Auditing Abhishek Sharma's 35-Ball Century and India's 118-Run Powerplay: Where the Third T20I Was Actually Decided

Abhishek Sharma made 113. Of those, 102 came in fours and sixes — 20 of his 38 balls reached or cleared the boundary, and the remaining 18 balls produced 11 runs. I ran that arithmetic three times on the scoring app, because a strike rate of 297.4 should mean at least some of the scoring came from running. It didn't. Ninety point three per cent of Abhishek's innings was boundary-dependent, which means it was built almost entirely on the one thing a bowler can most directly control — and he won that contest twenty times out of twenty.

That number will not survive into Wednesday's headlines. The headline will be the 35-ball hundred, and it will be correct: a 35-ball hundred is the cleanest, hardest measure T20 cricket offers. But the same innings holds a second number, and that one is misshapen. India chased 199 with six wickets down and forty balls unused. Six wickets and forty spare balls do not usually share an innings. A side that finishes 199 in 13.2 overs does not lose six; a side that loses six does not finish in 13.2.

That contradiction is the actual subject of the third T20I. Abhishek's hundred is its brightest surface, and the 3-0 series lead is its biggest consequence. But the match's structure was laid in the 118-run Powerplay that New Zealand conceded under a rule of its own making.

Context: what was pre-decided, and what was decided on the ground

Shreyas Iyer won the toss and chose to bowl — the decision that generates the loudest questions when it loses and disappears when it wins. New Zealand lost Tim Seifert early, then Finn Allen and Daryl Mitchell gave the innings its momentum. Allen fell in the forties, Mitchell finished unbeaten on 89, Bevon Jacobs contributed a useful knock. The hosts could not cross 200.

India's bowling was not uniform. Bhuvneshwar Kumar again failed to leave the impression his experience demands — I have written this before, and I also know that one match's economy figure is not an evaluation of a bowler. But Prince Yadav's comeback at the death, and the yorkers from Arshdeep Singh and Nitin Kumar Reddy, together kept New Zealand short of the score from which the match would have left India behind. A target of 199 was not difficult; had it been difficult, this chase would not have ended in 13.2 overs.

A methodological confession is needed here, because I never read an innings like this off a scorecard alone. Football taught me what the cricket scorecard had been hiding for years — that goals and runs are outcomes, not processes. Germany went out in the 2026 World Cup group stage with 72 per cent possession and 1.2 xG per game; I called it false dominance and carried that frame across to cricket. But the translation is inexact. Cricket's events are discrete, its innings asymmetric, the bowler–batter relationship adversarial, and its scoring density several times lower than football's. There is no continuous flow to model the way xG models one. So I built no composite index for this report; I kept three separate ratios separate and refused to fuse them.

Core: the 118-run Powerplay is the price of a decision

The figure is this: India made 118 in the Powerplay, the highest by any Full Member nation in men's T20 internationals. The opening stand yielded 133. India were 178 inside ten overs.

All of it began with a selection call — Vaibhav Sooryavanshi sent out to open with Abhishek. That looks like an easy decision and is not. Changing an opener means removing someone from a position, and a side leading a series 2-0 does not usually disturb settled ground. India disturbed it. The consequence was that New Zealand's new-ball plan — clearly built around one left-hander — broke in the first over against two of them.

Which brings us to Mitchell Santner, the most expensive decision of the match. New Zealand withheld their captain and most experienced spinner from the Powerplay because two left-handers were at the crease. The logic was clean: left-arm spin into left-hand batters is a favourable matchup, so defer it. By the time Santner bowled, India owned the contest. The hosts conceded 118 in six overs and 178 in ten.

It matters that this was a bad decision without saying anything about Santner's bowling. It is about the decision rule that professional T20 has standardised over the past few seasons: matchup first, over second. That rule arrived from football, where individual duels can be measured, and we imported it into cricket without asking whether a T20 innings is one optimisation problem or twenty separate ones.

The rule's flaw is that it optimises per over while an innings optimises per innings. By holding Santner back, New Zealand protected one over's matchup and paid six overs for it. Put into a metric, I call it Matchup Deferral Cost: the runs conceded in the phase where your best matchup bowler is not bowling, minus the runs that same bowler concedes when he finally does. In this match the gap is large enough that no fine-grained analysis rescues it.

I should admit that the first Pressure Index I built, in 2026, was inverted. I combined dot-ball percentage with boundary-conceded rate and found that the bowler conceding fewest boundaries appeared to be under the most pressure — because I was measuring how the batter was playing, not how the bowler was restricting. I later reversed it, built a separate baseline for overs 17–20, and published Jasprit Bumrah's economy of 6.2 with a 45 per cent dot-ball rate at a time when he was not yet a headline. Failure first, framework second; that is the only order I have ever trusted — which is why the decision to withhold Santner can look defensible in the language of numbers and indefensible in the language of cricket.

Now the other side of Abhishek's innings. A 35-ball hundred means he scored at roughly a boundary per two balls. It also means 102 of his 113 came in fours and sixes, and only 11 runs came off the other 18 balls. The reading is that Abhishek did not accumulate; he took one very concentrated, very narrow risk and was correct every single time. If three of those twenty boundary balls had gone to hand, the same innings becomes 140 off 60 and nobody remembers it. Two endings of one innings sit very close together in this format, and that is the format's central measurement problem.

He was dismissed eventually — Kyle Jamieson, caught by Glenn Phillips at deep mid-wicket. The damage was done. Sooryavanshi fell to Matthew Fisher's maiden T20I wicket; that is another figure that becomes a headline on a different evening.

The number nobody looked at: 21 runs, 20 balls, six wickets

India were 178 after ten overs. The target was 199. They needed 21 runs at 3.5 an over across six overs. The match was over at that point.

They took twenty more balls to get those 21 runs, and lost wickets doing it. I do not have the ball-by-ball breakdown, so I will not claim all six fell inside that window. What I can say is this: across the last twenty balls India scored roughly 21 runs and lost however many of the six wickets fell in that span. The cost per wicket therefore sits somewhere between 3.5 runs (all six inside the window) and 21 runs (none of them). I am leaving that interval open on purpose.

Which means this match can be used to argue that India's lower middle order is strong, or that it is fragile, and both arguments would be dishonest. The conclusion that survives is duller and sturdier: India lost six wickets in a chase that required 3.5 an over for its final six overs. That is carelessness inside a win. In a knockout, a target of 199 sits sixty runs away, not twenty-one; on that day the same six wickets become a twenty-run defeat.

The hundred count is a numerator without a denominator

Abhishek now has eleven T20 hundreds. He has moved past Virat Kohli's ten and sits third among men, behind Chris Gayle's 22 and Babar Azam's 13. Quinton de Kock and David Warner also have ten.

The achievement is real and I am not diminishing it. But a count of hundreds is a numerator, and its denominator is missing from that list. How many innings Gayle took for 22 and how many Abhishek took for 11 is not shown, and I do not have a verified figure for Abhishek's total T20 innings in front of me for this report. So I will not construct the ratio. I will only note that anyone writing tomorrow that Abhishek is a finer century-maker than Gayle will not have the denominator either.

The other thing the list hides is era. The pitches, the boundary sizes, the Impact Player rule and the two-bouncer rule that Abhishek plays under are not the ones Gayle played under in the 2010s. Lists like this are not measurements; they are memory lists with the dates left out.

Contrarian: four holes in the record

The first hole is the sample. A 118-run Powerplay is one phase of one innings. Treating it as proof of a principle is exactly the mistake I made in 2026, when I declared a model validated off a knockout and watched it miss across the next four games. What this match supports is narrow: India made 118 in six overs while the opposing captain withheld his primary spinner. What it does not support is that this repeats.

The second hole is causation. Abhishek scored quickly and India won — which does not mean the 118-run Powerplay won the match. The target was 199 and New Zealand did not reach 200, which is an equally large cause. Had Daryl Mitchell's unbeaten 89 become 105 and the hosts reached 215, those six wickets would have been fatal.

The third hole is my own bias. I work on the data that exists — IPL, international cricket, televised coverage. Ranji knockouts and women's domestic cricket are not where the cameras are. During the 2026 furlough I built a valuation engine on ten years of IPL data from 2026 to 2026 and recommended Moein Ali at 70 lakh; CSK made the buy and won the 2026 title, with Ali scoring 357 runs at a strike rate of 138. I am not claiming that model was infallible. I am claiming it was falsifiable — and everything I have said about this match should be held to the same standard.

The fourth hole is the least comfortable. We are writing about Abhishek's 38 balls and skipping the passage in which six wickets went down for 21 runs. A team that is 118 for none and still loses six wickets inside 13.2 overs is auditing its own depth. Abhishek's hundred covers that question up.

Takeaway: what I will watch in the fourth and fifth T20Is

I am logging three things before the next two matches, so that I cannot revise my conclusions afterwards.

One: whether Santner bowls in the Powerplay in the fourth T20I. If he does, New Zealand has read this match as a process correction. If he does not, the matchup rule is institutional and the six-over price will be paid again.

Two: how many wickets India lose after the tenth over. The number was uncomfortable here. If India pass 170 and still lose wickets in the next two games, it is a pattern; if they do not, it was one evening.

Three: Abhishek's boundary-dependency ratio. Ninety per cent is not sustainable. The question is not whether it falls, but whether he has a second game when it does.

Six wickets and forty balls left over. They arrived together in this match and nobody minded, because a 38-ball innings sat between them. Next month, when the same side chases 199 in a knockout, someone will ask where the six wickets went. The answer will not be available then. It was available to be recorded now.