The Pressure Over Index: A Ball-by-Ball Audit of Bangladesh's Middle-Over Collapse in Asia's Spin Fog

Hook

I keep returning to a night last month. Chasing 172, Bangladesh were 89 for 4 at the end of the 13th over, needed rate already above nine. Of the next seven deliveries, six were dots and one produced a single. From my room in Khulna I opened the ball-by-ball ledger — the file I have kept since 2026, where every delivery carries a timestamp, a batter, a bowler type, and the required rate at that exact moment. What the ledger showed, the scoreboard did not: the target was not growing, the pressure above the target was. Not a single wicket fell in those seven balls, yet the partnership's expected value dropped by twenty-six percent. That night fixed the question I have been chasing since — is Bangladesh's middle-over collapse in Asian conditions a random batting failure, or a recurring system state?

Context

Before any investigation I do two things: define the index, and write the hypothesis down first. This audit uses two. The first is the Pressure Over Index (POI) — an over counts as pressured if at least one of three conditions holds: (a) one or more wickets fell in the previous two overs; (b) the current run rate trails the required rate by more than 1.8; (c) the over contains at least two style-dots. The second is Recovery Efficiency (RE) — runs scored in the three overs after a wicket, divided by the baseline run rate for that phase. For the baseline I used non-pressure overs from the same venue, the same bowling attack and the same pitch, so venue effects are stripped out as far as possible.

Sample window: 2026 to 2026, 63 completed Bangladesh innings in Asian conditions (Bangladesh, Sri Lanka, UAE, Pakistan) where the spin share exceeded 35 percent. Variables were deliberately kept few; overfitting was my old sin, and I wrote the penalty for it into a fixed rule in January 2026. I do not chase outliers; I follow them until they confess.

Core Analysis

A Bangladesh chase splits cleanly into three strata. Powerplay (overs 1–6): run rate 7.8. Middle phase (overs 7–15): 6.4. Death (16–20): 9.1. In spin-heavy matches the middle phase drops to 5.9 and the dot-ball rate climbs to 42 percent. After the powerplay, boundary share falls from 28 percent to 14 percent, while single rotation drops by five points. The real centre of the collapse is overs 7–15, and the fault there is not tempo — it is strike rotation.

Recovery Efficiency tells the story more cleanly still. After a wicket in the middle phase, Bangladesh score at 5.2 in the following three overs against a same-phase baseline of 6.9, giving an RE of 0.75. Over the same window and conditions: India 0.94, Sri Lanka 0.88, Pakistan 0.81, Afghanistan 0.79. Bangladesh sit at the bottom of the Asian group, and the cause is not the number of wickets falling — strike change over the six balls after a wicket averages 4.1, well behind the baseline of 5.9.

Player level changes the picture by tier. In this window, Towhid Hridoy's middle-over RE is 0.91, Mehidy Hasan Miraz 0.86, Jaker Ali 0.74, Tanzid Hasan 0.68. Shakib Al Hasan sits at 0.97 — and on Shakib, the record of 4,000-plus ODI runs alongside 300-plus ODI wickets still belongs to him alone (source: ESPNcricinfo Records, updated 2026). The problem is that his middle-over investment time since 2026 has roughly halved against the previous decade. The gap that opened is not about missing boundaries; it is the slow path from dot to single.

The bowling side shows the mirror image, and that is the biggest information gain here. In Asian conditions, death-over economy is 8.1 for Taskin Ahmed and 8.6 for Mustafizur Rahman, while Rishad Hossain goes at 7.3 in the spin phase. Bowl Bangladesh at a rival's pressure points and the rival's RE also falls to about 0.78. That symmetry says the issue is not skill; the issue is routine.

Contrarian Angle

The numbers did not break the model; they exposed where the model was blind. My POI definition labels a wicket falling in the twelfth over as a "failure," yet taking risk in the twelfth over of a 172 chase is a reasonable cost. Recovery Efficiency also punishes a deep batting order unfairly, because a side that bats late is not recovering — it is saving resource.

The genuine leak is not in the fall of wickets but in the "silent pressure overs" that arrive with none: overs 12–16, where runs come at less than half a run per ball and the strike log records it exactly as it records a wicket. Mirpur, Chattogram and Dubai behave differently; on those low-bounce surfaces the baseline itself drops by 0.6 in run rate. My old mistake was entirely different — explaining 28 off 32 in one innings with 28 off 32 in another. Without a pitch report and dressing-room conversation this index is half a truth, so for this audit I checked separately with a coach and a physio — something a table of clean numbers can never do.

Takeaway

In the next bilateral series I will watch one number closely: dot-ball percentage in overs 12–16. Locking it in now — if that figure stays above 38 percent in a chase, Bangladesh's modelled chase-win probability in Asian conditions stays below 45 percent. The revision rule is pre-registered too: after the sixth over, if dew arrives and the pitch is low-bounce, the baseline must be revised down by 0.4. Expected truth is not a verdict; it is a timestamped estimate, and when it fails I point at the model first. The question now is not statistical — it is a question of routine. — Root: 2026, launching 'Expected Truth' in Khulna.