Speed Noise, Wicket Silence: A Data Audit of Bangladesh's Pace Attack

Twelve at night in Sydney, and I open eighteen months of ball-by-ball logs. Bangladesh's fast bowlers have added two and a half to three kilometres per hour in the first two overs with the new ball. But expected runs per ball in those same overs have not fallen; boundary rate is up nought point two percentage points. The speed went up. The control did not. Scoreboards never show that gap, because scoreboards only count runs and wickets.

In Mirpur it was an evening game. Under the floodlights the bowler was measuring the length of his run-up, while I sat in Sydney running three columns: speed-gun reading, length zone, shot outcome. One thousand eight hundred and forty-four deliveries. That is my first witness. The stadium itself is the second.

When those two witnesses disagree, I go back to the stadium.

Method: what expected value is, and why it is not the last word

What football calls xG, I lower to the delivery level in cricket. Every ball is split into six variables: speed band, length zone, line, match phase, innings state, and batter handedness. Each bin carries a baseline value built from my own logged deliveries and broadcast tracking. I call it a ledger rather than a model, because I know the size of its errors.

Three limits. First, speed guns carry a plus-or-minus one point five kph error, and movement tracking carries more. Second, I plot the pitch point by eye, so length zones can drift half a metre. Third, my death-overs sample is only two hundred and fourteen deliveries; one bad over in one match can flip the whole picture.

I refuse the trap where an analyst starts believing his own number is proof. A number is a claim, not a truth; it has to survive the conditions of the ground first.

The evidence: four findings, and which ones hold

One, pace and run rate rose almost together. In my log, the 135-140 band costs nought point eight nine expected runs per ball. The 140-145 band costs nought point nine four. The 150-plus band costs nought point nine seven. From outside it looks as though pace is being punished. But the samples are not equal: roughly seventy-four per cent of 150-plus deliveries came in the last powerplay over or the death phase, where batters are already forced to take risk. That is phase bias, not a verdict on speed.

Two, the real story is length. Good length costs nought point seven one expected runs per ball. Full costs one point zero two. Short costs nought point nine six. Back-of-a-length costs nought point eight three. The number that matters is this: thirty-one per cent of the 150-plus deliveries landed in the short zone. When the bowler reached for extra pace, the release point drifted back.

Pace that pushes the length back stops being pace and becomes a gift. On a flat IPL or Big Bash deck, a 152 kph short ball is just six runs of bat speed. I watched Mustafizur Rahman's four for twenty-nine for Chennai against Bangalore in the 2026 IPL more than once; the source of that success was never raw pace, it was the cutter-and-slower mix in the death.

Three, each phase is a different sport. Powerplay costs nought point eight three expected runs per ball in my log. The middle overs cost nought point seven four. The death costs one point five two. A strong phase can hide a weak one in a single result, but never across a series. Bangladesh's current attack is strong up front and exposed at the end; the damage usually lands in the last four overs.

Four, the wicket column is a lagging witness. I keep repeating one rule: I do not trust a number I cannot trace to a touch. Dropped catches, a ball seaming past the edge, a DRS error — none of it changes the process, only the result. One series with five wickets and the next with none may describe two identical spells of bowling.

Conditions and venues: Rawalpindi may be teaching us the wrong lesson

In August and September 2026, Bangladesh won a Test series in Pakistan two-nil, their first ever on Pakistani soil. Mushfiqur Rahim made 191 in the first Test, Litton Das 138 in the second, and Mehidy Hasan Miraz was player of the series. The internet settled on a phrase: a Bangladeshi pace renaissance.

I am sceptical of that phrase. Rawalpindi was slow and low. What worked there was not raw speed; it was seam position, care of the ball for reverse swing, and the patience to hold a fuller length. Two years later, at the 2026 Champions Trophy, Bangladesh lost all three group games, to India, New Zealand and Pakistan, and the tournament ended with a final in Dubai. On a flat deck the fast bowler's marginal edge evaporates inside an over, and evening dew turns a full ball into a gift.

A small sub-list in my ledger covers neutral venues with partial crowds and heavy dew: in those conditions Bangladesh's powerplay economy with the ball moves from six point eight to eight point four. Four years ago I found the same pattern in football's empty stadiums, where home scoring margin dropped by nought point two five goals. The model said one thing; the empty stadium said another. In cricket the equivalent is not an empty ground but a neutral one, plus dew.

The other side: what correlation would get wrong

Genuine pace in the first over and economy at the death move together in my log. That is a condition, not a cause. The bowler who can hit 147 in the first spell is usually the senior bowler, with more international matches and a better sense of how to preserve the ball for a second use. Experience sits behind both variables. Strip out experience and read only speed, and the forecast is weak.

There is a larger trap: confusing the bowler with the mood. Under workload pressure a bowler shortens his spell, takes an extra over off, and does not return for the final burst — and the scoreboard never shows it. Across franchise and Test cricket in 2026-26, injury history is quietly pulling that lever. Nahid Rana is still in his early twenties, yet his count of 140-plus deliveries per Test already exceeds that of our senior fast bowler.

I have written before that the hardest part of returning from injury was never muscle or ligament; it was indecision. For a fast bowler it returns as a stress fracture story: when the rhythm of footfall in the run-up breaks, pace falls on its own, and the bowler reads that as a loss of rhythm rather than a loss of pace, and starts changing his length. Athletes who have bowled for years know the return is quick but the old recklessness takes much longer.

What I will watch next: three explicit thresholds

I do not believe in leaving judgement until after the fact. So I write my checkpoints down first, in case I am tempted to defend my own numbers later.

First, expected runs at the death. If over the next five matches the figure does not fall below one point four zero per ball from one point five two, then the problem in this attack is strategy, not workload.

Second, the yorker rate. What share of death deliveries hit yorker or toe-crushing length? My log says seventeen per cent. The best T20 attacks sit between twenty-six and thirty.

Third, the link between pace and length with the new ball. If the short-zone share stays at thirty-one per cent while the speed band climbs, the bowlers are still bowling on adrenaline rather than to a plan.

The 2026 T20 World Cup is in India and Sri Lanka. On Sri Lanka's slower, spinner-friendly surfaces Bangladesh's pace weapon becomes a supporting act, and the spinners carry the load. Speed will be discussed far less; control will be discussed far more. That is the real test for me, because whether an attack built on pace can survive on length is still an open question.

One question to leave open, because it will take me time to answer alone. Small samples are loud; large samples are honest. Which sample is Bangladesh's pace attack walking towards: the four-over celebration, or the four-year accumulation?