, I do not rush to build a pretty story. I return it with one line: insufficient information, cannot assess.

Core: reading space as a unit of time

My career changed in 2026, when I was 23. Before the World Cup semi-final between Croatia and England, I confidently predicted Croatia would press high through the trio of Modric, Rakitic and Brozovic. In reality they pushed up for exactly 18 minutes, then dropped deep, let England control 57 percent of the ball, and still won 2-1 by exploiting the space behind England's back line.

I wrote a 1,200-word self-critique, admitting I had judged people instead of space. When Croatia came back, I understood that football is not mathematics but ethics. From then on I set a hard rule: every analysis must contain at least three figures about space, distance or line spacing. I stopped writing on feeling about player reputations and instead used coordinate maps for every notable passage.

In 2026, at 27, I was assigned to follow Morocco's entire World Cup run. They conceded only one own goal in the group stage; the other two came in the semi-final against France. I spent four weeks re-watching every match, counting how often Hakimi and Mazraoui tucked inside, recording an average gap of 12.4 metres between the two central midfielders, and seeing that the open space in front of the box was always covered by an inverted triangle.

The Morocco matrix was not there to block the ball but to choke the opponent's time. That was when I moved fully into the language of geometry: triangles, central trapezoids, player density per 10x10 metre square. I always place the system above the individual, even when writing about defensive stars like Hakimi.

For example, when analysing a counter-attack, I measure the distance from the final pass to the centre of the box, the opening angle of the defensive triangle, and the time from ball recovery to shot. Those three figures tell a story that a player's name cannot.

But because I systematised so far, I also learned the limits of systems. I believe in structure, but structure exists to collapse; a good analyst is one who predicts the exact point of collapse. And the point of collapse is usually not in the spreadsheet. It sits in the moment a player decides something against the model, and that cannot be measured by xG or PPDA.

Every tactical diagram is a confession: what a manager fears, he hides. A back three is a confession of fear of being cut open on the flanks. A deep double pivot is a confession of fear of losing control in midfield. When you can read that fear, you do not need player names to understand the match. You only need to know what the team is hiding and what the opponent will try to break.

The same logic applies to the transfer market. Transfers are the market of regret: those who wait win, those who rush pay. But to know who is rushing, you need data on fees, contract structure and timing. A club that buys a striker on deadline day usually pays 20 to 30 percent above market value, simply because it has run out of time. I once followed a deal in which a club bought a centre-back for more than 30 percent above valuation, only because an injury to a key player appeared on the final day. That decision was reactive rather than strategic. And reactions are always expensive.

Contrarian: the trap of the empty analysis

This is what I want to say plainly. In sports data analysis, the costliest mistake is not a wrong prediction. The costliest mistake is drawing a very confident conclusion from a data foundation that does not exist.

I have seen reports dozens of pages long, full of tables and diagrams, whose source, when traced back, was empty underneath: no original title, no source, no verifiable information. The frightening part is not the emptiness. The frightening part is its perfect appearance. A nine-dimension framework, neatly presented, can make readers believe a real analysis took place, when in fact it is only an empty frame polished to a shine.

That is when data gives us a map, but only chaos shows the real path. And chaos cannot be created from nothing. You cannot analyse a match you never watched, a team whose name you never had, a season for which you never had data.

Esports and football share one root: finding the smallest gap between the two biggest mistakes. But that gap only appears when you have a full record of both mistakes. Without data, you do not see the gap; you see only a blank.

Sports media runs on an emotional cycle. After a win, every manager's decision is called bold. After a loss, the same decisions are called errors. The analyst's job is to separate the emotional cycle from real data. Otherwise he is merely translating noise into a chart.

This is also where the views I hold become useful. I hold that referees treat giants and small clubs differently not as a conspiracy, but as real crowd and media pressure. I also hold that an amateur side reaching a final usually owes it to the luck of the draw and one explosive match, not proof that a system works. Neither view can be verified by an empty analysis. To talk about referee pressure you need data on cards and penalties across seasons. To talk about tournament surprises you need data on brackets and form.

If the data foundation is empty, the only honest conduct is to admit it is empty. Football without fans in 2026: every tactic was still correct, yet no tactic still mattered. I reuse that image for data: every analytical framework is still theoretically correct, but no framework matters without input.

Takeaway

So where is the lesson? In this: the value of an analyst is measured not by how many conclusions he draws, but by how many times he dares to say there is not enough data. In a sports industry that runs on speed and engagement, refusing to conclude is a small but necessary act of resistance.

I still remember the feeling of three weeks re-running the K League model. I did not find what I wanted to find. I found what I needed to know: that an off-pitch variable, the absence of the crowd, could rewrite the whole table. Had I published that result after one day instead of three weeks, I might be more famous, but wrong.

That is the standard I keep for every piece I write: verify first, conclude after, and never let a beautiful frame replace a verified fact.