Ledger Truth, Scorebook Lies: A Data Audit of Bangladesh at the T20 World Cup

I keep returning to one specific over from the 2026 T20 World Cup. In a group match, Bangladesh needed 27 runs from the last two overs. The broadcast win-probability graphic changed colour four times in six balls — 38 percent, 61, 42, and finally 57. A young colleague sitting beside me said, "What a dramatic match." I said drama belongs to the screen. My question was plainer: where did those balls actually land, who faced them, and how late did the graphic catch the truth? Before the match even ended, I ran my own model on a 15-second refresh and laid it beside the broadcast numbers. The gap was not something to hide. The scorebook was telling one story, the ledger another. The live xG model blinked first in Russia, and I learned to wait; so this time I held my discipline for three balls before drawing a conclusion. T20 cricket's two biggest lies are the scoreline and the word "intent." This piece is an audit of both. But first, a newer thought that interested me most this tournament. Cricket data still lives in scattered notebooks — every broadcaster, every fantasy platform, every coaching app builds its own definitions. I argue cricket needs a genuine data ledger: a shared, verifiable, tamper-proof record where every fact of every ball is written once and read identically by everyone. In blockchain terms, each ball is a block and each match a chain. But — and I will open this "but" at the end — an immutable ledger is not automatically a correct one. A tournament cycle is a strange thing. A whole year of work is compressed into four or five matches, and under that pressure everyone leans toward narrative. Flags, highlights, captain's speeches all gain weight, while numbers lose theirs. My experience says numbers are needed most precisely under that pressure. For more than eight years, from domestic leagues to World Cups, I have watched the same mistake: teams fail to set thresholds before the tournament, make decisions on emotion mid-match, and build explanations afterwards. I found the Rangpur newsletter in a drawer, still predicting the future. That was 2026. Sheikh Russel KC missed a playoff place by just 3 points despite outshooting opponents 87-64. I learned then that shot volume hides shot quality. From that accounting came "The Rangpur Data Monk" — a twelve-part xG and PPDA audit that eventually forced three clubs to adopt a single xG definition. I apply the same lesson to cricket today: the number of boundaries does not say how hard each boundary was. Here is the problem. Cricket has no single definition of expected runs, or xR. One platform measures what a batter would score from his own history, another what an average player would, a third separates line, length and field setting. For the same over, three platforms show three numbers, and the viewer is left confused. That is not data democracy; that is disorder. So my proposal is plain: cricket needs a common ledger. Every ball's data — pitch line, length, pace, the batter's footwork, the field placement — should be written once into an open, verifiable record. No one can alter it later, because each page of the ledger carries a hash. The core blockchain idea works directly here: verify rather than trust, and verify against a single source of truth. I think within three years at least one major league will move this way. Now to the field. I broke Bangladesh's whole tournament into four parts — powerplay, middle overs, death overs, and bowling load. In the powerplay the team's run rate was 7.4, which is not bad in itself, but they lost wickets inside two overs in four of six matches. That means the side scored while demolishing its own foundation. This is where the scoreline lies. Forty-four runs in the first six overs looks fine, but 44 for 2 is not 44 for 0 — priced properly, the first is worth about 38 given the value of holding two wickets for the next fourteen overs. By my xR model, Bangladesh's middle overs (7-15) yielded an average of 71 runs, while comparable sides yielded 78. The difference is seven runs, but the cause is not runs — it is dot balls. In the middle overs, 29 percent of Bangladesh's balls were dots, above the tournament average of 24 percent. Just as PPDA measures pressing intensity in football, dot-ball percentage measures middle-over pressure in cricket. More dots mean less pressure, which means a heavier accounting burden on the batters. In the last two overs, Bangladesh scored an average of 10.3 per over — a striking number. The inner picture is different. Success with yorker-led deliveries was 46 percent, while success with short or bouncer-led deliveries was 31 percent. The team's death-over plan rests on one specific tool, and the moment that breaks the arithmetic flips. Mustafizur Rahman's cutters were sharp this tournament, but when the pitch slowed, his economy moved from 6.8 to 9.1. Same bowler, same action — only the pitch changed, and the model got stuck. Take Litton Das separately. His strike rate of 147 sounds superb. But his boundary percentage in the powerplay was 18 percent, against an average of 23 percent among the tournament's best openers. He finds the ball, but sends it to the rope less often. This is where the word "intent" becomes hollow. Intent is an attitude; boundary percentage is an outcome. Pick attitude while mistaking it for outcome, and you err. Towhid Hridoy was this side's cleanest accounting. In the middle overs his actual runs beat his xR by an average of 9.4 per innings. Teams routinely underrate this kind of player because he does not make spectacle — he balances the books. And Rishad Hossain's leg-spin was a quiet weapon — an economy of 6.2 in the middle overs, though his wicket-taking ball rate was low. He creates pressure rather than absorbing it. Count only wickets and you undervalue him. Now the load question, which no one shows live but which wins and loses tournaments. Bangladesh's pacers averaged 3.7 overs across consecutive matches, but the gap between matches was just one day on four occasions. In my load index, Taskin Ahmed's pace in his third spell was on average 4.2 km/h below his first. That is not personal failure; it is a schedule's outcome. A 30-year-old fast bowler's body does not accept six matches in twenty days, and if the coaching staff do not plan for that in advance, there is no point regretting it after an injury. For all of this I want one number — a 0-100 efficiency score that captures each innings alongside load, strike rate, wicket value and situational difficulty. At tournament's end, Bangladesh's efficiency score stood at 64, below the 78 of the top four sides. But I offer this number as a witness, not a religion — a witness can be cross-examined, a religion cannot. Here I turn the other way. Many will praise my ledger proposal, because "blockchain" is magic now. But be careful. An immutable ledger does not mean its contents are true. Once false data enters the ledger, it becomes false forever — the ledger grants immortality, not truth. In football my rule was this: before trusting a model, let it survive a cold Tuesday. The same applies to cricket's ledger — verification at the door, immutability inside. The second trap is mistaking correlation for cause. Of the three sides with the highest strike rates to reach the knockouts, two exited in the group stage. Strike rate and winning correlate; they do not cause each other. The causes are holding dot-ball pressure down, conserving wickets through the middle, and stability of plan at the death. A side chasing only strike rate wins highlights and loses matches. The third trap is small-sample enchantment. Judging a player on six matches of data is like calling the weather from a coin toss. I keep a ledger of misses, because the hits already have press officers. Where my model erred this tournament — especially spinners' economy on slow pitches — I wrote it down, because that is where the next model learns. And then there is the toss story. Everyone says the side that wins the toss wins the match. In my count, toss-winners won 53 percent of the time this tournament — roughly a coin flip across nine or ten matches. People love to find patterns; the numbers say it is a game. So what do I watch in the next round? I will write my thresholds before the first ball: fewer than two wickets in the powerplay, dot-ball percentage below 26 in the middle overs, and yorker success above 50 percent at the death. Clear those three numbers and a side stays in the fight, whatever the strike rate. The team does not need more data; it needs one number it can defend. At sixty-eight I have understood one thing. A model can tell us which path raises the probability, but it never decides for us. People decide, and so people must decide in advance — which number to trust and which to discard. Blockchain or xR, both are only ledgers. A ledger does not speak the truth; it remembers. And remembering is what cricket needs most — because cricket forgets, and every tournament it repeats the same mistake fresh.