The Empty Cell: When Cricket's Data Desk Returns Nothing

One night in February, in a room in Hyderabad, a JSON sat open on a screen. Title: N/A. Source: N/A. Summary: empty. Article type: unclassified. And at the bottom, the list of information points — empty. One thing in the entire structure was still alive, a small tag: cricket_asia.

That night reminded me of my 2026 register. Hardbound, crooked handwriting, and in it 2,140 balls from Hyderabad's Ranji Trophy campaign at the Lal Bahadur Shastri Stadium. The most useful habit in that register was not that I wrote down every ball. It was that I kept a separate mark for a ball that was bowled and produced nothing — a dash. And a second mark, a blank, that meant I was not there to see it. Two different marks, and keeping them apart was the most honest part of the job.

The pipeline that now carries almost all cricket reporting has collapsed those two marks into one character. And that one character is quietly eroding thirty-five years of journalistic discipline.

Context: from the press box to the data desk

Through the 1970s and 1980s a cricket story was built from three things — the reporter's eye, the scorebook, and the next morning's paper. When I moved to The Hindu in 2026 the machinery was the same, only the prose was cleaner and the space longer. The scorebook lay on the desk and nobody touched it; the desk man looked at it and decided what the match had been about.

Now it is inverted. The modern cricket desk does not watch the match; it watches the feed. And the feed arrives in three layers. First, the board's official scorecard, which at Test and international level is reasonably reliable. Second, the broadcaster's graphics feed, which runs off cameras — where there is no camera, there is no feed. Third, the aggregator's API, which chews the first two layers and renders them as one more or less identical screen.

When all three go quiet, the news does not stop. It merely loses its documentation. And without documentation, journalism takes one of two roads: it reports that nothing happened, or it begins to fill the gap with inference. Neither is acceptable, but one of them is far more dangerous.

This is not hypothetical. The Ranji Trophy carries more than thirty-five teams and upwards of a hundred and fifty matches a season, and those without broadcast coverage do not enter any public ball-by-ball index. The match happens, teams change, bowlers change, a young man's career is made or broken — and its only evidence is a PDF scorecard that somebody sometimes uploads and sometimes does not.

Last season I ran a small test. I took the report of a Ranji match with no broadcast, and set its scorecard against the local paper's account from that day. The ball count matched. The field placements did not. Which fielder stood where, in which over, was nowhere. The thing that won the over had vanished from the record for good.

Core: three meanings of nothing

This is where the real argument begins. An empty cell is treacherous precisely because it can represent three different things. First: nothing actually happened — the ball was defended, no run came, no event worth logging. Second: something happened and was not captured — the camera was pointed elsewhere, the scorer had gone for tea. Third: it was captured and not parsed — the parse failed, the language detection failed, the page was an image.

In all three cases the screen shows the same thing: zero. And zero, as our data desks are fond of saying, never lies. It also never tells the whole truth.

My 2026 register still travels in my bag for this reason. The great advantage of paper is not that it is durable; it is that empty space does not crawl the way it does in a feed. When I left a blank after Azharuddin's name at eighteen, I knew what the blank meant. When an API today returns an empty field for a batsman's fielding positions, I cannot know whether he was not standing there or whether the collector never looked.

This is why I still keep the pencil. The pencil remembers what the software forgets.

Take one number from that register, the one that has never stopped bothering me. Arshad Ayub was a good off-spinner, not a miracle, a patient one. On my grid his cost was 1.9 an over with a slip and a short leg, and 3.7 with a sweeper. The variable that produced the difference was the field, not the ball.

A Wagon Wheel today shows where the runs went. It does not show where the fielders stood. The evidence of the mechanism that decided the outcome is simply not in the record. That is the root of my complaint: the pipeline gave us better run rates and worse reasons.

The next link in the chain is the argument we have all been having for eleven years — April 2, 2026, Wankhede. Chasing 275, India were 114 for 2 in the 22nd over when Mahendra Singh Dhoni walked out ahead of an in-form Yuvraj Singh.

I coded that innings live, having tracked the same decision across sixty-one previous chases. Dhoni took 14 from his first twenty balls, then 77 from his next fifty-nine, working the vacant long-on region while Muralitharan and Mendis kept shortening their lengths. India won by six wickets with ten balls left.

I wrote it as a decision tree — three options in front of him, the field each demanded, and why the third held. The report was built on a choice, not an investigation. And to this day I cannot tell you how any database recorded that choice. In the database the event became "Dhoni 91 not out" — the choice turned into a number, and the number erased the choice.

That is the structural flaw. The pipeline records outcomes, not options. And in cricket, everything worth analysing is an option.

Go back to the tag that survived that night: cricket_asia. One word that encloses an entire geographic market. In the rights business, categories move like money. In August 2026 the Board of Control for Cricket in India allocated the IPL's 2026-2027 media rights — roughly Rs 23,575 crore to Disney Star for television and roughly Rs 23,758 crore to Viacom18 for digital, close to Rs 48,000 crore together. Every category in that announcement carried a price. No category carried a line about what happens to the matches that do not fall inside any of them.

From here my hollowed pyramid argument starts, and I have repeated it for two decades. The IPL did not kill Test cricket in India — the grounds still fill. It made the Ranji Trophy administratively invisible. Invisible is worse than dead, because the dead are mourned and the invisible are never accounted for.

The consequence is that the hardest team in world sport is now India's second XI. Anyone can call that a surplus of talent. I call it a selection problem wearing the face of a surplus. When a batsman's best innings came in a match with no ball-by-ball record, that innings does not exist in the selector's list. He is not dropped. He is never added.

The winter of December 2026 brings me back to this place every time. The board had imposed a life ban on Mohammad Azharuddin — a ban the Andhra Pradesh High Court set aside in 2026. That winter I left the press box for eleven months and, from my desk, coded 1,180 one-day scorecards from 2026 to 2026, hunting a statistical fingerprint of a fixed match: run-rate stalls between overs 20 and 40, anomalous fields, abrupt strike rotation.

I found nothing robust, and I said so, in "The Unfixable Statistic" in Economic and Political Weekly. Somewhere in that winter, the statistic stopped obeying me.

That episode still teaches one thing, and it connects directly to the empty JSON. Had I not published my null result, three people would have mined claims out of the same data within a decade. If a negative result is never made public, the empty space fills itself — and the loudest voice fills it first.

The modern data industry breaks that rule. It publishes conclusions and never its own nulls. Which is why I say a model that cannot be wrong is not a model. It is a religion.

The Women's Premier League began in 2026, and after the 2026 women's World Cup the number of people talking about the level rose sharply in a place where almost nobody had been looking. I put those two facts together deliberately, because the room there was empty and therefore new structures could be built. In the men's game the room is full, and a full room leaves no space for an empty cell.

Contrarian: the data desk has won the argument

I should say this at full strength, because hiding it would be dishonest — the data desk has won the tactical argument. When an analyst explains why an old ball outperforms a new one in a fourth innings, or which slot produces the highest strike rate for an inswinging yorker, he is not speaking in newspaper cadence. He is speaking from ball-by-ball data. And he is right. The old commentary-box generation that dismisses him is throwing away a real advantage: it cannot name a bowler's changing angle mid-innings, so the captain repeats in the third over a mistake the data desk caught in the warm-up.

But there is a narrow delta, and it deserves a fair hearing. The data desk won the argument about reading the game and lost the argument about recording it. The better the models became, the narrower their input. Almost every model now runs on the same three fields — runs, wickets, balls. Ninety per cent of what happens on a cricket field is not in those three fields.

The cost is concrete. A captain burned a review last season because his slip catch was logged in a feed as not taken, when the ball had gone to a fielder who had been moved into that position in the same over. The chart did not record the switch. This is not a scandal story. It is a story of small empty cells, which in aggregate change a whole season.

And here I will admit one of my own weaknesses: I trust the crowd, and that trust sometimes does the thinking for me. But on this point, the terrace at Chepauk was naming a player two seasons before the index found him. The crowd was watching what was not being logged, because the crowd was sitting in the ground the camera never reached.

Takeaway: the next time you see an empty cell

The next time a feed shows you a zero, ask one question: did the match disappear, or did the camera? If it was the camera, the match was played and no evidence survives. And a match with no surviving evidence never reaches the list everyone is staring at before the next selection meeting.

Do this once this season. Pick a Ranji match with no broadcast. Take its scorecard, take the local paper's report from that day, and count the discrepancies between them. I make no claim that you will find a conspiracy. I only say that the count you end up with is the most neglected statistic of this entire era.

The ledger never lied. It just never told the whole truth.