Hook: One Over, Two Weeks, Three Ledgers

I opened the Khulna ledger, and the first column taught me patience.

Sheikh Abu Naser Stadium, Khulna. A first-class match in the 2026 National Cricket League. The final session of day two. Fewer than two hundred people in the stands, dust in the air, the hum of a generator. A laptop near my knee, a paper notebook beside it, because paper does not crash.

The online scorecard showed two runs off the last ball of the 32nd over. Then the first ball of the 34th. The 33rd over was missing. Six deliveries gone. The board's official follow-on sheet had the over. My handwritten ledger had the over: seven runs, one boundary, two dot balls, and in that over the left-arm spinner's four-over quota ran out. The digital record had nothing.

Two weeks later, reconciling the return-leg scorecards, it surfaced. Nobody had hidden anything. No conspiracy. One data entry had been dropped, and nobody caught it, because nobody was positioned to catch it.

The over that vanished was not six balls. It was the explanation of a tactical decision. Why the captain brought the seamer back, why the field dropped deep, why slip came out. Remove the over from the ledger and you remove the reasoning with it.

In Asian cricket this is the rule, not the exception. This article is the accounting of that rule.

Context: Who Keeps the Ledger, and Why Nobody Can Reconcile It

Asian cricket data lives in three layers, and the three layers do not speak to each other.

The first layer is board-level. India's Ranji Trophy, Pakistan's Quaid-e-Azam Trophy, Bangladesh's National Cricket League, Sri Lanka's Major Club Tournament, each with its own scoring software, its own entry team, its own archive. Ball-by-ball data from Bangladesh's national league is not reliably published in machine-readable form; where it exists, it arrives as an end-of-match PDF from which no column can be built. Across fifty-two years of covering this game I have sat in domestic grounds in four Asian countries and seen the same picture: paper sheets on the scorers' table, a laptop beside them, and someone either updating the score or not.

The second layer is international and governing-body controlled. Match referee reports, ball-tracking systems, Decision Review System archives. Ownership is centralised; publication is limited. DRS ball-tracking generates data on every delivery, yet researchers cannot see all of it. Broadcasters and league sponsors hold separate datasets.

The third layer is franchise and commercial. The Bangladesh Premier League, Indian Premier League, Lanka Premier League, Pakistan Super League. Here data is inventory. Whoever holds the tracking contract holds the data. When the tournament ends, the data goes into a locked room behind the website, and next season a new file starts over.

Between these three layers there is no shared record. One player's career is written three different ways in three places. One match's truth survives in three files, three ways.

When I launched an xG column for a Dhaka-based football site in 2026, my first act was to build a template that forced every claim to sit beside at least three metrics. Once the template circulated, junior writers worked in the same format. Asian cricket has no equivalent template anywhere, because before you can install a template you need a reliable, time-stamped, single version of the record.

Core Analysis: Three Regimes, Three Ledgers, Three Truths

The largest methodological error in cricket coverage is treating Test, ODI, T20I and franchise cricket as one regime. I take regime-splitting literally. Each format produces different physical demands, different bowling quotas, different samples.

Test regime: the largest sample, the weakest ledger.

A Test innings carries more than two hundred deliveries, a match four innings, a series four or five matches. Yet domestic Test data in Asia is the worst preserved. Bangladesh's National Cricket League runs eight teams per season, each playing at least six matches, roughly twenty-four first-class matches a season. Small gaps like the lost Khulna over accumulate: over a decade, several hundred overs go missing.

The consequence is specific. Spinner selection in Bangladesh is often justified by domestic first-class performance, but nobody keeps the pitch-level, phase-level breakdown of that performance. How fast does a Khulna or Rajshahi pitch break? How much does spin turn in the second session? How much seam movement returns on the third morning? With a continuous record, spinner selection would stop being guesswork.

When I made my T20I commentary debut in 2026, Bangladesh beat New Zealand 5-0 in the Mirpur series. Watching those spin figures, I thought the explanation sat in the Test ledger, and the Test ledger was incomplete. The skill that made a bowler effective in T20I should have been visible in domestic first-class records. It was not.

ODI regime: the middle overs, where data is richest and explanation poorest.

In fifty-over cricket the middle overs, 11 to 40, are the least discussed and most decision-dense phase. At the 2026 Asia Cup final at R. Premadasa Stadium in Colombo, India beat Sri Lanka by a huge margin and Mohammed Siraj took 6 for 21 in six overs. Coverage devoted itself to his swing. In my ledger the biggest number that day was not Siraj's. It was Sri Lanka's dot-ball rate in the first ten overs, which destroyed the tempo of their innings.

It took me two days to find that number, because Asian tournament scorecards do not publish over-by-over phase splits. I counted by hand. A hand count means a sample error risk.

Asian ODI cricket has a structural feature: spin-heavy middle overs. On subcontinental pitches, three or four spinners bowl between overs 15 and 35. How much the run rate drops, how many wickets fall, decides the match. Yet there is no standardised record of this phase. So those who say "Bangladesh is slow in the middle overs" cannot show a number. Neither can those who say "Bangladesh is sharp in the middle overs." Both sides argue from estimates, and estimates have no ledger.

T20 and franchise regime: fastest game, slowest accounting.

T20 carries the most decisions per ball, so the highest data density. It also carries the most concentrated ownership. Franchise tracking data sits inside commercial contracts. A researcher wanting a full season's powerplay dot-ball rate must key it in from broadcasts by hand.

In 2026 I built a live pressing model for France's World Cup run. France's passes per defensive action rose from 8.2 in the group stage to 14.6 in the final: they pressed less and waited more. That PPDA map was not a picture; it was a confession of where their pressure lived. Cricket's equivalent confession is powerplay bowling-quota usage: who bowls four overs in the first six, who bowls three, who bowls two. That ledger tells you where a side actually wants to build pressure.

Bangladesh's franchise cricket does not keep it. Player valuation therefore runs on total runs and wickets rather than context. The bowler doing the hardest work in the powerplay has it buried under easier middle-over spells in his economy rate.

This is where a standing objection of mine attaches. The soft language of workload management in Asian cricket is often a curtain over commercial calendar pressure. When a national series sits inside a franchise league, the player rested is rested because of the schedule, not because of a medical report. If it were genuinely load management, every delivery's speed data, spell length, and gap between consecutive matches would sit in an open ledger. It does not. What exists is the announcement.

The ledger question: what a shared record actually requires.

I am not arguing for blockchain as a technology; I am not a technology enthusiast. I am arguing for a method: a record that is time-stamped, signed, tamper-evident, and replicated in many hands rather than one.

To catch that lost Khulna over you needed only two things: a timestamp and a cross-check. Had the scorer written an entry at the end of the over, had the board's server signed it, had an independent node verified it, those six balls would never have gone missing.

A shared cricket ledger would require five layers:

Per-delivery timestamps with bowler, batter, runs, wicket probability, and field-placement codes.

Two signatures per over, scorer and referee, from two different hands.

A hash at the end of each match chained to the previous match's hash, making mid-ledger deletion impossible.

Pitch metadata: pitch age, session temperature, humidity, grass ratio, and the preparation timeline.

Format tags on every record, so no analyst blends regimes into a wrong conclusion.

With those five layers, what changes in Asian cricket is not the number of metrics but the quality of decisions.

Take spinner selection. Today it runs on total domestic wickets. With a shared ledger it would run on fourth-innings turn per over, right-hand/left-hand matchups, and third-session stamina.

Take T20 opening pairs. Today selection runs on strike rate. With a shared ledger you would see the quality of bowling faced in the powerplay: runs against the easiest bowlers, or against the best two with the new ball.

Take workload. Today rest is announced two days before a match. With a shared ledger you would see deliveries bowled in the last 28 days, spell counts, gaps without an over-long break, and the percentage drop in pace. Rest would then be a calculation, not an announcement.

One clarification is necessary, because praise of ledgers may suggest I think the problem is solved. It is not.

Contrarian Angle: A Permanent Ledger Makes Bad Data Immortal

My largest fear is the proposal's own failure mode.

A shared ledger makes data hard to alter. But verifiable is not the same as true. If a scorer writes an error at the ground and it receives a time-stamped signature, the error becomes permanent. A blockchain-style record does not delete mistakes; it preserves them.

I think of that Khulna over. I was certain it contained seven runs. But my handwriting could have been wrong. I was the only witness. One witness is not a ledger, it is a claim. Two independent witnesses agreeing is evidence. The entire value of a shared record lives in the second witness.

The second problem is the gap between evidence and interpretation. A ledger can say the 33rd over produced seven runs and the spinner's quota ended. It cannot say why the captain brought the seamer back. That decision lived inside a coach's head, and no hash captures it.

In 2026 I tracked Abahani Limited Dhaka across 14 matches. They scored 28 goals from 21.4 xG, an overperformance in finishing. I published a regression warning. Three of their next five matches were draws. Someone said my model had predicted it. That was a misreading. My model did not predict; it said the overperformance was not sustainable. The gap between correlation and causation does not close because you carry a ledger.

The third problem is commercial. Franchise owners are not interested in open ledgers, because open data reduces their bargaining power. If a player's true value is publicly visible, the auction's dark room loses its leverage. Boards are equally uninterested, because an open ledger exposes the weaknesses of domestic structures.

The fourth problem is not uniformity but reconciliation. Test and T20 ledgers must stay separate, but keeping them entirely apart breaks the bridge. Understanding how a player converts domestic first-class patience into T20 aggression requires crossover notes. Without them, two ledgers are two islands, and islands do not add up.

The largest caution: missing data and silence are not the same thing. Sometimes a record is absent because nobody kept it. Sometimes because somebody prevented it. Sometimes because nothing happened. Failing to separate these three turns analysis into a chase after its own shadow. In the Khulna case the event occurred and someone forgot to write it: the first kind. Many gaps in Asian franchise data are the second kind, and that is a different conversation.

When the stadium emptied, I audited the silence and found the game still breathing. But silence is not always proof of breathing. Sometimes it just means nobody is speaking.

Takeaway: What to Watch Next Season

I am not petitioning any board. I am marking a signal that can be tested over the next twelve months.

First signal: if an Asian board publishes domestic first-class ball-by-ball data as a dataset rather than a scorecard PDF.

Second signal: if a franchise league publishes a full season's powerplay bowling-quota usage at the end of the campaign.

Third signal: if workload rest announcements attach deliveries bowled and spell lengths over the previous 28 days, and publish that before the announcement rather than after.

If any one of those three holds, I will revise my accounting. A clean row of data will outlast a thousand hot takes, and inside that survival Asian cricket gets its lost overs back.