The Empty Cell: The Cricket Data Nobody Ever Collected

On Tuesday morning a file arrived at my desk. Eight sections, rows of cells beneath each heading, and inside every cell the same three words: insufficient information. No format, no venue, no innings, no player name, no assessment of timeliness. The article I was supposed to analyse never showed up. At first I read it as a failure. Two days later I understood it was the most honest document ever produced about cricket.

Because the most dangerous thing in cricket is not a wrong number. It is a number nobody ever collected, which everyone nevertheless discusses as settled knowledge. If an empty cell stays empty, the system survives. If it gets filled with a story, that story becomes the reference point for the next decade. I am not going to open with the twenty-two matches; I am going to open with one empty cell, because in every broadcast booth in the world the same thing is happening every night — someone is filling the cell, and nobody is asking where the filling came from.

Cricket does not suffer from a shortage of numbers; it suffers from the numbers that were never generated and then became the basis of decisions.

The conventional framing says cricket now lives in the data age. That is not wrong, only incomplete, and the incompleteness is the dangerous part. The truth is that cricket operates in two separate data economies, and there is no bridge between them.

The first economy is men's international cricket. Since the early 2000s, essentially every Test, ODI and T20I has a preserved ball-by-ball record. What happened on each delivery, who was fielding, who bowled which over — all of it. Broadcast cameras, Hawk-Eye and layered scoring systems have produced a near-continuous archive. In this segment the analyst's job is easy: ask, and you receive.

The second economy is enormous and almost invisible. The ICC has more than one hundred members, of whom only twelve play Test cricket. Domestic cricket in the rest of those countries, Under-19 tournaments, a vast share of women's cricket, and nearly all associate matches before 2026 exist as scorecards and little else. Who scored how many, who took how many wickets, who dropped a catch — nothing. Which over turned the match, which bowler cracked under pressure, which fielder ran three metres to save a single — none of it exists.

Standing between these two economies, cricket journalism makes a decision every day, and almost always makes it unconsciously. The analytical vocabulary developed for the first economy gets applied to subjects from the second. The result is a sentence with a scientific shape and no foundation. "His economy has risen over the last five matches" — fine, but against which teams, on which pitches, for how many overs? Those cells are blank. Nobody looks at blank cells, because the sentence already appears complete.

I came into this profession by a different road. In 2026 I took a bus from a small town to Dhaka and talked my way into an unpaid video-coding role at a club. Twenty-two matches, by hand, alone. One thousand one hundred and forty possession sequences, forty variables each. That spreadsheet produced something nobody on the coaching staff expected: sixty-one per cent of goals conceded arrived within twelve minutes of a turnover in their own third. The head coach ignored the number. The assistant did not. But the lesson from those four months was not about numbers. It was about cells. The value of a table lies not in its filled cells but in its list of empty ones. As long as you know which cells are blank, you are winning. The day you forget, your table stops being data and becomes the skeleton of a story.

Applying that lesson to cricket stopped me in three places.

The first stop: May 2026, Malahide. Ireland playing their first Test against Pakistan. In the second innings Kevin O'Brien made 118 — Ireland's maiden Test century. Pakistan won by five wickets. Those facts are complete, verifiable, and they are what everyone wrote.

But the question nobody asked: what was O'Brien's best innings? It looks like an easy question and it is actually unanswerable. Where there is no ball-by-ball record, there is no scale on which to rank "best." Ireland played first-class cricket for two decades before Test status, played the Intercontinental Cup, played World Cups. Scorecards survive; context does not. Which pitch, how much wind, how much pressure — nothing. So how would I claim the 2026 World Cup win over Bangladesh was not his finest hour, and that 113 against England in the 2026 World Cup was? I cannot. Claiming it would be a lie.

Something strange happens here. When data is absent, people fall back on instinct, and instinct always rates the most recent event highest. The 2026 innings became the centre of discussion because it is the only one with a camera, a highlights package, a historical label attached. Two decades of innings disappeared for want of data. The real damage of an empty cell is this — it does not steal information, it steals attention.

The second stop: December 2026, Chittagong, against India. Mohammad Ashraful, seventeen years old, 158 not out. The youngest Test centurion, a record at the time. The scorecard records everything: runs, balls, fours, sixes, how many against which bowler.

What the scorecard does not record is the physical condition behind that innings. How a young body performs, what fatigue it tolerates, what load it absorbs — none of it was captured anywhere. Sports-science data collection had not begun in Bangladesh. Analyse that innings and you hold an outcome without a process.

The effect is long-term. We read the career curves of Bangladesh's 2000s batters through an incomplete injury record. Who pulled a hamstring and when, who spent how many weeks off the field, who played injured — no full list exists. So a data analyst looking at the numbers and concluding that a certain batter declined after 2026 is in fact making an editorial judgment, not a cricket one. The distinction matters.

Looking across Bangladesh's Test history from 2026 to 2026, a pattern emerges that I have checked by hand. I have counted twenty-two matches by hand; the spreadsheet remembers what the injury erased. But I also know that the more empty cells that spreadsheet contains, the less confident I should be. I will not fill a single one myself.

The third stop: 2026. The world shut. The Bangladesh Premier League was suspended, the international calendar uncertain. In that void I built something that remains my most-used reference. A dataset of one thousand two hundred matches across twelve leagues from 2026 to 2026, of which four hundred and twelve were played behind closed doors.

The result: home win rate fell from 44.8 per cent to 37.6. Home penalty awards dropped nineteen per cent. Both numbers are clean, the direction is clean, and this is the one case where a variable changed and the record caught it. Covid was a brutal natural experiment, and cricket was fortunate enough to preserve the sample.

But here I hold myself to a strictness I never relax. That the gap from 44.8 to 37.6 was caused by the absence of crowds, I cannot prove. I can show only that the gap exists and coincides in time. Because at least three other things changed simultaneously: the schedule compressed abnormally, the five-substitution rule arrived, and travel restrictions disrupted preparation. None of those can be separated out.

That is where my real unease lives. When data exists, people assume the questions are finished. The opposite is true. Data starts the questions, and discipline decides where to stop. Where data is absent, the questions never start — only opinions do. And cricket media lives in the market for opinions, not questions.

The expectation gap is what shows most clearly. At the 2026 T20 World Cup, Afghanistan reached their first ICC semi-final, beating Australia by twenty-one runs in the group stage, then collapsing for 56 against South Africa in the semi-final. In the same tournament Bangladesh reached the Super Eight by beating Sri Lanka, the Netherlands and Nepal. At the 2026 ODI World Cup, the Netherlands beat South Africa and Bangladesh — two full members.

Nobody predicted these results, and it is now obvious why. Performance data for associate sides is so shallow that no model can be built. So the market misprices, and we call the outcome an upset. But "upset" is a confession — a confession that our expectation was blind. What we call an upset is usually a match whose data we never collected.

Here a counter-argument is needed, because the easy conclusion is "more data, better outcomes." I do not believe it.

First, more data narrows cricket's memory rather than widening it. The format that generates data is the format that gets attention. Every ball of a T20 league is recorded, so T20 leagues get analysed weekly. Domestic first-class cricket of the 1990s barely exists in data, so it has become history rather than subject matter. We believe we know more. We know the same amount, distributed differently.

Second, an empty cell does not grant permission to fill it — it creates the temptation. And the temptation is manufactured by editorial pressure. The deadline is tonight, collecting a player's injury history takes two days, so the easy path is explanation. "He is out of form." Three words concealing five unknowns: injury, role change, batting position, opposition planning, luck. Not one is known, and the sentence is already complete.

Third, correlation is not causation. I cite the 2026 study repeatedly for that reason — not for the result, but to show that a gap exists between result and explanation, and that gap must be closed in the language of uncertainty, not confidence. I do not trust any analysis that does not print a confidence level beside every number. The Croatia piece I wrote in 2026 proved correct. But the Croatia piece was right; the market simply wasn't ready. The distinction matters to me, because being vindicated and being understood are not the same thing. What I learned from that piece is that a prediction coming true is a test of the method, not a personal victory.

I don't trust a narrative until I have counted it by hand. For me that sentence is a professional rule, not a moral position.

So what should be done with the empty cell?

The first task is procedural. I now install a null gate on every project. No analysis proceeds without at least three verifiable information points and a specific format label. Without a format, analysis cannot even begin — Test economy and T20 economy are not measured on the same scale, and where the format is unstated, comparison is itself an error. This rule has halved my output. It has taken my correction count to zero.

The second task is journalistic: publish the empty cells. If I lack match-by-match data for a player in a given year, I say so in the piece rather than hiding it. The reader deserves to know that part of the analysis is weak, because the reader will quote it tomorrow.

The third task is long-term and hardest: investment. Everyone can identify where data is missing; almost nobody builds it, because the market there is small. Associate cricket, women's domestic cricket, Under-19 tournaments — whoever first builds ball-by-ball records in those spaces will set the analytical vocabulary for the next decade. That is a cricket-operations decision, not a journalism one.

The last question is for myself. The file that arrived empty at my desk was, in fact, a test. And I know most people fail it. Because an empty cell does not require you to write anything — only to drop in a placeholder and move on. And placeholders slowly become statistics, then beliefs, then history.

I want to know how many of cricket's established truths are standing on an empty cell. Answering that needs no new model — only the nerve to count the blank cells in every table, once.