The 71 Percent Puzzle: Khulna's Unwritten Scorecards and the Real Measure of Bangladesh's Spin Dominance

In three four-day matches at Khulna's Sheikh Abu Naser Stadium last season, I hand-coded every ball. Spinners took 71 percent of the wickets that fell. Same season, same competition, near-identical grassless surfaces: in Rajshahi the number was 48 percent. Twenty-three points of gap. The first explanations that arrive, the pitch and the humidity, do not survive contact with the curators' reports, which are almost word for word the same.

The gap is not in the pitch. It is in the filing cabinet.

In 2026 my first job at a Dhaka digital startup was coding all 44 matches of the 2026-17 Bangladesh Premier League football season by hand, 14,200 events. The data said Abahani Limited Dhaka had scored 23 goals from 15.8 xG across their first twelve games. The piece never ran; my editor's reasoning was that tactics talk is for the boys. Three weeks later Abahani scored nine goals in their next eight matches and dropped eleven points, and the same piece ran under a staff byline. The numbers were not lying; they were waiting for a better question.

The next lesson arrived from the opposite direction. Before the 2026 World Cup I spent twelve days coding 1,240 goals and published one claim: 43 percent of knockout-stage goals would come from dead balls. The tournament delivered 73 set-piece goals from 169 — 43.2 percent. A claim and a guess are not the same thing, provided you write down in advance what result would prove you wrong.

Those two experiences set my method. In domestic cricket I now walk the reverse route: question first, data second, verdict last. The part of the National Cricket League that never reaches television, that has no footage, whose scorecard never appears online — that is my primary archive. In Khulna I learned that silence is also a dataset. And the story of Bangladesh's spin dominance, recited now almost as scripture, is at least partly a product of that silence.

Let me put the question plainly. That spinners can bowl in this part of the world is not a theory, it is an observation. I want something else: how large the spin wicket-share actually is in domestic first-class cricket, and how much of it we are capable of measuring. Starting the measurement, the first obstacle is that the matches are not recorded evenly. Some carry full ball-by-ball logs, some only a scorecard, and some only a short newspaper paragraph — who took how many wickets, no overs, no maidens, no session-level breakdown.

The first pattern emerges right there. Sorting matches into three tiers — full log, scorecard only, news report only — the best-documented matches carry the lowest spin share, and the worst-documented matches the highest. Khulna 71, Rajshahi 48: not a venue difference but a documentation-density difference. Where every ball is logged, you can separate a pace change, the direction of the breeze, the state of the ball. Where only a results row survives, everything collapses into one vessel and produces a comfortable story — the pitch was helpful to spin.

My control group was the Dhaka Premier League, where scorecards are complete and ball-by-ball coverage far denser. In that 50-over competition the spin share sits between 38 and 42 percent, and barely moves when the venue changes. Same country, same bowlers, same heat — only the record-keeping is better. The distinction then becomes clear: the rise in spin share in my dataset is the sum of two separate things — a real pitch effect, and an inequality in record-keeping.

The second element is workload. In the subset I coded, spinners who bowled more than 40 overs in a domestic first-class innings averaged under 21 years of age. At that age the load tolerance of shoulder, elbow and lumbar joints is not yet built. On the international calendar a spinner's burden is measured in overs per match; in domestic cricket it is not measured at all, because no central workload register exists. So one bowler sends down three hundred overs in a season, appears on the injury list the next, and nobody files that as a failure of injury management — they file it as luck. That bowlers like Taijul Islam or Nayeem Hasan were forged in the long sessions of the domestic league is true; how many walked the same road and stopped halfway is a number nobody keeps.

The third layer is age verification. My position is unambiguous, and it is not a personal suspicion but a systems question. If the age of a young bowler rests on a school certificate and a birth registration, and the competition's age bands rest on that information, then a single bad record can bend the peak curve of an entire generation. My dataset shows that many spinners bowling regularly at twenty in the league have short active careers — and that is not a shortage of talent, it is an accounting problem.

The fourth layer is the least discussed. The standard peak-age curve, now used in almost every analysis, was built on English, Australian and South African conditions — five or six months of season, a winter for the body to recover. Bangladesh's domestic first-class season runs in a compressed four-to-five-month window, in heavy humidity, and an away match means eight to ten hours on a bus. Under those conditions the peak arrives later and flattens wider. What we call late maturity is really a body being compared against an imported curve. A system's output becomes credible only when the shortfall in its information is measured too.

There is a trap here, and it is my own. When the Khulna numbers showed a gap that wide, the most comfortable verdict would have been that spin dominance is entirely a manufactured story. That would be wrong.

A real pitch effect cannot be argued away. On a dry, grassless surface the ball will turn slowly — that is physics, not opinion. In the third and fourth-day sessions of a domestic match, when the colour of the wicket changes, a spinner's economy improves inevitably. The real question is not whether the effect exists; it is how large the effect is, and how much of it we capture.

My estimate runs like this: roughly two-thirds of the wicket fall is pitch and conditions, and roughly one-third is documentation variance — the matches we do not write up in detail are the ones whose explanation takes the easy road. That is not an exact split, it is a range, and the width of the range is my confidence limit. Let me be explicit about what my method cannot see: how much pace a bowler was losing within an over, what a spinner changed to buy impact, or how a side's calculation shifted when a session was washed out — none of that is captured. I use silent data; passing silent data off as proof is fraud.

One more thing has to be added. Most of the spinners coming through the domestic league are coming out of that same undervalued archive in Khulna, Rajshahi and Bogra. Even if someone calls all of this merely a filing story, one fact survives: many of those bowlers cannot be seen in advance because a large part of their work was never written down for us. Absence is also information.

Three things will hold my attention next round. First, the over-count of spinners under 23 in four-day matches — if a bowler crosses forty overs across two consecutive innings, that is a pre-injury signal for me, whatever the result. Second, who is keeping the log at those matches — I weight the pattern of documentation shortfall above the ball-by-ball data itself, because where there is no log there is no analysis, only narrative. Third, whether that twenty-three-point gap between Khulna and Rajshahi returns next season. If it does, it is not a truth of the pitch but an entry in a ledger; and ledgers can be rewritten without touching the square.

I will leave the last question open. If a large part of the spin-dominance story we love so much belongs to the matches whose scorecards nobody wrote — then whose explanation is the success actually?