The Empty Column in a Football Data File: When an Analyst Must Learn to Say “Not Enough”

An empty column in a data file makes no sound. It throws no error, turns no red, blinks no warning. It simply sits there, silent, among hundreds of filled numbers, waiting for someone patient enough to notice its absence. That night, before the squad deadline, I opened the scouting file for a player the club was targeting. The column for “pressures inside the box” was empty. The column for “touches in the second half” was empty. I could guess. I could interpolate from the previous three matches. I could make the report look complete and convincing, so that no one would ask another question. Instead, I left the cells blank, and wrote in the conclusion a sentence no coach wants to hear: “Not enough data to conclude.” Three days later, I received a short email. The club had signed a different player, based on a more “complete” report — smoother, more confident. I don't know whether that player was better. None of us did. And that is exactly what I want to say in this piece. Over the past fifteen years, football has gone through a quiet but total revolution. Analytics departments have sprung up in almost every club in the major leagues. Metrics such as expected goals (xG), passes allowed per defensive action (PPDA), and positional tracking data have become part of everyday language in transfer meetings. A modern sporting director can hardly decide on a signing without opening a spreadsheet. Based on my experience watching matches, I can see a paradox that is rarely spoken about. The more data there is, the greater the pressure to give an answer. And when the pressure is great enough, the gaps in the data get filled — not with truth, but with narrative. People are uncomfortable with emptiness. When a number is missing, our instinct is to invent it, or worse, to invent a plausible-sounding explanation for its absence. There is a technical detail few fans know. For the same shot, two different data providers can give two different xG values, sometimes differing significantly. Their models differ in how they handle defender positions, shot angle, strong foot, and even match context. So when you read “this player's xG is X,” whose model are you reading? That question is almost never asked in television debates. I once watched a match in which the positional data system failed for the entire second half. The post-match analysis was still published in full, with a small note at the end: “Second-half figures are estimated.” Almost no one read that note. What people remember is the conclusion at the top: player X ran twenty percent less than his average. That number may be right. It may also be the product of an interpolation algorithm that guessed wrong. But right or wrong, it became part of the player's file, a piece in a transfer decision that might take place months later. I started the blog “Data Corridor” in 2026, when I was a second-year student, with an analysis of seventeen key passes by a Premier League midfielder. I was told I was “a girl who knows nothing about football.” I didn't delete the post. I added three more charts. But after many years, I realised that adding data is only half the story. The other half, the harder half, is knowing when to stop. This is the crux I want to go deeper into: in football analysis, the hardest skill is not finding the number, but knowing when the number does not exist. Take a striker. He scores five goals in thirty matches. The stat sheet says: failure. The press says: finished. But if you open a deeper layer of data, you see his xG is eleven. He created chances good enough to score eleven, but scored only five. That gap is not in his feet. It is in the quality of the passes, in the positions his teammates choose, in how the team builds play before the ball reaches him. There are numbers that do not appear on the stat sheet; they live between two touches. A season is not the sum of thirty-eight matches, but the repetition of seventeen forgotten passes. But the reverse story is also true, and few are willing to say it. A striker scores fifteen goals from an xG of only eight. Is he “lucky” or “clinical”? The honest answer is: not enough data to tell them apart. In a small sample, luck and skill look identical. People only learn the difference after the season ends, when it is too late to change a contract. And during that waiting period, the market will pay for a beautiful story, not for a truth left open. I once witnessed an internal debate that lasted weeks over a defender. One side presented figures showing he was one of the most prolific clearers in the league. The other pointed out that he cleared so much because his team allowed opponents to reach the box far too easily. Both were right. The “clearances” number means nothing without knowing the context in which it occurred. And no data provider sells you that context. They only sell you the number. I have a persistent observation about goalkeepers, and I will say it plainly. Distribution is being sanctified. Clubs pay enormous sums for a goalkeeper who can play accurate long passes, while his basic reflex metrics — the thing that actually decides points in big matches — are declining without anyone noticing. A spectacular save is replayed on television and spread across platforms. A well-chosen position that avoids the need to save is not. Post-shot data can show who is truly good, but it demands a patience the transfer market does not have. In a corridor, if you only look toward the light, you will miss what stands in the dark. In 2026, working as a part-time statistics assistant for a football site in Singapore during the World Cup, I coded every action of the Spain 3-3 Portugal match. Cristiano Ronaldo reached a top speed of 9.8 km/h, below Portugal's team average of 11.2 km/h. If you looked only at the speed figure, you would conclude he was slow. But all five of his shots on target came from situations close to goal, where speed is no longer the most important variable. My analysis of the “unusually narrow pitch” drew over two hundred thousand views. The lesson I took was not “Ronaldo runs slowly,” but “the speed number does not answer the question I was asking.” Then there is refereeing. The subjective space in VAR is larger than people think. The “clear and obvious error” standard sounds objective, but it is itself a vague clause. For the same collision, two different VAR teams can reach two opposite conclusions, and both can justify themselves with “video data.” Here, data does not remove ambiguity. It merely dresses ambiguity in a more scientific coat, making viewers believe a correct answer exists, when in reality only a decision was made. And there is a deeper layer, where data is used to conceal rather than to illuminate. Satellite club systems allow the giants to sidestep domestic training rules. A young talent in a small league becomes a “satellite asset,” registered in one place but playing in another. On paper, everything is legal. In the spreadsheet, every number matches. But behind those numbers is a young player, and his story does not appear in any data file. This is where I want to go against the crowd. We usually worry about missing data. I believe the greater danger is data that is present but wrong, or data that is right but placed in the wrong context. A number filled into an empty cell by guesswork carries the full credibility of a real number. It does not announce itself as a hypothesis. It sits there, in the spreadsheet, waiting to be cited, waiting to be quoted in an article, waiting to convince a club president that this money is worth spending. My job, for many years, has been to resist that instinct. And I have learned that it is not rewarded. Clubs prefer a confident wrong answer to an honest, hesitant one. A coach needs to know how the opponent will play on Saturday, not that the sample is still too small to be sure. The humility before uncertainty that I regard as the core virtue of an analyst looks like weakness in a meeting room waiting for a decision. The irony is that this very confidence is the source of the biggest mistakes. A report that admits “I don't know” makes people ask more questions, check more sources, wait for more data. A report overflowing with numbers makes people stop, trust, and sign. Clubs dissolve, football stops. But data never stops telling stories. The problem is that it can tell a wrong one, and that wrong story will outlive the club that produced it. I do not think the future of football analysis lies in collecting more data. We already have far too much. It lies in learning to respect the gaps — to tell the difference between “I don't have the data” and “the data says there is nothing.” Between those two sentences lies an entire decision-making culture, and perhaps an entire next generation of analysts. And sometimes, the most valuable thing in a spreadsheet is precisely the empty cell that no one wants to fill.