The Death-Overs Ledger: Three Rows Bangladesh's Scorecard Never Prints

Zahur Ahmed Chowdhury Stadium, Chattogram. One evening of the domestic season, sitting thirty yards from the dugout, I logged ball by ball — which delivery was a yorker, which wide yorker became a full toss, whether the batsman swung or was beaten. At the end of the innings the scorecard recorded 4-0-31-2 beside one bowler's name. Anyone in the stands would call it a good spell. The third column of my notebook, the one no scorecard ever prints, says otherwise: fourteen of those balls landed in the slot. Six executed the plan; four had a plan but missed the spot; four had no plan at all — just a ball flung under pressure. One of the two wickets came from that fourth category, where the batsman made his own mistake by reaching for a big shot.

That night I opened the old ledger. In 2026 I started a Bengali-English data blog from Chattogram and hand-logged 1,847 shots across 132 Bangladesh Premier League matches. A local betting syndicate told me a woman could not keep this account. I did not close the sheet. Today those 1,847 rows are my strongest witness, because one row from that evening never reconciled — and the Chattogram desk taught me that a missing row is a louder story than a headline.

Context: why overs 17 to 20 are Bangladesh's real examination

In a regular season the death overs decide nothing about the table, and everything about how the next squad is built. Franchise bowling quotas shift year to year, the national fast-bowling pool stays wedged at four or five names, and the entire season's death workload lands on those few shoulders. What happens to a bowler who sends down death overs in a year — in his shoulder and in his head — appears in no column.

My method is plain. Beside every death-over ball I fill four cells: the plan (line, length, type), execution quality (three tiers — executed, nearly executed, missed), batsman intent (swung, defended, beaten), and whether the field setting matched the delivery. Four columns over one match's death overs yield 24 to 48 rows. A season of twenty-odd matches reaches roughly 900 balls. Only past 900 balls do I write a verdict.

Years of watching tell me there is a grey zone between what a spectator sees in the death overs and what an umpire records — no scorecard reaches it. I try to measure it. I never publish a line without sample size, data source, and error bars.

At the 2026 World Cup in Russia I calculated France's PPDA at 15.8 against Argentina's 8.9 in the 4-3. I followed France not because they won but because their pressing structure was stable. My ledger afterwards showed Argentina's three goals came from 0.9 xG in total. Since then PPDA has been a fixed column in every match preview I write.

It is worth stating precisely what PPDA measures: passes allowed per defensive action. Cricket has no direct replacement. I built a proxy — in the death overs, the ratio of the opposition's aggressive shots to each intervention (dot, boundary prevented, wicket ball). I call it the Death-Over Pressure Index. The disanalogies must be stated or the analysis becomes dishonest: football pressing is continuous and 11 v 11, while bowling is discrete, six balls to an over, and fielding restrictions shift by phase. So I treat the index as a signal, not a claim, and I record its falsification condition: if across a series the index moves with no relationship to death-over economy, I change the framework.

In 2026 I studied 83 Bundesliga matches before and after Project Restart and found home win rate falling from 43.2% to 33.8%. I cut home advantage in my betting model by 18% and tested it across 27 matches. The lesson is simple: crowd, venue and environment are ledger rows too. I apply the same caution to pitches in Chattogram and Dhaka.

At Euro 2026 I refused the Pedri hype despite 629 minutes and 92% pass accuracy, because only three of ten teenage midfielders since 2026 sustained elite output beyond 900 minutes. In cricket that filter becomes balls: 900 death-over balls. The 900-minute rule is a monastery bell: it calls you back from magical thinking.

The execution column: the gap between plan and delivery

The three deliveries Bangladesh bowlers use most in the death overs — the cutter, the wide yorker, the slower ball — raise a question that is not "what did he bowl" but "did he bowl what he intended." In my hand-written ledger, a large share of domestic death-over balls had a plan for a full yorker in the slot, while execution landed near half. A ball that becomes a half-volley on the way to being a yorker is recorded the same way on a scorecard — four or six. Forty runs appear beside the bowler's name. Which part of those forty belongs to a failed plan and which to failed execution, the scorecard never separates. I keep two budgets: a wrong plan is the coaching staff's debt, a wrong execution is the bowler's. Collapse the two into one verdict and nobody learns anything.

This separation has a practical consequence. In domestic cricket the same bowler often wins a match with superb yorkers one night and is dropped the next after being hit off full tosses. That variance is a function of sample size. Six good yorkers followed by three failures is not a decline in skill, it is the noise of a small sample. The scorecard turns that noise into a permanent row.

The intent column: the batsman hidden inside the wicket

The next column holds batsman intent. Every death-over delivery is one calculation — who errs first. So I record whether the batsman chose to attack or was pushed into the shot by the ball. That distinction is the centre of death-bowling analysis.

Consider Germany. At the 2026 World Cup in Qatar they lost 1-2 to Japan despite 26 shots, nine on target and 1.95 xG, against Japan's 1.36 xG. Germany's PPDA was 7.2 — pressing high, leaving transitions open. My ledger showed Japan's two goals came from 0.4 xG. Some called it a collapse; I called it process dominance without ledger balance. Cricket repeats this exactly: a side bowls six dot-ball overs, then concedes two sixes off one bad length and loses. The statistics say the process was good and the result bad. At the betting table only the result is counted.

So I added a further cell to the intent column: the "forced shot." In the death overs a large share of aggressive strokes come from required rate pressure, not the batsman's plan. That is my bowler's success, invisible as a wicket or a dot. While assessing Lamine Yamal during Euro 2026 and the Paris Olympics I applied the same caution — one goal and four assists in 507 minutes, and still I waited for 900. When youth and sample size move together, patience is the only instrument. Shakib Al Hasan's 606 runs and 11 wickets at the 2026 World Cup, by contrast, is a complete row: both columns filled. Complete rows are rare, which is why they cost more.

The field column: nine fielders who were never there

The third row nobody prints is the field setting. A scorecard only says whether the ball was good. But a ball is good or bad precisely because of where nine fielders stood.

An example. A fast bowler sends down a wide yorker, with third man and point set. The batsman plays straight, for four. The ball was not poor. The delivery type and the field setting simply did not match. In my field column that event reads "execution complete, structure failed." On the scorecard it is only four.

My suspicion about field settings runs deepest in domestic cricket because coaching information is scarce there. International matches give practice footage, drone views, fielding maps. In domestic matches I sit near the dugout with my notebook. That notebook is my only drone. After years of filling it, a pattern surfaced: in the death overs, deep midwicket stands two to three yards finer in domestic cricket than in the IPL. As a result the slower ball does not work, because the batsman reaches it with a pull or a lofted drive. That is not an individual bowler's error. It is a structural setting.

The contrarian angle: the crisis may be real, but we are blaming the wrong place

Now I must stand against my own argument. Suppose Bangladesh's execution rate in the death overs is genuinely low. Does that prove a shortage of bowling talent? Not necessarily. My first doubt is context contamination. Death-over economy depends heavily on how many runs a side is defending. Under the pressure of defending 120 a bowler sends down different deliveries than when defending 230. Averaging the two situations into one table produces a number that recognises no individual bowler.

My second doubt is sample and selection. Franchises habitually use four or five fixed bowlers at the death, producing enormous workload on few shoulders and high variance in the record. In a small sample, variance always looks like a lack of skill.

The difference between correlation and causation must be honoured here too. Death-over economy correlates with winning, but claiming good economy causes victory is unfair. After cutting home advantage by 18% in 2026 I learned that adjusting a variable makes a model more honest, not more accurate. Death-over analysis needs the same correction — at least a 15% context adjustment, with scoreboard situation sitting in its own column.

There is a human dimension easy to forget at the data table. In Chattogram a young bowler conceded heavily across two straight death-over spells and was dropped. In my ledger his execution rate was better than the team's average; the failures belonged to field setting and context. The scorecard tied him to one row. Nobody brought him back. A missing row never costs less than a career.

The Soumya Sarkar byline, 2026

In 2026, while at The Daily Star, I interviewed the rising Soumya Sarkar; the piece was later picked up by Prothom Alo — my first verifiable byline. One sentence from it I have never revised: a young player needs more time than he needs data. Bangladesh's first Test came in November 2026 at the Bangabandhu National Stadium in Dhaka, against India. Almost nobody in that side was ever given a run of 900 death-over balls. The structure was the same then as it is now.

What I will watch in the next series

Three marks sit in my notebook. First, the percentage of death-over balls that landed in the planned slot — not the runs. Second, the ratio of field settings that match the delivery type, especially on Chattogram pitches. Third, the delivery count of any bowler making his first death-over appearance, so that nobody is hanged before the 900-ball bell rings.

A player who has handled the death overs in domestic cricket for years may carry no asterisk on any scorecard. His ledger is still full. Staying professional requires an archival patience that headlines almost never print. On squad announcement day I will reconcile the numbers again. If they do not reconcile, I will change my framework — because evidence is the only loyalty I keep.