The Empty Column: Cricket's Thinnest Data Sits Where the Game Is Played Most

Last week, at my desk in Manchester, I opened a spreadsheet and tried to put two domestic first-class seasons into a single frame — a full Indian Premier League season on one side, a Ranji Trophy season on the other. Three days later I had to close the sheet. The question was not wrong; half the data needed to answer it had never been recorded. The IPL side has every ball, every phase, every matchup. The Ranji side has scorecards, ball-by-ball commentary for a handful of matches, and then a large empty column. Eleven thousand rows later, the clearest thing was this: cricket's data is thinnest exactly where the game is played most.

This is not a technical complaint. It is the oldest evidence problem in a sport we measure obsessively and understand only partly.

The County Championship has run since 1890 — eighteen first-class counties, two divisions, a calendar stretching from April to September. The Ranji Trophy began in 2026-35 and today runs roughly thirty-eight teams. Both systems manufacture the raw material from which Test teams are built: Dukes-ball seam movement, fourth-innings pitch deterioration, the wicket ratio in the first six overs with the new ball, the swing spike on an overcast April morning. James Anderson's entire career belongs to that document — debut in 2026, farewell in 2026 with 188 Tests and 704 wickets — yet the largest share of it sits in Old Trafford scorecards that no analyst opens regularly.

Cheteshwar Pujara's 2026 county season for Sussex, in English conditions against the red ball, is evidence of something Indian domestic reporting never records: how technique changes in foreign conditions. The season is in the scorecard. The phase-by-phase record of that change is not publicly available.

This is where the real gap sits — between scorecard data and event data. A scorecard lets you rank: who scored how many, who took how many. Event data lets you explain: on what line, at what length, through which shot, against which field. Ranking is not enough for selection. "52 wickets this season" is a number; "6 wickets in 14 overs in the fourth innings, three seam-up, two stump-to-stump" is an argument. Indian and English selection panels have decided on the first number for years, because the architecture for the second was never built.

In the IPL, that architecture got built. Franchises stood up data desks and turned the auction into a pricing problem — sorting age, role, matchup and cheap desperation onto one sheet. Zonal selectors once did that work; scouting desks do it now. The IPL's abundance proved one thing: where the record exists, cricket's investigative logic wins and the commentary box falls behind.

But the same intellectual architecture was never built for the format that actually picks Test cricketers. A Ranji season runs four months, contains more than a hundred matches, and ball-by-ball coverage of those matches shifts from season to season. That inconsistency is the real problem. The data exists, but its shape changes every year, so a five-year comparison cannot stand. A record that is not the same year to year is not a record; it is an accident.

In 2026 I tagged every ball of Lancashire's home fixtures from the North Stand at Old Trafford — two columns, seam movement and pressure deliveries. The finding that came out that season is still the template for my work: in the twenty overs after a wicket fell, that attack's economy rose by 1.4 runs an over. That is a thing only phase-wise data shows, and it is the thing selection needs most — because it says when the pressure breaks, not who took the most wickets.

I keep a column for what the data cannot see, and it is rarely empty. In this work it is crowded. How many overs a domestic seamer actually bowled across a season is on no public sheet. How the pitch was prepared, whose wrist was swollen, which match a team deliberately asked to be slow — all as unmeasured as field placements. So you fix the question first, then say plainly what the data cannot see. An honest "unmeasured" beats a confident misfire.

An INTJ does not trust a narrative until the rows survive a second sort. So I reopened the unfinished sheet and kept only matches where ball-by-ball coverage existed on both sides. The surviving sample was small enough that any trend drawn from it would be self-deception. This is where most data stories die — and, instead of dying, get published.

There is a reverse risk I want to raise against my own argument. More data does not automatically produce better decisions. The IPL's abundance carries a trap: a matchup resting on forty balls starts to look decisive when it is noise. Apply the same rigour to a thin Ranji dataset and the result is confident nonsense. So I have fixed the sample threshold in advance: if an effect rests on fewer than a few hundred deliveries, it is a question, not a conclusion.

But the alternative to thin data is not better data — it is anecdote, and anecdote has been picking domestic cricketers for decades. Here the eye test deserves one genuine concession: no public sheet records a seamer's workload or the slow decline of a body. A coach who sat in the stand all season can see how a quick's pace dropped in February. That is the eye's legitimate job and I would not replace it with data. The problem starts when that same eye looks at a wicket count and makes the entire selection call.

There is a third layer, and it is less comfortable. Where the measurement framework breaks, the market fills it — but only where there is money. T20 league data is dense because broadcast revenue, fantasy and betting attach to that ball. Nobody's market cares about a Ranji Tuesday morning. The result is that the format which earns money gets measured better automatically, while the format that makes players runs on old guesswork. This is not a technology gap; it is an incentive gap.

It touches the Indian story directly. 19 November 2026, Ahmedabad — the World Cup final, Australia winning by six wickets. The debate that followed was mostly conducted in the language of emotion. For me it was a systems question: we had the framework to measure that day's conditions, but not the framework, built over the preceding four years, to select players who could hold up in them. The scorecard told us what happened. It never told us why.

What to watch next. Through the coming Ranji season I will track one thing: how many matches have ball-by-ball coverage available for a second consecutive year. If that number halves from one season to the next, then any five-year trend claim is not a trend but selection bias. The question is not how much data we have; it is whether it will have the same shape next year. The day a domestic seamer's fourth-innings economy gets sorted the way an IPL death-over specialist's is, we will not need a new hero — just a sheet that looks the same two years later.