The Cricket the Machine Cannot Read: Scorecards, Regional Commentary and the Politics of an Incomplete Archive

11:40 p.m. On a Mumbai digital desk, a screen turns green. There are six fields. Five of them read: not available. The sixth holds two words: cricket_asia. Title: blank. Source: blank. The list of information points: zero. Article type: unclassified. Time sensitivity: never assessed.

The dashboard did not turn red. It turned green — and that was the real find of the night.

I have seen screens like this. In 2026, I began my career on a Mumbai digital desk writing 250-word agency-fed match reports. For four years I was told that tactics were 'for the boys', in press boxes where I was often the only woman in the room. I learned the game in 250-word fragments, then started listening for the silence between them. That silence became my real pitch notebook — the small things that never made the agency feed: how the slip cordon drifted a step wider in a particular session, when the crowd's tone shifted after tea, how much slower a spinner's run-up became in the evening.

The dashboard that went green that night was listening to the same silence. The difference is that it did not hear anything. It simply reported that there had been nothing to hear.

Two stages, and the link that snapped

The architecture is simple. Stage one reads an article and extracts what is concrete: headline, publisher, date, player names, teams, format, numbers, and the sentences that can be read as claims. Stage two sits on those extracted fragments and analyses them: what the format was, how the pitch behaved, which phase decided the match, where luck ended and skill began, and what remains unknown.

The entire structure rests on one condition — that stage one returns something. On the night it returns nothing, stage two has two roads. The first is to invent a story that sounds plausible. The second is to state honestly that the record is empty.

The first road is fast, bright, and in this profession the most in demand. The second is slow, and it requires you to admit you know nothing. I have been in this trade for twenty-one years. I have watched people who took the second road get shown the door. I have also watched the ones who took the first road stay in the room a long time — and fill the record with things that never happened.

The real discovery that night was not technical. It was that when every content field collapsed, one field survived: cricket_asia. No language, no byline, no date, no format — only a geography tag. Which tells you something precise. What caught that article was not its subject. It was its channel. The classifier had read a URL or a feed category, not the body of the piece.

In isolation this could be one software's fault. In cricket journalism it is not isolated. It is a pattern, and the pattern cuts away exactly the part of the game that is already at the margins.

The scorecard nobody can read

Open any state association website and try to find a Ranji Trophy scorecard. You will find one of three things. A scanned PDF, its figures printed in a typeface that screen readers cannot parse. An image, from which no character can be lifted. Or an HTML table, which any conventional readability parser throws away at once, because those parsers are built for sentences, not for rows and columns.

The Ranji Trophy has been played since 2026. India's first Test was at Lord's in 2026. Eden Gardens has stood on its ground since 1864. The country's cricket rests on roughly nine decades of domestic competition. Every season produces hundreds of matches, four innings each, ten wickets and more than two hundred deliveries per innings. It is the largest and most continuous sporting record this country has — and in the digital world it is embarrassingly thin.

Now look from the other side. A long English-language feature, cleanly paragraphed, on a major publisher's site, is entirely machine-readable. Every claim can be lifted, every name linked, every number tabulated. The same report written in Marathi, Gujarati, Tamil, Bengali, Kannada or Telugu often sits on a page of small type, cramped columns, text wedged between advertisements and a newspaper scan. Its subject is cricket. Its form is a wall.

Which means that when a system decides what we know about 'Asian cricket', it is not reading the cricket most played in Asia. It is reading the cricket most easily parsed. And the most easily parsed cricket is usually the cricket with the most money and the most reach behind it.

Regional microphones, vanishing voices

I am a commentator, so I know where cricket's democratic reach actually begins. It does not begin in a board press release. It begins at a Tamil microphone, where a commentator calls a lower-order batsman by the name of his village. It begins in a Hindi live feed where you can hear the crowd breathe as the bowler starts his run-up. It begins in a Bengali studio where an old captain's story is born again in his own language.

The voices are growing. Their written record is not.

Here is something I learned sitting in the ground, something no database keeps. Crowd noise is a character with its own measurable presence. At Eden Gardens, when a large partnership breaks, the hush makes a sound that appears in no scorecard. At Chepauk, the rhythm of the crowd changes before the ball leaves the spinner's hand — a signal the local commentator catches and no algorithm ever will. I once, in a Mumbai newsroom at three in the morning, counted 47,000 spectators singing through a final whistle using only radio audio and a fan's live stream. The truest number of that night was not the count. It was the sound that would not stop after the last ball.

Now consider what happens when that sound passes through a speech-to-text tool whose training corpus never contained the language. The sound is neither cancelled nor recorded. It simply falls away — and the dashboard prints 'not available' in its place.

What does not enter the record does not enter history — and what does not enter history does not survive in memory.

Women's cricket, and promises left unkept

The Women's Premier League began in 2026 with five teams; Mumbai Indians won the inaugural title. It was the moment women's cricket in India finally got the broadcast weight that had been demanded for decades. The numbers appeared on television, audiences grew, and the board was right to call it a success.

My question is not about the broadcast. It is about after the broadcast. Where were the regional-language commentaries stored? Where will the scorecards, the pitch reports, the moments that a local commentator named in his own language, be found five years from now? If the season we are calling a turning point can be read only in English, only inside a streaming platform, only with the owner's permission — then how much of that turning point actually turned?

This is the same question I ask about the Ranji Trophy, and the same question I ask about every tour whose coverage thins out in the winter months. A fixture list is not a neutral document. It is a statement of institutional priority — who gets the day, who gets the night, who gets the camera, and who gets one line.

The scout's spreadsheet, and the darkness around selection

Here the matter gets more tangled, and this is the biggest question of the second innings.

I have spent years doing data work on digital desks and had long conversations with domestic players. What I took from those conversations is that India's bench strength is not a shortage of talent. It is a selection failure. And selection no longer works the way it did.

There used to be a zonal selector with a thick notebook, a few names, a few recommendations and a long memory. Franchise scouting has quietly replaced him — a spreadsheet on a laptop, video-tagging software, and a list of batsmen deemed likely to trend during a single auction week. The Right to Match rule, the base-price slabs, the retention regulations: together they have built an economy in which a player's price is set before his ability is settled, and in which that price is entirely public.

The reverse is also true. All of that data is private. The scout's report belongs to the franchise. The minutes of the selection meeting are recorded nowhere. Which is why we have mountains of data on how players performed, and almost no public record of why one was picked and another dropped.

Every transfer has a body, a biography and a bill. And when the basis of selection is filed neither in a newsroom, nor in a board document, nor in any public database, the player knows least about his own future and the audience guesses the most. Rumour grows in precisely that gap. I should admit that I am part of that gap: who briefs me, on whom I depend for information, and when and why that information reaches me — I cannot record any of it here, and until I can, my own account is incomplete too.

Why a green dashboard is more dangerous than a red one

Now the least comfortable part.

Whenever cricket and technology are discussed, we all share one fear — that a machine will start writing about the game. My fear is different, and I think more urgent: that a machine will stop writing about the game, and we will treat that as correct.

The damage is done when a system can read nothing and still appears successful. The report produced that night contained no invented name. It contained no fabricated figure. It said only 'not available' everywhere. And that is the most dangerous sentence of all, because somewhere downstream a review meeting can read it as 'no risks identified'.

Absence of evidence is absence of evidence. It is not evidence of absence of risk.

This domain has three separate formats — Test, ODI and T20 — which is why the ICC maintains three separate ranking tables. Numbers from one format cannot be carried into another. If an analysis lacks the courage even to state which format it was, then every sentence in it is suspect. The same applies to bowling figures, to selection, and to arguments about legacy.

And one thing belongs at the very top: cricket_asia is a geography tag, not the name of a story. There is no single thing called 'Asian cricket'. To assume an India–Pakistan fixture from a tag, or an IPL auction, or a Bangladesh board dispute, is not analysis. It is the laziness of guessing. And laziness is where blind spots harden into policy.

Whose record will it be

That green dashboard at 11:40 p.m. is still pasted in my notebook, because it left behind a question no software can answer.

When the next generation remembers this 2026 tournament cycle, whom will it remember — the matches whose scorecards can be lifted cleanly, or also the innings whose story was told into a single regional microphone and recorded nowhere? The final whistle does not close the story; it opens the door for questions. And my worry is that we will leave that door open, with no script written on it at all.