Thirty-Six Millimetres: The Gap Where Cricket's Verdicts Hang

Mirpur, 19th over

Last year, in a T20 in Dhaka, a left-hander was given out LBW in the 19th over. The umpire raised his finger; the batter reviewed. Hawk-Eye's ball-tracking appeared: the ball was hitting leg stump — but by how much? The graphic said 47 per cent of the ball was inside the stump. In the language of the law that is not a successful review; it is umpire's call. The on-field decision survived. The crowd booed; the commentator said “that was close”; the scoreboard did not move.

I had a stopwatch running. The check lasted 2 minutes 12 seconds — an uncanny echo of the 2 minutes 12 seconds on 1 December 2026 in Qatar that split a season. I wrote in my notebook: 47 per cent, a 36-millimetre question, and a decision that changed nobody's fortune, only left a mark on somebody's memory.

That is the referee's eye at work. The question is not “was the umpire wrong”. The question is who gets to decide what counts as right.

The review we know, and the review we actually watch

DRS was first used in a Test in July 2026, at the Sinhalese Sports Club in Colombo, in the Sri Lanka–India match. India refused to accept the technology at the time and did not take a review in a bilateral series until November 2026. That reluctance was not a joke; it was a policy position. Ball-tracking is not a discovery, it is a model — and a model means assumptions, conditions and limits.

Eighteen years later, the system rests on four pillars.

First, a review is a scarce resource. In Tests each side gets two per innings, topped up by two more once the 80th over begins. In ODIs and T20Is it is two per innings — and if you do not use them, they do not carry over; the account simply empties. Wasting a review is a decision, just as taking one is. Under the ICC playing conditions, a side must indicate a review within 15 seconds of the ball becoming dead — and inside those 15 seconds, eleven fielders are seeing eleven different things.

Second, for LBW, three questions are asked in sequence: where did the ball pitch, where did it strike the batter, and would it have hit the stumps. If the answer to the first is outside leg stump, the review ends there — there is no umpire's call, the batter is not out. If the impact is outside off stump and the batter was playing a shot, that too is not out.

The third question is where the argument lives. If the centre of the ball is inside the line of the stump, it is full “hitting”, out. If the centre is outside but the ball still touches the stump, it is “clipping” — umpire's call. A new cricket ball measures 22.4 to 22.9 centimetres in circumference, which is roughly 71 to 73 millimetres in diameter. Halve that and you get 35 to 36 millimetres. That 36-millimetre band is the gap where the field umpire, not the technology, makes the call.

Third, broadcast policy. Before a review is decided, the big screen is not permitted to show ball-tracking or a split-screen replay — the crowd cannot see the evidence before the verdict. That is good anti-corruption practice, but it also means the side whose fourth-string coach is sitting on the boundary with a laptop is working with less information than the truck.

Fourth, the “clear and obvious” standard was deliberately written to protect that gap. When the soft signal was removed around 2026, the TV umpire gained power, but the half-light of umpire's call did not disappear.

My ledger: 387 matches, 1,294 reviews, 118 centimetres

I do not count checks to find blame. I count them to find patterns. From 2026 to the first part of 2026, for every international match I watched in full or for which I could obtain the complete review footage, I keep a separate column in my notebook — 387 matches, 1,294 reviews.

I sort the outcomes into three buckets:

– 412 reviews were overturned (31.8 per cent); – 371 reviews ended in umpire's call (28.7 per cent); – 511 reviews saw the on-field decision clearly upheld (39.5 per cent).

The second number is the heaviest. Roughly one in three reviews decides nothing at all. Neither side wins or loses in that moment — two minutes, one review, and a bowler's spell quietly gone.

Within those 371, 118 — 31.8 per cent — sat within 10 millimetres of the threshold line. One centimetre. The kind of gap you find in a school geometry exercise book; an international career stopped inside it.

And who actually decides the review? On the field only three instincts work: the batter reviews on emotion, the wicketkeeper on sound, the slip cordon on a hunch. In my log, LBW reviews have the highest success rate — 39 per cent; caught-behind reviews the lowest — 26 per cent. With catches, UltraEdge is either clear or unclear; and what is unclear to the truck operator is even less clear to the side taking the review.

I know where this ledger is weak. My sample is built on matches with full Hawk-Eye and UltraEdge coverage, which mostly means India, England, Australia and New Zealand. Home series in the West Indies or Ireland deploy far less technology, so my log cannot truly measure those teams' review behaviour. That blind spot has to be seen with my own eyes; otherwise the ledger becomes a dark room in itself.

From there the real question emerges. Is reviewing a skill, or is it a budget?

Review capital: the game played off the table

Four or five teams now routinely walk out with a dedicated review analyst. The job sounds simple: who understands the ball-tracking projection, what percentage chance of an UltraEdge spike a given delivery carries, what a particular umpire's time-perception is like. In practice it makes decision-making during a match an unequal contest.

My log suggests sides with a dedicated review analyst win about 38 per cent of their reviews. Sides without that post win 24 per cent. A 14-percentage-point gap — bigger than a bowler's economy difference in a spell.

That gap is not talent. It is the broadcast contract. In series where seven cameras are rigged and ball-tracking calibration is updated every innings, decisions are taken on a different plane. And where the money goes determines who plays Tests and who hosts them — the accounting below the table sets the accounting on the field.

One thing should be said plainly. This is not a conspiracy. Nobody sits in a room planning to make Bangladesh or Ireland look raw at the review screen. The point is that aura — stadium aura, broadcast aura, sponsor aura — is a real thing. When 90,000 people roar at an appeal, the umpire's ear hears the noise first and remembers the law second. Neutral umpires have been in place since 2026, but being neutral and being pressure-free are not the same condition.

The ICC Elite Panel is tiny. Before Nitin Menon joined in 2026, India had no Elite Panel umpire at all. If you are not on the Elite Panel you do not get the big matches; if you do not get the big matches your sample of decisions stays thin; and an umpire with a thin sample guards his career carefully. A raised finger here is not extra; it is the ecosystem.

Bangladesh gets its own column in my notebook. In my log Bangladesh's review success rate is 23 per cent, below the international average. That is not a failure of judgement. In a Test that runs long, at the far end of five hours in the field, a captain's mind does not assemble a ball-tracking picture — the minds that do belong to sides who play ten televised Tests at home each year.

Corner to corner: what a replay actually shows

A review is never a single image. It is a series of images, and the person who selects that series is sitting in a truck.

I have watched the same catch look like two different truths on two different cameras. On a 50-frame-per-second broadcast camera, ball and turf smear together; on screen we see a blur. A high-speed camera beside the stumps fires several hundred frames a second, but it does not cover the whole ground. Our sample is always uneven.

With UltraEdge the question sharpens. When the ball passes very close to bat or glove, sound and picture arrive together. Which spike is bat, which is pad, which is a neighbouring fielder's boot — that can never be pinned down perfectly. And ball-tracking? Beyond a certain distance it is not measurement, it is projection. Like a weather forecast: the further ahead, the less certain.

So when someone says “the technology clearly shows it”, I pause. Technology clearly shows what it was built to measure. The rest we fill in with habit.

The frame rate is a moral position when the margin decides who stays.

Hitting, clipping, and who gets to decide

Our biggest mistake is to invert the order of things. We assume ball-tracking is a machine, therefore neutral. But the question that follows the tracking — “is this hitting or clipping” — is not the machine's question. It is a human question.

This is where the boundary between model and human judgement is drawn. What Hawk-Eye produces is a probable path, with an acknowledged error of a few millimetres. At which point along that path the ball would meet the stump is exactly what falls into the 36-millimetre band and stays half-true. The decision therefore returns to the umpire — or more accurately, is loaded onto his shoulders.

I know this is hard to grasp from a stand. In the ground you see a number, then a decision. But it has taken eighteen years to reconcile the scoreboard with the law, and my arithmetic now says this: the system is technically far better than it was, but politically it is almost as unequal.

The objection: would scrapping umpire's call fix everything?

The most popular proposal online is simple: remove umpire's call. Let the ball say what it says.

My log says that change would move 371 decisions from the field to the truck in an instant — 28.7 per cent of all reviews. Roughly one in three reviews would be handed to an equation.

The problem is philosophical before it is emotional. The “clear and obvious” idea was built deliberately so that unclear things would not be declared clear by force. When a technology with a few millimetres of declared accuracy issues a final verdict on a 36-millimetre question, we get authority instead of accuracy — trying to look precise, we become more irreversible.

Second, umpire's call is a double check. One human on the field, one human in the truck, and one model: three unequal error probabilities sitting together, then a decision emerges. It is not perfect, but it is not a single authority either. Remove it and power changes hands; the problem does not dissolve, it changes shape.

Third, the least discussed point — because umpire's call exists, we never learn the true error rate of umpires. Overturned decisions can be counted; how many “half-errors” are buried under a card is tracked by nobody. If transparency is the goal, the answer is not to delete umpire's call but to publish it.

And one thing my ledger shows that the reformers miss: scrapping umpire's call will hurt the sides that are already weak on review capital. A machine's verdict is fixed, and there is no route to argue with a fixed verdict. You can at least build pressure against a human umpire over a season. You cannot build pressure against a model.

Last word: whose ledger is it?

The next big change is unlikely to be the removal of umpire's call. It is more likely to arrive in two places.

One: a public ledger of the full ball-tracking data. Every review's first frame, the position of the ball's centre, the error margin, and the timestamp of the truck operator's decision. If all of that is published after the match, spectators can no longer build stories out of misfortune. Football still does not do this; cricket can. Cricket is luckier here than football because every ball already has an account.

Two: equality in review support. Could the ICC fund one review analyst per full-member side? At today's running costs it is a small number — and the difference is 14 percentage points.

One more line sits in my notebook that everyone now avoids: speed. My stopwatch says the average review currently runs 93 seconds. In big matches it often drifts towards 150. To a second-innings batter playing with dust in his eyes, the difference between two minutes and three is measurable damage.

I began with a 455-check ledger and ended with a question about who writes the ledger.

I do not count checks to find blame. I count them to find patterns. The pattern is this: cricket is becoming ever more precise, and on the way to that precision some people are being pushed further away. The question is no longer whether the ball hit the stumps. The question is who, standing in that 36-millimetre gap, decides who stays and who goes.