Five Columns, Four Empty: The Audit of a Football Report That Was Never Written

A file opened on a desk. It had twelve lines, and eleven of them said the same thing: not applicable.

This was a football report. Or rather, it was the shell of one — the kind of artifact that should have been produced in a first-stage deconstruction, after an actual article had been broken down into its component claims.

Title: N/A. Source: N/A. Type: unclassified. Core summary: blank. Information points: an empty list. Entities: "to be identified from the information points above" — except there were no information points above.

The only living thing left in the entire document was one English word: football.

I keep a database of things people swear never happened. This is the first time a document has arrived on my desk swearing that nothing happened at all. That is a form of humility, and it is rare. But humility and failure are printed on the same sheet of paper, and separating them is the reporter's job — not the system's.

So this report is not about a match. It is about the process that turns matches into reports, and the day that process threw up its hands.

First, what this thing actually was

What football coverage calls "deconstruction" is really an evidence table. An article, a press release, a federation statement or a match report is broken into pieces and its claims are placed in separate boxes — what is the claim, who said it, when was it said, which document carries it, and what is the basis underneath. Stage two joins those boxes into analysis. Stage three writes the piece.

In Indian football this process is not optional, because information does not arrive in one place. An AIFF statement, a state association circular, a club press note, a league media guide and a broadcast feed all speak different dialects and maintain different standards of honesty. The journalist's job is to stand all four in the same line.

I have a habit that dates to 2026. That year I was a 21-year-old journalism student in London who talked his way into a media pass for the FIFA U-17 World Cup in India. Fifty-two matches. I logged every card, every penalty and every video review into a spreadsheet. Referees had already awarded 14 penalties in the group stage; I published the tally, a Kolkata daily quoted it, and I spent the following month checking each decision against the Laws of the Game.

The lesson from that month was not technical. It was procedural: an empty box is not a neutral object. It is an alarm.

In 2026 I watched all 64 matches of the Russia World Cup and built an audit file of 29 penalties and 20 official reviews. That was the year I fixed my five-column template: minute, incident, on-field call, VAR outcome, correct decision. A twelve-page disciplinary consistency report came out of the same frame, and a London sports data agency handed me a three-month junior analyst contract on the back of it.

In 2026, from the empty stands of Project Restart to an empty ISL season in Goa, I started treating pitch-side microphone audio as primary evidence. Forty-seven incidents logged there, sixty-three in India. Every one with a minute, a Law and an outcome.

I am recounting all of this because the template is the subject of this piece. When the empty shell landed on my desk, I put it through my own five columns. The stadium was not empty — the document was, and so was its audio log line. It is the inverse case.

Before the verdict, the timeline. Minute by minute. No gaps allowed.

Column one: minute

Not applicable.

In football reporting a minute is not a time, it is an address. When I write that a penalty was given in the 67th minute, the reader can find that incident in a recording, watch it with their own eyes and argue with me. Take the minute away and there is no way to replace evidence with anything except another assertion.

This document has no minute. It has no date. It has no publication time. A document that does not know when it was made cannot tell you how long its contents were true.

That is not a small defect. A timestamp is the thing that makes a claim retractable. A claim with no time cannot even be corrected.

Column two: incident

Not applicable.

This sounds absurd, and it is the single most important finding. The document's entire purpose was to say what happened. It said nothing happened. It did not say whether that was because nothing happened or because nothing was found.

The distinction matters enormously. "Nothing happened" is a fact. "Nothing was found" is a failure. Filing both under N/A hands the responsibility for separating them to the next stage — and the next stage, as we will see, was not equipped for it.

Column three: on-field call

Something is recorded here, if you look closely.

The on-field call was this: the pipeline issued an output. It did not claim to have finished. It called itself unclassified and left the boxes empty. That is honesty, and it should be logged as such.

I know how rare that honesty is. From July 2026 to March 2026 I was Director of Communication and Media at the All India Football Federation — a journalist moved inside the federation. In those two and a half years I saw hundreds of templates with every box filled and every box filled without a source. The FIFA suspension of August 2026 and the compliance chain beneath it existed in documents and in dates. But the communication built around them was written in a different dialect, one in which every blank space is filled with a reason, because a blank space looks like failure.

That is my real concern. Empty boxes do not merely need attention. An entire industry exists to fill them.

Column four: review outcome

Not applicable — because a review requires an incident. That is true of VAR as well. If the on-field referee did not see it and the fourth official did not log it, the review room knows nothing. Everyone remembers the whistle; nobody remembers the fourth official. It is football's oldest blindness, and it exists in writing too.

Column five: correct decision

Here I stopped looking at the table for a while.

Because the only column in the whole document that contained anything was not filled by a pipeline. It was filled by a critic — by the source analysis itself, which concluded that the artifact is not analysable, that it is an empty shell, and that no honest evaluation is possible.

I pulled five columns, and only one of them told the truth — and it came from outside.

In a football match this would sound strange. It does not happen to referees. A referee at least has ninety minutes on a watch. Here there was no watch.

The line that points to something which does not exist

Now the actual defect.

In the entities field, the instruction read: "identify from the information points above." The list of information points was empty.

This is not a data problem. It is a structural defect in the template. The template requested an output field whose input was not guaranteed to exist.

Which means — the template contains no gate capable of stopping an empty input. And whatever a gate does not stop will look complete at the next stage.

I read it through a referee's lens. A Law only functions when someone applying it can see a violation. If the Law says "inside the line" and nobody has drawn the line, the Law is inert no matter how clearly it is printed. A rule is only as strong as the person holding the stopwatch. A field is only as strong as its mandatory status.

There was no mandatory status here. There was only a polite clay figure that had broken itself and was informing the room.

One word survived: football

The source analysis raised a possibility worth repeating, because it is the most practical thread in the whole episode: the football label may have been inferred from metadata or a URL, or it may be a hard-coded default.

Neither option is good.

Here is the counter-intuitive part. Having no label looks worse than having one. Having one unsupported label is worse than having none at all. With no label the system does nothing. With a wrong label the system does a great deal, with total confidence, and is caught very late.

An example from my own file. In the 2026 audit I tagged a penalty as "correct." Three months later I reopened the column and found I had tagged the right decision but cited the wrong Law — I had written "serious foul play" on a handling offence. The tag was full. The tag was wrong. Because it was full, the error survived nine months.

That is the single-document trap, and I apply it to myself first. A document is never a truth; it is a testimony. It has to be cross-checked. If four of five columns are empty and the fifth was filled from outside, the honest move is to say four are empty — not to treat the fifth as settled.

False precision, and the version of this that could have existed

The source analysis listed three risks that would have materialised had it guessed: an invented author stance, invented viewpoints, and false-precision credibility scores assigned to sources that do not exist.

That list is correct and incomplete. The fourth risk is larger: downstream contamination.

A fabricated analysis is more dangerous than an empty one, not only because it is false but because it is harder to catch. An empty box looks empty. A filled box looks exactly like a correct box. By the time the article publishes, no individual sentence will read as wrong. It will all read as plausible. And plausible falsehood is what an editor cannot stop without an outside party walking the claim back to a source.

I watched a version of this trap for years inside Indian football administration. On sponsorship, licensing and promotion files, the phrase that always caught was "due process was followed." The question is which process, on which document, signed by whom, and who received the licence afterwards. I am in favour of auditing the result, not merely the filing.

And I have to apply that strictness here to myself. It would be easy to write that the pipeline performed correctly by returning N/A. That is proceduralism as an alibi. A system capable of emitting an empty shell into a desk, with nothing stopping it, does not belong in the category of systems that work. A system that needs a human to stop it is failing on its own terms.

The contrarian angle: the fault is not that the pipeline broke. The fault is that N/A was allowed to travel.

The first suggestion from anyone who sees this file will be to re-run the pipeline.

That is the wrong diagnosis.

Consider it in football terms. Suppose a foul is committed in the 82nd minute. The referee misses it — that happens, referees are human. But then something else happens: the incident is never written on the fourth official's sheet, and the minute is left blank in the match report. Reading it the next day, you would conclude nothing happened in the 82nd minute. One miss occurred; five links in the chain failed.

This is a specific failure profile, one I encounter in my five-column template every season: the first error can be stopped; the eighth never can.

Here the pipeline did its job. It returned what it had, which was nothing. The real failure was that a file marked unclassified reached a desk with no sender, no date and no gate. Which means there was no gate at all.

The second contradiction is harsher and bothers me more: an empty shell is more honest than a full one, and the industry prefers a full one.

Any content operation measured on volume will not see this distinction. Two pieces beat one; a piece that looks complete beats one that looks empty. Under that incentive, the empty shell's honesty is an accident. It is evidence that the system has not yet learned to lie, not evidence that it chose to tell the truth. A system that prefers integrity and a system that simply has not yet covered its blank page produce identical output on day one — and very different output after three months.

The third contradiction is for readers who will object that I have spent five thousand words on a document containing nothing. The question is not what is inside the document. The question is what kind of system manufactures such a document, and whether that same system is writing your match report next. Writing about an empty document is not saving it. It is auditing the class of which it is a specimen.

What survives from the source's own table

Beneath the source analysis was a small table — goal and recommended action, three rows. Row one: to analyse a specific article, re-run Stage 1 with real text. Row two: to debug the pipeline, treat this output as a failure artifact. Row three: to test the analyzer — confirmed, the analyzer should refuse to produce substantive output on empty input.

All three are sensible. All three miss something.

Row two comes closest and is still half-finished. "Treat it as a failure artifact" means we throw the file away and build a similar one again. A bin is not a debugger. The real question is which gate the artifact passed through, and why it passed.

I remember the line number in the 2026 spreadsheet. Its use was not in counting penalties. Its use was that when the number fourteen appeared, its abnormality was visible. Fourteen forced me to read further. Numbers only work when they carry context; a number alone is worth no more than an empty shell.

The most useful quantity in this file is numerical and there is exactly one: nine of ten fields marked not applicable. That is a meaningful figure. It tells you this is not a conclusion but an interruption. A report in which valid fields are outnumbered by invalid ones is not ready to publish; it is ready to be returned.

Where the gates belong

First gate, at the extraction stage, before this file is even born: any deconstruction with a blank title, source, date or information-point list should be rejected outright. That is not severity, it is a minimum standard. I think of it like a scorecard: if the batsman's name is missing, the innings is void, however neatly the runs are written.

Second gate, inside the template: every field should declare its source field. A field whose source does not exist should fail loudly rather than pass quietly as a blank. A line like "identify from the information points above" should not exist in any template, because it is a promise the template cannot keep.

Third gate, on the desk, in a person: "no source" is never a neutral note. Printed in front of a reader, it is a claim. A piece with no date tells the reader this could have been written at any moment. A piece with no source tells the reader someone said it somewhere, and you do not need to know where.

That is my most basic objection, and it applies to the journalism I grew up in as well. Five years ago I began checking decisions against the Laws of the Game because I assumed that if I did not, nobody else would. That was not arrogance; it was fear. Today the same fear applies to the whole system, and faster than before. When content pipelines produce articles in minutes, the old question — is this true? — becomes a different one: can this claim be attached to a document, and which document?

Until that question has an answer, an empty box is an honest box.

One thing I am changing after this report

I have had a template since 2026. It is no longer a five-column template. I have added a sixth, in first position: who filled it.

The five columns tell you what is true about an incident. The sixth tells you who knew it, and whether that person was in the room when it happened. The day we start filling that column is the day no empty shell reaches our desk — because in front of a blank field we will write, without knowing why it was written.

Until that day, when the next fully-populated report arrives, do not ask first what it says. Ask first which column was filled by whom.

Four of my five columns are still empty. That is the answer, and I will carry it forward — with the next date attached, not without one, the way this report arrived.