The Honesty of an Empty Cell: Cricket's Unlogged Half

The Empty Cell

Eleven in the morning, Delhi. I opened the laptop and found a row in the spreadsheet. The row existed — it just had nothing in it. No match name. No player. No format. No venue. Every cell carried the same three characters: N/A.

My first instinct is the one the ESFP half of me always wants — fill the gap at any cost. Grab the nearest plausible match, drop in the nearest plausible name, start the story. Nobody on a panel wants to look at an empty cell. They want a number, even if that number belongs to a different match, a different season.

But I opened the notebook before I opened the spreadsheet. And the notebook had exactly one line written in it that day: what I do not know, I do not know.

This piece is about that line. In cricket, N/A is not a failure. N/A is a finding. And very often that finding tells you more about the sport than any number we filled in to avoid the discomfort.

A scorecard is a receipt; a log is a ledger

2026, Feroz Shah Kotla. Sixteen years old, a lumbar stress fracture, and a scorer's assistant chair. Delhi versus Karnataka, Ranji Trophy. Four days, 1,412 legal deliveries, 38 maidens, 12 edges to slip. Every ball's line, length, bowler type and release point, measured off broadcast stills.

Karnataka's Mayank Agarwal made 176 in that innings. The scorecard said: a big hundred, a match-defining knock. My notebook said: Delhi's seamers beat his outside edge 19 times.

Both statements are true. They do not say the same thing.

A scorecard is a receipt. A ball-by-ball log is a ledger. The receipt confirms a transaction happened. The ledger tells you how it happened, in what order, and where the road turned. When a match has no ledger, we pronounce judgement on it using only the receipt — and then we call that judgement analysis.

That is my whole complaint. A highlight is not a fact. A highlight is a claim. And claims get audited.

Cricket's three layers

When I built my first spreadsheet, I thought I was assembling all of world cricket in one place. A few years later I understood I was only assembling the cricket that had data attached to it. And cricket's data sits in three layers.

The first layer is where every ball is logged. The IPL, major internationals, World Cups. Here you get ball-by-ball, phase splits, matchup grids, dot-ball pressure, death-over economy — everything. Analysis is cheap here because the data was never expensive to collect.

The second layer is where only the scorecard is logged. Large parts of the Ranji Trophy, domestic one-dayers, under-19 tournaments, a good chunk of women's domestic cricket. You get runs, wickets, overs. You do not get which ball triggered a field change, which over the ball got old, against which batter the yorker plan was abandoned.

The third layer is where only a line gets logged. A paragraph on an inside page, a scoreboard, a local reporter's note. Associate cricket, district cricket, club cricket, and a great deal of women's cricket live here.

The ball is the same in all three layers. The length is the same. The human being is the same. What changes is the camera, the commentary box, and a scorer's salary.

When you see N/A in a database, it is mostly telling you about money, not about the sport. That is cricket's biggest hidden story, and it matters more than any match report.

14,400 balls and an empty stadium

In 2026 I built a full log of the IPL played behind closed doors in the UAE. Sixty matches, 14,400 deliveries across both innings, and more than 12,000 logged events. Plus one more column — broadcast crowd-audio level, over by over.

I did not add that last column out of curiosity. I added it because I needed to know whether the mathematics of a match changes when nobody is in the ground.

What came out: chasing teams won 57.5% of matches that season, against 51.2% in 2026. A six-point swing. Here the context-adjustment hawk in me wakes up, and I hold it back, because I have pre-registered the rule — a variable I do not fix in advance becomes an excuse later.

Team strength, toss, pitch, dew — none of them alone explains that six-point gap. But it is also true that six points does not equal a cause. Cricket inherited the habit of turning correlation into causation from broadcast graphics.

The empty stadium still had a heartbeat, and I logged it. I just refused to call it a cause. I called it a signal that needs three seasons of data, not one.

That is the difference between me and the table someone sends me on Twitter. There, every correlation becomes a cause instantly, because causes sell easily and questions take time.

Tokyo at 2 a.m.

2026, the Tokyo Olympics. Two in the morning in Delhi. I logged India's 117 athletes and seven medal events into a 1,204-row dashboard. Neeraj Chopra's javelin went 87.58 metres and became gold.

That night I learned something that transfers directly to cricket. On a remote screen you watch the event, not the atmosphere. What an athlete in the stadium heard, I did not hear. What I saw was decided by a camera operator.

At 2 a.m., Tokyo looked like a column I could not sum.

Apply that to cricket. When you read a Ranji scorecard, you are not watching that match. You are watching the portion of it somebody chose for you. And when you read only the result of an associate match, you have not read the match — you have read a notification about the match.

Bench strength: the thing with no data at all

Every season India's bench strength gets debated. Every debate contains the same sentence — "he has the domestic record."

I take that sentence apart. What does a domestic record mean? An average of 45 in the Ranji Trophy. On what surface? Against which bowling attack? In which match state? How many catches went down? How many times did he shield himself in the first ten overs against the new ball?

The data needed to answer those questions does not exist. And when data does not exist, the debate parks itself on the average. An average is a number that satisfies everybody and informs nobody.

This is not a talent story. It is a measurement story. We cannot pick the player we never properly measured.

Take the Women's Premier League. In the IPL a player carries ten seasons of phase-wise data — powerplay strike rate, rotation against spin in the middle overs, boundary percentage at the death. In the WPL that layer is still being built. That does not make WPL cricket lesser. It means our ledger for WPL cricket is thin, and judgements drawn from a thin ledger are thin judgements.

The same happened with the Impact Player rule. It arrived in 2026 with the stated purpose of changing one phase of the match. The question is whether that phase actually changed. Answering it requires over-by-over data from immediately before and after the substitution in every match. In the IPL that data exists. If a domestic tournament copies the rule, all you will have is a scorecard — and a scorecard will never tell you whether the phase moved.

More data, less truth

Now the part where I have to argue with myself.

The easiest slogan in cricket data is: "we need more data." I do not agree with it, at least not in the form it is sold.

An empty cell is honest. A filled cell containing a borrowed number is a lie with a decimal point on it. I have seen plenty of models where every player had a number for every situation, and the number had been lifted from another league entirely.

On the other side sits context-adjustment creep. Pitch, dew, match state, bowling matchups, crowd — add enough variables and any outcome can be explained afterwards. It looks like rigour and functions as immunity. A model that no result can falsify is a model that does no work.

My rule is strict. Put the sample size inside the sentence, and fix the variable count in advance. If a pattern is showing up across a three-over spell, do not call it a pattern — call it a question, and state the threshold at which it becomes a claim.

I have one more rule. Every piece needs one scene no column could have produced. Fourth day at Kotla, when somebody put a cup of tea behind the scorer's chair and it went over, and I missed an entire over trying to save my notebook — that over is blank in my log. I have never filled it with an estimate.

Sixty-four matches fit in one spreadsheet, but the room did not.

What to watch next

Next season, when you read a match report, do not go to the scorecard. Check whether that match has a ball-by-ball log. If it does, the report's claims can be audited. If it does not, the report is an opinion dressed as a statistic.

I followed the decimal point until it became a story. I also learned that some decimal points never become stories, because nobody ever measured them.

Next time someone tells you India's bench strength is deep, ask one question — measured in which ledger? And if the answer is a scorecard, then understand that we are still playing half the game in the dark.