The Empty Data Sheet: Lessons From a Football Report With Nothing to Say

The PDF file weighed 2.4 megabytes and ran to forty-two pages. It had a table of contents, evaluation tables, an analysis-conclusion cell, a risk-warning cell, and a glossary of technical terms on the final page. The analytical framework was crafted down to the last ruled line. And every cell was empty. Not empty in the way someone forgets to fill something in. Empty in a systematic way: every cell was stamped with the same sentence - insufficient information to assess. The chapter on tactics was empty. The chapter on club finances was empty. The chapters on league standings, on rules, on the dressing room, on public opinion, on the transmission chain of the entire football industry - all empty, all with the same sentence. What made me stop was not the emptiness. My profession lives on emptiness. What made me stop was the perfect frame around it. A frame complete enough to be cited, professional enough to slide straight into a transfer meeting, and polished enough that nobody asks a question. There are numbers that never appear on a stats sheet; they live between two touches of the ball. But this time, even the corridor between two touches was empty. That file came out of a two-stage analytical pipeline. Stage one does the decoding: read a source article, extract information points, identify the entities named - which player, which club, which competition - assess source quality, and determine whether the matter is time-sensitive. Stage two takes that output as its raw material and runs it through nine analytical dimensions: tactics and technique, club finance and the transfer market, results and the opinion cycle, league landscape and team positioning, rules and compliance, management and the dressing room, risk profile, media narrative and expectations, and finally the transmission chain of the whole industry. A pipeline like that is nobody's private toy. In the 2026 season, almost every club in the top divisions has a data department, whether it is two people or twenty. That department takes requests from the coaching staff, from scouts, from the board. And it has to answer. Answer before the match, answer after the match, answer during the transfer window, answer when a player gets injured and nobody knows how long he will be out. The crucial point is this: stage one did not run. Or it ran and failed. Or it ran and returned an empty file. No article, no title, no summary, no information points, no entities. Only cells marked "not applicable" and a few instruction lines left over in the template, such as "identify from the information points above" - while above there were no information points at all. And stage two, instead of stopping, ran all nine dimensions. It built tables, built a risk matrix, built a transmission-chain diagram, built a glossary of technical terms. It did everything a professional report is supposed to do, except one thing: say something. Based on my experience of watching matches, I have seen thirty-page scouting reports on a player the author had never watched live. I have seen comparisons of advanced metrics assembled from two different data sources that did not even share a definition of a key pass. But I had never seen a forty-two-page document say it knew nothing - and still be presented as though it knew. An analytical pipeline fails in three ways, and each leaves its own trace. The first: the source article cannot be retrieved - the server does not respond, the page is blocked, or the article has been taken down. The second: it is retrieved but cannot be parsed - an unusual format, mangled encoding, or content sitting inside an image rather than text. The third: it is parsed but no entity can be extracted - no player name, no club name, no competition name. Those three leave three different traces. But there is one shared trace, and it is the most dangerous of all: the form still gets filled in. It is not empty like a blank page. It is empty like a tax return that has already been printed, stamped, and neatly stacked, missing only its contents. Among those forty-two pages there were lines like "identify from the information points above" sitting inside the "related entities" cell. That is not an analytical result. That is an operating instruction left in the template, forwarded downstream unchanged. Which means the step before it either never ran, or ran without anyone checking. In statistics, a missing value is not zero. A player with no data on ball recoveries is not a player who is bad at recovering the ball. A match with no PPDA data is not a match without pressing. PPDA - the number of passes a team allows per defensive action - only means something when you know how it was measured, over how many minutes, in what game state. But in the day-to-day operation of football, missing values are almost always treated as zeros. Metric leaderboards still rank. Charts still get drawn. Reports still go out. And the reader - a coach who needs a decision in forty-eight hours, a technical director weighing a contract - has no way to tell a real metric from an empty cell that has been coloured in. This is where an empty report becomes more dangerous than a wrong one. A wrong report can be caught by cross-checking. An empty report, presented to standard, makes no claim that can be caught. It does not lie. It simply lets the reader lie to himself. In June 2026 I worked as a part-time statistics assistant for a football website in Singapore. My job for the Spain 3-3 Portugal match in Sochi was to code every action, every pass, every shot. When the dataset finished running, one number made the whole team stop: Cristiano Ronaldo's top sprint speed in that match was 9.8 km/h, while the Portugal team average was 11.2 km/h. It is very easy to build a story out of that number. Very easy, and completely wrong. Because the same dataset also showed that all five of his shots on target came from situations close to goal, in a range where sprint speed is never triggered. He did not need to run fast. He needed to stand in the right place. When Arnold Schwarzenegger says "I'll be back", he is not talking about speed. Nor is Ronaldo at 9.8 km/h. My analysis that day included a section on the unusually narrow width of the Sochi pitch, and how it squeezed the space Portugal had to defend. The piece drew more than two hundred thousand views. But the lesson I kept was not the view count. The lesson was: if I had simply exported the raw dataset and let it speak for itself, it would have spoken wrongly. In 2026, football stopped because of the pandemic. The club where I was doing my data internship was dissolved. In a period with nothing to analyse, I volunteered to do performance analysis for a women's U19 national team - a team that played only twelve matches all year. Their goalkeeper had a penalty save rate of 43%. In professional football, that number sits in a zone that borders on fantasy. But what I found was not in any data export. It was somewhere else. I sat through eleven opposition penalties in slow motion. She was not guessing the direction of the ball. She was reading the shooter's belly step - the rotation of the hip before the foot met the ball, a movement shorter than a tenth of a second, with no column in my coding sheet to record it. Clubs dissolve, football stops. But data never stops telling stories - including the stories it was never programmed to record. The team's coach said something to me that day that I have carried for six years: you see what other people do not see. I do not think it was praise for my eyesight. I think it was praise for my capacity to endure not seeing - and to keep looking anyway. There is an almost perfect parallel to that empty report file, and it sits inside the Laws of the Game. The VAR intervention threshold is set by IFAB using the phrase "clear and obvious error". The phrase sounds very solid. It has two affirmative adjectives; it seems to draw a line anyone can see. But in practice, "clear" and "obvious" are not technical categories. They are judgements, and they are always made by a person, in a room, under a specific pressure. In more than a decade of following refereeing controversies, I have seen the same challenge read as "clear" in one league and "not clear enough" in another, in the same season, under the same law book. Nobody is lying. It is simply that nobody has defined it. And when a system runs on undefined phrases, every conclusion it produces carries an empty space behind it - a space named after the person who decided. That is why I keep an odd habit: whenever I read a refereeing report, I read the reasoning section, not just the conclusion. The conclusion is the published part. The reasoning is the part that shows how much emptiness sits behind it. xG - expected goals - is not a number. It is the output of a model, and that model is built on thousands of shots coded by human beings. Every time someone assigns a shot to a "big chance" or "small chance" zone, that person is making a subjective judgement. Accumulate a few hundred thousand of those judgements and you get a metric that looks perfectly objective. UEFA's FFP and the Premier League's PSR are the same thing at a different layer. They are financial rule systems operating on accounting data - and the accounting data of a football club is among the hardest data in existence to standardise, because it depends on how commercial revenue is recognised, how contract amortisation is allocated, how loans between owners and clubs are booked. What all these systems share is this: they are only as reliable as their input validation. And input validation is the least discussed step of all, because it does not produce a chart. It only says yes or no. It only says: does this data exist. Back to the forty-two-page file. What I want to say is not that it was useless. It has one very specific value, and that value lies in the fact that it did not fill itself in. If stage two of that pipeline had chosen to invent a club, a player, a transfer fee, a tactical system, it would have produced a document that looked far more useful. It would have numbers. It would have charts. It would be shared. And it would poison every decision made after it. Its refusal to do that is the single bright point in the entire document. There are numbers that never appear on a stats sheet; they live between two touches of the ball. And there are empty spaces that do not live between two touches - they sit right there on the stats sheet, in the exact cell somebody should have filled. A season is not the sum of thirty-eight matches; it is the repetition of seventeen forgotten passes. The same is true of a report: its value lies not in what it concludes, but in what it repeats consistently enough for you to trust it. But I do not want to stop there, because that comfortable story would hide the more uncomfortable part. This document does not exist in a vacuum. It exists because there is a market that pays for documents that look complete. Club boards want reports before the deadline. Scouts want documents to defend their decisions. Agents want numbers to negotiate with. Journalists want quotes. None of them wants to receive a forty-two-page file saying there is nothing to say. Silence does not sell. That is the whole problem. In this industry, a report that says "insufficient data" is treated as the analyst's failure, not as a finding about data quality. Analysts are judged by page count, chart count, and the smoothness of the prose. And so the incentive structure does not sit on the side of honesty. It sits on the side of confidence. I know this from myself. In 2026, as a second-year student, I wrote my first piece on the Data Corridor blog about Mesut Ozil's seventeen key passes in the Premier League. I showed that Arsenal's xG ranking dropped in the matches he did not start. The response came fast and it was not gentle. One comment on a large football forum that day consisted of a single line: what does a girl know about football. I did not delete the post. I added three more charts, with per-match data notes. But if I had chosen that day to add a conclusion I had no basis for - just to make the piece look a little fuller - the pressure would have won. That is why I read that empty report file with understanding, not contempt. Someone in that chain stopped at the right moment. But it must also be said plainly: stopping only at the end of the road is far too late. What should have stopped was a gate in the middle - right where the input data was empty. Another trap deserves a mention: do not use uncertainty as shelter. "Insufficient information" is a legitimate answer if and only if you have tried enough ways to obtain the information. If you call everything uncertain because you are too lazy to check, you are not humble. You are simply someone who has not done the work. There is one more layer, and it concerns the markets that get least attention. In Southeast Asia, where I live and work, a great many clubs run their data operation with one person, or with nobody. They have no budget for a two-stage pipeline. They have a scout doing two jobs, a coach doing two jobs, and a spreadsheet. It is precisely in those places that a beautifully presented empty report is most dangerous - because there is no second person to cross-check, and no third person to say that this number is meaningless. The next cycle of the season will begin, and it will again produce thousands of analytical documents. Among them will be files with exactly the beautiful frame of the one I just read, and they may be missing exactly one thing: data. The one thing I want to carry into that cycle is a single question, placed in front of every table: where did this data come from, and who checked it. In a corridor, if you only look toward the light, you will miss what is standing in the dark. But if the corridor is empty, the first thing you should do is not to imagine a light. It is to switch one on, and confirm that it is empty.