The 3.6 Centimetres of Umpire's Call: Seven Layers of a DRS Audit

The 3.6 Centimetres of Umpire's Call: Seven Layers of a DRS Audit November 6, 2026. Arun Jaitley Stadium, Delhi. Bangladesh against Sri Lanka in a World Cup league fixture. Angelo Mathews walks to the crease and, at exactly that moment, the strap of his helmet snaps. A new helmet arrives, the strap is fitted, he takes guard again — and the two-minute limit has expired. Shakib Al Hasan appeals. The on-field umpire raises his finger. International cricket records its first timed-out dismissal. That evening I did not watch the match. I audited it. I ran eight broadcast camera angles frame by frame, cross-checked them against the scoreboard clock's time stamps, and wrote one question in my notebook: is this decision hateful, or is it wrong? The answer is uncomfortable. By protocol, the decision was almost flawless. Yet what the cricket world argued about for the next three weeks was not a law at all — it was its own feelings. Timed out, umpire's call, soft signal, overthrow. At the centre of every one of these controversies sits the same question: are we following the rules, or are we in love with them? I do not watch cricket. I audit the moments where cricket watches itself. The DRS entered international cricket in 2026, through a series between India and Sri Lanka. In its early years it was limited and experimental, and the Indian board resisted its most contested component for a long time — ball-tracking. The objection was not unreasonable. Ball-tracking is a projection of the future, and overturning an umpire's decision with a projection means placing an algorithm on top of a human eye. Sixteen years later, the picture is different. Tests, ODIs, T20Is — two reviews per innings everywhere, retained on success, burned on failure. Bowling in the thirteenth over, the final afternoon of a fourth day, a knockout at a World Cup: every setting moves through the same sequence. I arrange that sequence into seven layers: the on-field call, the player's fifteen seconds, the technology chain, the threshold calculation, the third umpire's interpretation, the match context, and finally the outcome audit. Questioning the order of those layers is my job, because a review in cricket is not a judicial system. It is a finite resource that must be spent inside fifteen seconds. This is where the first contradiction hides. A batter standing in the middle has no ball-tracking. He sees where the ball pitched, where it struck his pad, and reviews on that basis. Nobody whispers into the captain's earpiece, 'the ball is hitting less than half the stump'. So the review decision is almost always a reflex, not an analysis. Layer One: the on-field call is the foundation Most viewers believe the DRS is an independent trial, as if the third umpire decides from zero. The truth is the opposite. Before the review even begins, the on-field umpire's decision carries a weight. In LBW it creates a band of opinion; for catches it was, for years, the soft signal. So the real question is not out or not out. It is: what did the umpire say, and is there enough evidence to overturn it? That is a design principle with teeth. It means different umpires carry different bands. An umpire who raises his finger the moment ball touches pad needs very little evidence behind his 'out'; a generous umpire's 'not out' can still be overturned when ball-tracking shows the stumps hit. Same ball, same track, two umpires — two different outcomes. That weight later becomes the most important evidence in any control case. Layer Two: the strategy of fifteen seconds The law says a review must be requested within fifteen seconds of the ball becoming dead. In practice those fifteen seconds are a political animal. An opener has just been given out and is walking towards the pavilion; the captain is watching the last replay on a dressing-room screen. Some review out of conviction, some out of fear, some simply because the team's biggest batter is waiting at the other end. In my own stored dataset, a large share of LBW reviews across the last five IPL seasons returned to the on-field decision. The reason is not complicated. Teams were reviewing based on what happened at pitch level, without knowing where the ball would end. A batter does not know how much the ball is turning in; against a leg-spinner, that inward angle is almost impossible to measure from the batting end. Strategically, the best reviews therefore come against pace, where the track is comparatively simple. The worst come against spin, where a batter gambles on a trajectory arriving at over a hundred kilometres an hour. The retention rule itself is a behavioural test. A successful review survives, a failed one burns. Which means the side that shouts loudest runs out of chances, and the side that maintains information discipline survives longest. The law is not a fairness mechanism. It is an asset-management system. Layer Three: the weak link in the technology chain Here I made a mistake for years — I treated ball-tracking as final truth. That frame collapsed for me in 2026, at a domestic derby, when I reviewed a disputed 89th-minute penalty across fourteen camera angles and realised every angle tells its own story. In cricket the chain is longer: UltraEdge, real-time Snicko, HotSpot, ball-tracking, and time synchronisation. Synchronisation is the most neglected layer of all. If there is a slight offset between the audio window of an UltraEdge spike and the video frame, a spike belonging to the bat can be displayed beside a ball that struck the pad. If any single link in that chain is weak, all seven layers go wrong, because the output is never loyal to the last link — it is always loyal to the first faulty one. The bigger crisis, though, is not in the chain but in how the chain is marketed. On broadcast we watch the third umpire staring intently at a green screen, but the viewer never sees how confident the software is in its own answer. Here I keep a lesson from my football work: across 83 behind-closed-doors matches in 2026, the home team's penalty rate fell. The whistle did not change; only the noise did. I am not going to assume home-crowd pressure works the same way on cricket's video umpires, but that pressure is never publicly measured. A protocol that hides its own uncertainty will always find it easy to defend itself. Layer Four: 3.6 centimetres This is the centre of the storm. When ball-tracking shows the ball striking the stumps, but the amount of impact is less than half the ball's width, the decision becomes umpire's call. A cricket ball is roughly 7.2 centimetres wide; half of that lands at about 3.6 centimetres. People love to read those 3.6 centimetres as an insult: the ball hit the stumps, so why is it not out? This is the day's biggest misunderstanding. Umpire's call is not mercy, and it is not a doubt-goes-to-the-batter doctrine. It is an admission of error. The tracking system has its own uncertainty — pitch surface, seam position, revolutions, wind speed, camera frame rate. If that uncertainty band is wide enough that the ball's contact with the stumps can be described differently, then overturning means replacing one uncertainty with another. The protocol says: where the uncertainty itself is in doubt, the on-field decision stands. This is also the commercial version of the argument India's board made for years. I initially sided with the principle, then realised two very different positions exist — 'the technology has error' and 'therefore the technology should not be used'. The first is an engineering statement. The second is a political one. I am with the first, provided the protocol publicly displays the size of its own error on screen. Layer Five: the third umpire and the death of the soft signal For years the soft signal on catches was the most frustrating thing for viewers, because it tried to express two levels of decision at once. The on-field umpire, unwilling to take a risk, would say 'out', and the third umpire would preserve that out when evidence was insufficient. Cricket has found a slow remedy: the soft signal has been withdrawn, and verdicts now rest directly on evidence. That has not solved the problem, only displaced it. The interpretive pressure now falls personally on the third umpire — some want speed, some want every frame examined. Of the seven layers, this is the least transparent, because nobody on the field explains the why. Only an image floats up on a screen. Layer Six: match context This layer sits outside the protocol and yet is its loudest silent driver. A review in the twelfth over at 45 for 3 and a review in the fortieth over at 310 for 4 are identical on paper and nothing alike in reality. Under knockout pressure a batter's pad is four centimetres further forward, a fielder's elbow bends slightly differently, the wide line shifts a fraction. I wrote this down before the tournament: context pressure will not change the protocol, but it will change the application of the protocol, and that is the real danger. The DRS therefore stays at the centre of every cricket argument, because the law is one and the hands applying it are several. Layer Seven: the outcome audit This is where my work separates from other analysis. I audit the process and the outcome separately, and I record both — even when they contradict each other. Control case one: 14 July 2026, Lord's, the World Cup final, England against New Zealand. Off the final ball of the Super Over, the ball deflected off Ben Stokes' bat and ran to the boundary through third man. The batters had not crossed for the second run, yet six runs were awarded. By the law as it was written that evening, the decision was correct. Four months later, in October, the law was amended — no crossing, five runs. The protocol was right; the law was wrong. Control case two: Delhi, 6 November 2026. Mathews. The law correct, the application correct, the outcome unbearable. One lens, two completely different returns. One says, 'we followed the rule but the rule itself was flawed.' The other says, 'the rule was fine but we wanted an interpretation of it.' Now my counter-argument, and my own admission of failure. In 2026, at the Russia World Cup, I audited 22 review incidents across 64 matches and built a seven-layer structure that now looks over-confident to me. I had assumed a clean protocol run would produce a correct decision. With Mathews it did — and the audience could not accept it. My structure had no space for that disagreement. So I amended my own framework: separate boxes for the process audit and the outcome audit, plus a third — the 'public cost audit', meaning which decisions broke a spectator's heart, and why. I no longer treat crowd anger as noise. The clearest measure of a cricket decision's value sits in the hands of a boy in a small town, who may not have bought a twenty-taka ticket but carries the entire innings on his shoulders. The strongest counter-question comes from the other direction. If umpire's call is a product of uncertainty, then uncertainty should cut both ways. Right now it cuts one way only. The on-field umpire is protected; the batter is not. The same 3.6 centimetres rules against the batter but never against the bowler. When a bowler protests a wide, nobody says, 'that was only a three-centimetre error.' That asymmetry has barely been discussed. My record shows I have publicly mis-predicted review controversies twice — once in the IPL, once in a day-night Test — and I published both corrections myself. A framework's authority comes not from its accuracy but from the honesty of its admissions. So, three proposals going forward. First, the measurement of ball-tracking error — those 3.6 centimetres — should be displayed live on broadcast during every review, so the viewer knows exactly where the uncertainty sits. Second, rather than adding reviews, teams should receive a review budget: the resource would be allocated as a single pool for the entire innings rather than refunded on success, which would curb reckless reviews and reward information discipline. Third, and most importantly, a prediction with an expiry date. I am stating publicly that before the 2027 World Cup, the ICC will either reduce the umpire's call margin or rename the concept altogether, because frame rates and ball models are improving. I have written the date in my notebook. If I am wrong, I will show my own working — because a protocol that refuses to publish its errors is less trustworthy than any umpire. An umpire standing in the middle says in one breath what an analyst behind a screen can take seven layers to answer. The difference is only time. Whether time is the only message cricket has — that question still hangs, at a distance of 3.6 centimetres.