High Domestic Runs, Low Test Averages: The Gap That Emerges When You Read the NCL Ledger

Hook: One Name, Two Ledgers

On the last week of February, in the scorers' box at the Sheikh Abu Naser Stadium in Khulna, I was staring at two numbers. One was 914 — a Bangladeshi batter's aggregate for that National Cricket League season. The other was 23.8 — the same batter's Test average. Same man, same hand, same eyes. But the two ledgers refuse to speak to one another.

I have read cricket as a ledger for years. Every ball is a block. Every innings is a chain. A series is a record that, once closed, you cannot alter — you can only interpret. When play ends, the scorecard's arithmetic does not change; what changes is our reading, our definitions, our venue notes. So I never release a claim without a hash: what the sample is, what the source is, what the error margin is. Without those three written down, it is not analysis, it is just commentary.

The question that circled me at Khulna that afternoon was old: how much of 914 is signal, and how much is venue, pitch, attack pace, and the debt to the calendar?

Context: How the Ledger Was Built

1. I was thirty-five, working as a club licensing assistant in Khulna, coming home at night to hand-code all 132 matches of the Bangladesh Premier League — every shot, every xG value, every defensive action — into one spreadsheet. Nine months, unpaid, on the conviction that the spreadsheet catches what the eye misses. I built the 132-match spreadsheet to find what my eyes kept missing. It was read 40,000 times: champions Abahani Limited Dhaka converted at 0.19 xG per shot above the league mean, while Sheikh Russell KC generated more chances but shot from an average of 19.4 metres.

After that thread, my writing changed. I dropped match reports and began writing "how we know" pieces — slower, but with readers who stopped arguing with my numbers and started quoting them.

Earlier still, in 2026, I began on the sports desk of The Daily Star as a cricket reporter. That desk taught one thing: you can write under deadline, but you can never edit the scorecard's arithmetic.

In 2026, aged thirty-six, three weeks before the Russia World Cup, I ran a PPDA regression across the 32 qualified teams. It flagged Germany as the tournament's most fragile seed — their pressing intensity had drifted from 8.1 in 2026 to 13.6, meaning fewer pressures and more progressive passes conceded per 90. The PPDA regression named Germany before the broadcasters had a clue. Germany exited in the group stage. In interviews I refused the word "prediction," calling it "a description of a trend with a stated error bar." Since then I add a standing paragraph to every preview: what would prove me wrong.

2026, aged thirty-eight. When the Bundesliga returned without crowds, I logged all 83 remaining fixtures. Home advantage collapsed — home goal difference fell from +0.42 to +0.09 per match, and yellow cards issued to away teams dropped roughly 24 percent. Eighty-three closed-door matches shook my confidence in every crowd-driven metric. I published the raw dataset openly but refused conclusions until I had a full control season, a delay that cost me three weeks of coverage. My sentences changed shape: not "the data shows" but "the data shows, given these conditions."

Now I applied that same discipline to domestic cricket. The last three seasons of the National Cricket League, first-class, four-day format, ball-by-ball coding. Six venue clusters separated, because an average never names its venue. Variables capped at four, and the final sixteen matches held out in a validation slice so the model cannot praise itself.

Why the NCL? Because it is the fixture nobody watches. No cameras, no crowd, no social media noise — which is exactly where crowd-driven metrics cannot even be born. Sample-size archaeology is the job: recover signal from those matches, then use it to stress-test the crowd-driven metrics everyone else trusts.

Core: Where the Gap Is Manufactured

In my ledger, after venue adjustment, the finding is blunt: the distance between domestic first-class batting averages and Test batting averages is not player-specific, it is artefact-specific. The gap is manufactured by attack pace distribution, pitch flatness, ball brand, and the fatigue arithmetic of the calendar.

Start with the fastest variable. In the NCL, the share of deliveries above 135 kph sits below single digits. In Tests against Bangladesh, that share in the new-ball overs is several times higher. The consequence is that a domestic top-order batter's front-foot discipline is never tested under real stress. He scores 900 runs in an environment where his reflexes never touch 140. The average rises, but the skill that average represents is never the skill actually used at international level.

A domestic average does not predict international conversion unless you treat attack pace distribution as a separate variable. That is my central finding.

Second: spin. The NCL is genuinely spin-dominant — but of a specific pace band and trajectory, largely left-arm orthodox. A large share of the domestic elite bowl at the same pace with the same release. A batter grooves to that rhythm for three years and believes he has learned to play spin. In Tests he meets wrist spin, 95 kph off spin, and leg spin — against whom his domestic sample is close to zero.

Third: innings par. The gap between a flat deck at Fatullah or BKSP and a turning surface at Sylhet or Chattogram is large. Without innings-par adjustment, an average is really a venue's name, not a batter's identity. In my ledger, where the flat-deck sample is heaviest, the conversion rate's collapse is steepest.

Fourth — the factor everyone skips — the calendar. Four-day format, compressed winter schedule, frequent travel. A domestic bowler is rarely asked for more than 35 overs in one spell-cycle. So a 50-wicket domestic haul never tests the durability of his fifth spell. Test cricket's last session of the second innings demands exactly that.

The reverse also sits in the ledger. A domestic seamer's 40 wickets spread across three seasons and two pitch clusters is a far stronger signal than a 50-wicket single-season storm. A bigger number is not bigger proof — the spread of the sample is the real weight.

Now from the ledger to the market. The BPL auction prices an incomplete dataset. Franchises routinely price a domestic player off a one- or two-innings sample on domestic indices. In the transfer market, I learned to wait for the third source. One number comes from the player's camp, one from the story of a club's interest, and the truth comes from the quiet row of timestamps and fee columns. A deadline-day deal is a story told in timestamps and fee columns, which nobody reads in full.

And here an invisible cost surfaces: the noise manufactured by managers and intermediaries. That noise scrambles the valuation process so that what moves the column is not the player's actual index but who can shout loudest. I keep a ledger of every rumour that died without a receipt; after each auction I reconcile which price came from on-field data and which came only from volume.

A structural Bangladesh problem is the "domestic tax" — picking a local player for budget, quota, or marketing reasons, a decision that is expensive on the field. When I matched the quota arithmetic, sides that filled the domestic slot on genuine competition had both better strike rates and better economy. Quota arithmetic is one thing; decision arithmetic is another.

Then the youth funnel. In my ledger the path from an Under-19 World Cup squad to the national side divides into five gates: age-group side, NCL, A-team series, BPL, national team. Attrition rises at every gate — and there is no safety net for those who fall.

Here sits my second standing position. Scout networks discover genius, but they also create a lottery economy: a family invests in a child, uses one brilliant domestic innings as collateral, and the expectation inflates unevenly. One succeeds as a talent; ten return carrying receipt-less debt. In my writing I treat the family investment structure with the same cold eye as agent selection — because the numbers are cold.

There is a tactical parallel too. In football, the revival of a back three is not progress; it is the safe route around the reputational risk of a back four. Cricket's analogue is the safe six-bowler structure whose purpose is not to win the match but to share out the blame for losing it. The side does not lose, but it cannot take the wicket it needs in the last five overs either. In my ledger, teams using this safety structure most often show consistently weak tie-breaking indices.

Finally, depreciation. How long does a domestic average predict? Its shelf life depends on two things: ball brand/seam and the pitch-preparation calendar. Change either and the weight of the old sample decays. The conversion rate you trust today may be a dead metric next season — and that can be declared in advance if you read the calendar.

Contrarian: Correlation Is Not Causation

Now the turn against myself. My ledger also holds names whose domestic numbers understated their international capacity. A few middle-order batters posted slow domestic averages because the side loaded them with scoreboard responsibility and gave them no freedom to play at their own tempo. Change the role at international level and the numbers change. Where the eye test beat the model, I log each instance separately — otherwise the model stays pleased with itself.

The alternative explanation must stay on the table. Perhaps the problem is not the batters but the pitch-preparation schedule and the behaviour of the ball — that what we call a "skill deficit" is actually the imprint of a preparation failure. If NCL decks were deliberately prepared closer to Test conditions, the conversion arithmetic would itself change. That possibility cannot be dismissed; it has simply not been measured yet.

A further doubt: BPL strike rate may be a better conversion index than NCL averages, because overseas quicks are present. A weak sample is not automatically a useless sample — the question is what the sample was taken against.

What would prove me wrong? Three conditions. One, if across the next two seasons the NCL's 135+ pace share rises materially and the conversion rate still does not improve, my pace-division argument weakens. Two, if a full control season shows the conversion gap for high-average domestic batters collapsing below ten percent, then the gap is generational, not structural. Three, if the same result appears without venue adjustment, my innings-par factor is a redundant variable.

Takeaway: What I Will Watch Next Round

I do not write predictions; I write a review date. Next NCL season I will watch three things: first, the count of batters scoring outside the flat-deck cluster; second, the list of seamers consistent across two pitch clusters over two seasons; third, how much of a domestic player's BPL auction price is set by on-field indices and how much by camp noise.

My ISTJ habit is simple: audit the row, then trust the trend. The 914 cannot be erased, just as the 23.8 cannot. The question is which ledger you accept as the true record — the one written on the field, or the one we have not yet learned to read?