In the summer of 2026 I hand-coded three thousand three hundred and sixty deliveries — and nobody replied.

In a rented room in Jayanagar, four kilometres from the M. Chinnaswamy Stadium, I had a laptop, a spreadsheet, and the recordings of fourteen Royal Challengers Bangalore matches from IPL 2026. That year the free stats sites offered no phase splits. They gave you totals, averages, strike rates — but not which passage of play those runs went to, off which length, to which side of the ground. So I started coding it myself: every legal delivery, its over, bowler, length, shot type. It ran night after night. By the end of the season the sheet was telling me something the scoreboard never had.

Bangalore finished last that year — three wins. That was the scoreboard's story, and it was not the whole story. The sheet said: death-over economy of 12.1 runs per over, the worst in the league. And 61 per cent of those runs came on the leg side, off full-length balls. Three wins was not caused by a 'bad team.' It was caused by something very specific and very fixable — the length repeatedly landing full and leg-side in the death overs.

I turned that sheet into a nine-page PDF and sent it to the franchise's general enquiries address. Nobody replied.

Eight years later I understood that the missing reply was not someone's negligence. It was a system design. In Indian cricket, what is not recorded is not considered to exist. And what is not considered to exist is never acted upon.

---

Context: the recorded surface of cricket, and what lies beneath it

Cricket is the most measured sport in the world. A Test match contains more than four hundred deliveries, and each one carries at least eight data points — runs, batter, bowler, mode of dismissal, field position. And still the game cannot answer its own most important questions, because those questions sit below the surface that nobody records.

For a long time I assumed this was a technology gap. I now think it is a priority gap. Cricket's data runs on two economies. The first is the broadcast economy — totals, hundreds, five-fors, milestones. That economy is visible, sellable, and gets put into graphics. The second is the ledger economy — phase-wise strike rate, economy under required rate, wickets as a function of match situation, opportunity cost. That economy is invisible, unsellable, and therefore unrecorded.

I let the ledger speak before I trusted the scoreboard — and I still keep that order. This is not stubbornness, it is method. The scoreboard is a marketing artifact; it tells you how a match ended. Phase-level data tells you how a match was made. These are different questions, and in Indian cricket coverage the first always buries the second.

The clean example is the hundred that cost the chase. Or the five-for that came in a lost cause. Those go into the milestone column and never into the match-situation column. That is the unrecorded data — and it does not go missing by accident. It is kept missing.

---

Core analysis: four ledgers, and one empty column

Ledger one — 2026, Chinnaswamy. Three thousand three hundred and sixty balls later, the pattern stopped being noise. A death economy of 12.1, and 61 per cent of runs off full-length balls to leg. One clarification matters here: that figure does not say 'the bowlers were bad.' It says there was a structural skew in the length distribution, and that skew was most expensive in the death overs. The difference is large — the first is an opinion, the second is an actionable instruction.

What I learned that season was not about technique. It was that the most expensive thing in cricket is not poor performance; it is misdiagnosing poor performance. If you believe the problem is 'quality,' you change players. If you know the problem is length distribution, you change practice sessions. The first route is expensive and often wrong. The second is cheap and often right.

My sheet pointed at the second route. But the second route needed to reach someone willing to abandon the first. That person was not on the other side of the email.

Ledger two — 2026, an empty stadium. September to November, sixty matches, Dubai, Abu Dhabi, Sharjah — no spectators. I was twenty-one, on a remote coding internship with a data vendor, coding live. The empty stadium taught me that absence is also a variable. No crowd changed bowling behaviour — crowd noise, field reading, batter pressure, even the appeal in an umpire's head.

But the real finding came from the pitches. In Dubai, for any side losing a powerplay wicket, the scoring rate between overs seven and fifteen fell by 1.4 runs per over. Sharjah's short boundaries inflated totals by roughly 19 per cent. One tournament, one rulebook, three different variables.

The crowd is a dataset, but never the whole dataset. And when the crowd disappears entirely, what is left is the venue. I wrote it up in forty pages. The outlet that commissioned it ran nine hundred words and cut the methodology completely.

That was the most useful blow of my career. I understood that media wants the conclusion, not the method. And that is where Indian cricket's real data crisis begins: if the method never gets printed, nobody can reproduce it. And what cannot be reproduced is not science — it is an opinion wearing a spreadsheet's mask.

Ledger three — February 2026, the ninety-crore sheet. The IPL mega auction in Bangalore: ten franchises, a purse of ninety crore rupees each, six hundred names in the pool. I was twenty-two, finishing my degree. I built a rupee-per-win-share model off three seasons of phase data — how many wins a player's runs or wickets actually convert into, and what that conversion is worth. The sheet valued an uncapped nineteen-year-old left-hander at four and a half crore rupees. Mumbai Indians bought Tilak Varma for 1.7 crore, and he made 397 runs that season.

A spreadsheet can hold ninety crore rupees and still miss the point. My sheet was pointed the right way. The sheet stayed with the franchise; the lesson stayed with me. The lesson was that valuation was never the bottleneck. Access was. A wrong number and a right number walk through the same door, and there is no bouncer at that door.

Ledger four — the column that was never filled. Last week I watched an automated analysis pipeline run a score-based deconstruction. Its output came back — the title fields empty, the list of factual points empty, the entities empty, the source undetermined. And yet the full report skeleton was present, every heading in place, every table prepared.

I sat in front of that screen for a long time. It was frightening not because the system failed, but because after failing, it still looked complete. No fake name was invented, no fake number inserted — and that is precisely what makes it most dangerous. An empty column that reads like a clean sheet.

This is the purest form of cricket's data problem. Our problem is not that our data is wrong. Our problem is that we never find out that the data was never there.

---

The measurement gap: what India bothers to record

I have spent ten years auditing who gets counted and who does not — and the pattern is identical at every level.

Take women's cricket. The WPL launched in 2026, and in 2026 India won the World Cup at home. Both are large historical facts. But the distance between winning a tournament and measuring a tournament is unchanged. Phase-level data, length maps, fielding-position data for women's matches remain thinner than the men's game. If you do not have three years of ball-by-ball data, then in three years you cannot extract a single trend — not in selection, not in strategy.

Take the Ranji Trophy. Four months of season, more than eighteen teams, hundreds of innings — and coverage is nearly invisible before the knockouts. That invisibility is not the result of fan interest; it is not even the cause. It is a circular arrangement: what is not recorded gets no coverage; what gets no coverage is not recorded.

Take the second team — the team with no name. In any sport in the world, the hardest team to make should be India's second XI. You have to score in Ranji, play India A, take wickets, then sit on a bench and start again. And still the data from those second XI matches is about as scattered as Chinnaswamy's death-over data was in 2026.

India's problem is rarely talent. The problem is what the system bothers to record. A million deliveries are bowled every year, most of which nobody codes. What is not coded does not enter the statistics. What does not enter the statistics does not enter the selection meeting. What does not enter the selection meeting does not show up. It is a clean pipeline, and it comes out empty at the bottom.

One thing must be said honestly here. An empty column and a zero figure are two different things, and merging them is the easiest way for this entire argument to fail. In some cases data was recorded and the result was null — that is a finding. In some cases data was never collected — that is a limitation, not a finding. I do not put the two in the same column. Keeping measured and negative separate from unmeasured is the whole of the analysis.

---

Contrarian angle: 'no problem found' does not mean 'no problem'

This is where I turn my own argument, because it needs turning.

Last week's empty report told me something loudly. The full table was present — every row, every heading. And every cell read: insufficient information. That was correct behaviour. But if someone skimmed it, they would read 'no risk.' A table being empty and a risk being empty are worlds apart.

I have made this mistake myself, and paid for it. Early on I began reading holes in data as signal. A player not selected looked like a conspiracy. A match with no coverage looked like something being suppressed. It is an attractive pattern, and it is almost always wrong. Sometimes data is missing because someone suppressed it; far more often data is missing because nobody bothered to pick it up. The second is more common and less thrilling.

The second trap sits inside my own signature. Counter-intuitive discovery gets rewarded — the piece travels, the argument builds. And in that reward loop an auditor slowly becomes a debunker who needs one improbable finding per column. That is my largest trap.

I know only one cure: publish the null results. If I coded three thousand balls and found that phase splits showed nothing meaningful, that should be printed too. A ledger that only ever reports surprises is not a ledger — it is a column with a theme.

The third trap is sample size. If a pattern rests on one twelve-ball spell, it is an anecdote with a spreadsheet attached. So I set the threshold before I run the query, and I publish the threshold. I do not chase outliers; I interview them until they confess. If they do not confess, I stop calling them outliers.

And the fourth, the one most uncomfortable to write about: the broadcast box. In Indian cricket, the last resistance to the data argument is the commentary panel. Getting there requires softening your sharpest lines. I have decided that access and argument are separate things. Methodology and conclusions get published first; the conversation about the chair comes later. If the chair comes on condition that the argument changes, that chair belongs to someone else.

---

Takeaway: the next signal is not in a number, it is in a column

The next time you read a cricket report, do not read it for the data. Read it to see what it is not telling you.

I am not saying Indian cricket lacks data. There is plenty. I am saying that in Indian cricket a few things are over-measured and a great many things are measured at zero, and both are painted the same colour. How thoroughly a team's top order is measured, and how thinly its second XI is — that is the real inequality.

Over the next two years, the franchise or board that understands its advantage lies not in buying players but in filling columns will move ahead. Because players are visible to everyone. Length distribution, phase breakdown, venue-adjusted run value are not. And what is not visible attracts no bids.

The question is not whether India has the data. The question is: the next time a seventeen-year-old sits four kilometres away and codes three thousand balls, who reads his email — and should replying to it be the system's duty, or his luck?

Every clean column began as a messy argument with reality. We have stopped arguing, and the columns look beautiful.

---

Note: the phase data, auction model and venue analysis referenced in this article come from the author's own hand-coded datasets — the 2026 IPL ball-by-ball ledger, the 2026 UAE tournament dataset, and the February 2026 auction valuation sheet. Method and sample limits are both open; no conclusion is quoted here without its method.