Zero Information Points, Eight Sections: The Cricket Report That Should Never Have Been Filed

A document landed on my desk last week. Eight sections. A risk matrix with six categories. A governance checklist running five rows. Star ratings across four dimensions. A transmission map with upstream, midstream and downstream nodes all drawn in. In format, it was exactly the kind of report that gets written about the fifth day of a Test, or an IPL auction, or a disputed run-out.

Every cell carried the same sentence: “N/A — insufficient information.”

That was honesty. And that honesty is the biggest story in the document — because the pipeline that produced it allowed an input with zero information points to pass forward. The report did its job. The system did not.

We recognise this distinction in cricket. A third umpire can reach the correct decision and the match can still end with the wrong result, if the protocol lacked the gate where it should have stopped. The rulebook said one thing, the replay said another — I went looking for the third answer. This is a Laws & Officiating case file, only it belongs to a news desk instead of a match.

To see how this is a cricket story, start with the demand the coverage sits under.

An IPL season is 74 matches — ten teams, league stage and playoffs. On top of that sit the Women’s Premier League, domestic tournaments, India’s bilateral calendar and the overseas leagues: BBL, SA20, ILT20, CPL, MLC. Every match needs a preview, a mid-innings report, a match report, player ratings and a next-match preview. In India alone the seasonal total runs into the thousands.

Now the second number: how many journalists are actually present at the ground. At a domestic Test, the cricket writers in the press box are often in the low double digits. The IPL is better staffed, but not at a ratio that produces 74 previews and 74 reports.

So a large share of what a reader consumes is built from three inputs: press-conference quotes, agency wires, and a draft produced by a person or a script sitting at a desk.

That is not a moral panic. It is a supply-and-demand description. The Associated Press began publishing automated game recaps in 2026, starting with Minor League Baseball on the Automated Insights engine. Those recaps carried accurate scores, hits and key plays, because the data was present in the input.

That is the real question. The issue is not that automation has arrived. The issue is which gate has been placed in front of it.

Now the forensic examination of that document. Three findings, each at a different level.

First: the missing validation gate. The input behind this report had a title of N/A, a source of N/A, an article type of “unclassified”, and an information-point count of zero. It was not poor material. It was no material. The output still cleared the schema.

That is the most dangerous class of failure — the kind that passes the system’s own tests. Imagine a scorecard where every batsman has nothing but a dash beside his name, but the format is complete, the columns are all present, and the header even has room for the team names. On paper, that scorecard is valid.

A schema can check whether a field exists. It cannot check whether the field contains anything. That needs a separate rule — a very cheap rule, one line long: no payload with zero information points passes forward. Had that rule existed, this report would never have been built, and a clean error would have surfaced instead, fixable in twenty minutes.

I have re-watched a phase of play forty times so you only have to read it once — here it ran in reverse. The document wrote “insufficient information” in dozens of places and still could not be made worth reading.

Second: the taxonomy mismatch. The input carried the domain label cricket_world. The output contract required Cricket. The two were arriving from different taxonomy versions.

Cricket has a direct parallel, and it is not trivial. The MCC owns the Laws of the Game, but every tournament applies its own playing conditions. The same run-out can produce two different answers under two different condition-sets, especially around the use of the soft signal. If the third umpire is reasoning from an old edition while a new one governs the field, the error belongs to the protocol rather than the umpire — and nobody reviews it, because nobody checks whether the same book is open on both sides.

The same thing happened here. The label did not even establish whether the subject was international cricket, a franchise league or administration. cricket_world is a topic tag, not a format tag. It could not separate Test from T20 conclusions, or home from away performance.

Third: the false negative. The report carried six risk categories — sporting, personnel, commercial, rules and integrity, public opinion, and systemic. The first five read: “could not be assessed.”

That is the correct vocabulary, and it is the safeguard. Had those rows instead read “no risk identified”, a reader would have taken them as a clean bill of health. The distance between the two is the distance between two different worlds.

Cricket produces this distinction daily. A report says “no evidence of a no-ball was found.” The question is whether anyone looked at the front foot. If nobody looked, then “no evidence found” does not mean the bowler was clean. It means no check was performed. That is this document’s genuine contribution — it kept “not assessed” apart from “nothing found”. Most cricket coverage erases the difference.

Now what the document does prove. A title of N/A, a source of N/A, a type of unclassified, and not one resolvable entity — that combination points to a failure at the fetch layer rather than the reasoning layer. The underlying article probably existed and simply never reached the pipeline.

Three explanations sit at the top: the page refused with a 403, a paywall or a geo-block; the page was JavaScript-rendered with no server-side text at all; or the input was never prose — a scorecard, a video page, a social post. That distinction matters because it fixes the location of the repair. A reasoning fault would mean replacing the model. A fetch fault means checking HTTP status, rendering and paywalls.

Here I have to state plainly what the available material cannot show. This document does not establish what the underlying article was about. It could have concerned anything. All that is proven is that nobody read it — and that a format was filled in which reads as though somebody had.

Now the part where I have to argue against the fastest reaction. The instinctive conclusion is that the template is the enemy, automation is the enemy, and these eight-section reports have hollowed out cricket writing. I find that wrong, and close to inverted.

Look at what the document actually did. It wrote “insufficient information” in dozens of places. It refused to issue star ratings. It refused to assign a ranking. It refused to say everything was fine.

Cricket journalism’s real disease is the refusal to refuse. A desk on deadline never leaves that space blank. It pastes last match’s preview onto the next one, changes two names, and writes the line that fits every team: “the bowling attack needs discipline.” This was happening before automation. It was happening when every desk had a human at it.

So the template is not the fault. The fault is that the pipeline’s default was “generate”, and there was no default for “decline”. In cricket’s language, it had no threshold at all — and most arguments are really about where the threshold sits.

DRS writes that threshold down and calls it clear and obvious error. The on-field decision survives until the evidence is clear and obvious enough to overturn it. Umpire’s call sits a step further back: the technology reports that the ball is clipping enough of the stump that the matter is too close, and the field call stands.

This pipeline had no such threshold. A zero-information payload crossed over. That was not umpire’s call. That was the situation where the umpire never saw the ball and still retained the authority to give a decision.

And this is where the letter of the law is easy, and the spirit of the law is where careers get decided. The letter says the output format is correct. The spirit asks how the reader was supposed to know there was nothing inside it.

My proposal is cheap and dated today, so that nobody can later claim they saw it and stayed quiet.

First: the first gate of any cricket content pipeline should be a minimum of three information points and at least one named entity. Without those, the report does not get built. Second: publish the failure rate by source domain. Sites that keep failing at the fetch stage will quietly disappear from coverage, and nobody will notice, because disappearing makes no noise.

Third, for readers: put one question to any cricket preview. Which delivery, which innings, which clause, which frame. If there is no answer, you are reading a template, not a report.

The question is not whether a script can write cricket. The question is who stops it from stepping into the gap, at a moment when the press box is shrinking and the demand for copy keeps rising. That document stopped itself. Three years ago a person would have sat in its place, and there was no rule stopping that person at all.