The Expected-Runs Autopsy: Where T20 Powerplay Narratives Break Down

The Expected-Runs Autopsy: Where T20 Powerplay Narratives Break Down

[Hook]

In September 2026 I was sitting in the press box at Mirpur's Sher-e-Bangla National Stadium, staring at a number that refused to match the arithmetic in my notebook. Bangladesh was winning a T20I series against New Zealand — their first at home against New Zealand, and my debut series on the T20I commentary roster. On the big screen the powerplay strike rate glowed like a verdict. The senior colleague beside me said, “The boys have finally learned to attack.” I told him this was not a story about attack; it was a story about balls. The dot balls the crowd was reading as caution were, in my model, debt — and debt is repaid with interest in the last five overs, when a strike rate can no longer buy back the past.

By the end of that series a file sat on my laptop named powerplay_myth. It had three columns: expected runs per over, ball-by-ball wicket probability, and phase-adjusted impact. Whatever the six-over strike rate said, those three columns told a different story — the powerplay is really settled in overs sixteen to twenty.

[Context]

My first assignment after joining the Mumbai new-media outlet The Field in 2026 was to build a model for the Champions League final. Real Madrid beat Juventus 4-1, but the model showed Madrid generating 2.6 expected goals against Juventus's 1.2, while Juventus pressed with a PPDA of 7.1 in the first half. I wrote that the final was not a 4-1. Once that piece circulated through Indian football circles, my method changed: no more opening with scorelines and quotes. Every analysis began with xG, PPDA and shot maps. “I performed the first xG autopsy in Indian new media; the body was a narrative.” That sentence is still my statement of work.

The following year, at the Russia World Cup data desk, I worked on Germany's 0-2 defeat. Germany held 70 percent possession, took 26 shots, generated 2.7 expected goals — and pressed with a PPDA of 6.8, leaving space behind. South Korea produced 1.1 xG from two counters. I had published before the match that Germany's possession was a warning, not a virtue. Three European outlets cited the model after their exit. ( — Root: Experience 2, Germany )

The method does not transplant cleanly to cricket. Football's event space is continuous; cricket's is discrete and low-scoring. But the skeleton survives. First isolate the claim. Then build a baseline — the same pitch, the same quality of opposition over three seasons. Then stress-test the claim with expected runs, wicket probability and phase-adjusted impact. A claim that cannot survive the stress is a myth.

Three instruments do the work. Expected runs (xR): what a delivery should have yielded given pitch type, line and length, field setting and the quality of the bowler. Wicket probability: who is likelier to fall in which phase, and crucially who replaces him. Phase-adjusted impact: how much a run was worth when it was scored, not merely how many runs there were.

[Core]

Powerplay strike rate is the most used and most distorted number in the game. In the first six overs only two fielders are out, the ball is hard and new, and attack is easier — but easier is relative. On Mirpur's slow, low-bouncing surface a new ball does not come onto the bat; a 120 strike rate there is evidence of aggression, while the same 120 on a batting-friendly deck is surrender. When one number means two things in two places, it stops being a benchmark and becomes an illusion.

My model carries an entry called dot-ball debt. Suppose a side makes 45 off 36 balls in the powerplay — seven and a half an over. If the baseline is 54, the nine-run shortfall is not merely nine runs; it must be recovered across the remaining fourteen overs by taking extra risk. And the worst time to take extra risk is overs seven to fifteen, when spinners bowl and wicket probability peaks. A dot ball in the powerplay is therefore a small shove toward a wicket two phases later.

Here sits the second error. We weight every wicket equally. In reality the weight shifts by phase. Losing Litton Das in the second over costs comparatively little, because the incoming batter's first ten balls carry the lowest expected runs — and those ten balls fall inside the powerplay, when fewer fielders are out. Losing Mahmudullah in the sixteenth over costs the most valuable batting overs of the innings, and the new batter is not asked to settle; he is asked to strike immediately. One wicket, two phases, two prices — and the scorecard renders both identically.

The third layer is pitch and conditions. Watching from press boxes across several seasons, I have seen visiting sides' powerplay strike rates dip on Dhaka's slow surfaces, because the ball does not sit up and spinners are operational from the second over. The question is whether that dip is caused by the crowd or by the pitch. Across several seasons I set empty-stadium matches beside full-stadium matches; in some leagues home advantage barely moved with nobody in the ground. What we call home pressure is largely pitch preparation and familiarity; only a smaller share is crowd noise. Where a board changes the surface, no crowd can change it back.

The fourth layer is selection and auction pricing. BPL pitches are slow. A young batter posting a 140 strike rate there gets a big franchise price, yet that 140 does not translate away from Bangladesh. India's IPL has a vast sample, so its modelling errors surface quickly; Bangladesh's sample is small, so its errors survive for years. In squad construction, youth potential is systematically overrated and dressing-room continuity systematically underrated. Runs in a single innings come from the bat; runs across a season come from clarity of role — who owns overs one to six, seven to fifteen, sixteen to twenty. Without that clarity, a powerplay advantage evaporates at the death. The true value of a Shakib Al Hasan, a Mushfiqur Rahim or a Mahmudullah never appears in a runs table; it appears in the stability of the batting order.

The fifth layer is cross-border comparison, where my experience sits at two poles. Much of Bangladesh's cricket journalism remains narrative-first — player as villain, player as hero, and nothing in between. India's new media can open with a table and a field map because the data supply chain exists. Germany's Bundesliga culture taught me that data means pre-match warning, not merely post-match explanation. That warning is written less often in Bangladesh, because publishing a model before the toss risks annoying the audience. My 2026 experience says the warning is the product.

The sixth layer is a specific trap. The easy way to raise powerplay strike rate is the aerial shot, and the aerial shot raises the probability of a boundary catch. If a model does not price the cost of the wicket against the gain in runs, we push players into a style that lowers the team total. In a powerplay over, taking seven to nine without losing two wickets is often worth more than taking ten to twelve with them gone, because a set batter's presence from overs seven to fifteen rewrites every subsequent calculation. The eye does not see that gain; it sees the total.

The seventh layer is time. In T20 the relationship between form and surface shifts with the calendar. Winter evening dew makes the ball hard to grip and reduces spin, which changes powerplay maths too. A September night at Mirpur is not a January night at Mirpur. Selectors who read only powerplay strike rate, without dew schedules, humidity and the date of the last rolling, are deciding on half the information.

[Contrarian]

My method's greatest enemy is my own model. Success at football autopsies can suggest the instrument is universal, but cricket's event space is different: lower scores, higher variance, and a match turned by a handful of decisions inside 120 balls. On small samples any model can look confident, and an INTJ temperament indulges that confidence. So I write down a falsifiable hypothesis in advance — and the condition under which I would call it false.

My current hypothesis: Bangladesh's T20 powerplay problem is mostly procedural, not personal. The test: if reordering the batting or changing intent does not durably raise powerplay xR within ten matches, then the problem lives in pitch preparation, the domestic calendar and the training pipeline — and changing players is wasted time. If the evidence goes the other way, I will move my own hypothesis aside without complaint.

A second caution matters more. Correlation is not causation. Teams that score heavily in the powerplay do not always win — on slow, low-scoring surfaces the relationship weakens, because the side batting second gains an advantage and defensive bowling decides the match. Television panels present that relationship as law. Before reading the top line of a scorecard we must ask: which pitch, which humidity, which attack?

One uncomfortable truth remains. My autopsy can be so cold that the game disappears for the viewer. The beauty of a Kane Williamson or a Devon Conway innings does not appear in any xR column, and pretending otherwise turns analysis into arrogance. Numbers are not the last word; they are where the objection can be raised.

[Takeaway]

Three things I will watch over the next twelve months. First, powerplay dot-ball percentage — the most honest pre-match signal, because rhythm, not speed, is built there. Second, whether selectors convert domestic strike rates into pitch-adjusted expected runs. Third, and most far-sighted: ball-tracking data will eventually be stored in immutable, verifiable ledgers, at which point Bangladesh, India and Germany can be placed on one standard — and no journalist will be able to build a story alone on the claim of “my theory.”

There is no column for luck. But why the television panel's story about “intent” has survived year after year without ever being opened — that is the puzzle I cannot yet solve.