The Pendulum of Umpire's Call: Where Ball-Tracking Measures and the Law Does Not

Umpire's call is not a rule. It is a process whose only function is to transfer the uncertainty of the technology onto the shoulders of the on-field umpire. Two decades of reading cricket's laws brought me to that conclusion, but the real evidence came from my own notebook: the deliveries that cause the loudest storms are rarely the closest calls. They are the calls where what the technology measured and what the law asked are not the same thing.

One image returns again and again from the reviews I have tracked this season. A left-arm spinner pitches the ball outside leg stump; the batter is struck on the pad roughly two metres from the crease; the projection says the ball would have grazed the outer edge of leg stump. On screen, the centre of the ball is outside the stump and a sliver of it is inside. If the umpire said out, the review fails. If the umpire said not out, the review succeeds. Same ball. Same graphic. Two verdicts.

DRS is not a second chance here. It is a second guess — and the decision goes to whoever guessed first.

History matters. The first player review in international cricket arrived in July 2026, in the India-Sri Lanka Test at Colombo: experimental, restricted to selected matches, resisted for years by several boards. The 2026 World Cup used UDRS, and from there it crept into every format.

Then the protocol itself: three questions. Where did the ball pitch? If outside leg, no tracking is needed and the answer is not out. Where was the impact, and was the batter playing a shot? And would the ball have hit the stumps? If any of the three carries uncertainty, the advantage flows back to the on-field call. That is umpire's call.

Two sub-clauses have accreted over time. The wicket zone: even when the projection says the ball hits, if most of the ball is outside the stump line, the on-field call stands. And the three-metre clause: if the point of impact is three metres or more from the stumps, the on-field decision survives regardless of what the tracking shows. Midway through 2026, the ICC abolished the soft signal, removing the on-field umpire's first instinct as the structural basis for catch decisions.

Translation is the real fault line. Ball-tracking infers a path from a series of video frames, then forecasts the post-bounce segment with a model. The final metres are projection, not measurement. The technology provider's own published figure puts mean tracking error on a millimetre scale — small against the width of a stump, but not zero. And the bounce itself is the least reliable moment: the ball changes shape, hits the seam, kicks up dust, and whatever happens between two camera frames has to be estimated.

The law does not know about any of that. The law asks a binary question: hitting, or not hitting. The technology answers on a continuum: 48 per cent, 53 per cent, 61 per cent, plus a confidence interval. Umpire's call sits between the two languages as the interpreter.

And that is where the definition of fairness shifts. Fairness stops being a property of the ball and becomes a property of the sequence.

Reviews are a scarce resource, which tightens the knot. A side gets two per Test innings, replenished after 80 overs; in the shorter formats the number is smaller. The pattern I see most often in my spreadsheet is waste: a review burnt after a boundary, just before a break, to break a partnership — spent precisely when the temperature is highest. The ICC added a 15-second timer to curb it. The instrument worked, but it treats a symptom. The cause remains: a team does not know whether its review will succeed, and that darkness is where strategy lives.

The three-metre clause is intellectually honest. If a ball spins and bends over a long flight, projection beyond the impact point becomes less dependable, so the field call is treated as the baseline. The consequence, though, is asymmetric. In bowling cultures where the ball turns most, the sharpest deliveries are trusted least — precisely the ones that use a long crease to trap a batter.

I was born in Dhaka and I work in Liverpool. That distance shows me something nobody sitting inside the ground sees: common sense is not a neutral term. What counts as natural in an English committee room, built on a large slice of Test history, does not always travel. So when someone says the ball was obviously hitting, that is not an argument; it is a cultural assumption wearing a badge. A large share of subcontinental fans believe their teams' reviews are scrutinised more harshly. I cannot prove a conspiracy with numbers. I can show the architecture. No conspiracy; asymmetry.

Abolishing the soft signal for catches in 2026 was a significant admission: the on-field umpire's first instinct should not be the foundation of a decision. The ICC accepted that — for catches. For LBW, that admission has still not been made. Same technology, same error, two philosophies. The topic has reached the ICC Cricket Committee table, where former players such as Anil Kumble have sat, at least once. The clause survives for now.

The popular proposal is simple: scrap umpire's call and let ball-tracking decide outright. I disagree — not because the consensus is wrong, but because the premise is.

As long as ball-tracking means a frame-based projection model, removing umpire's call will not erase the error; it will relocate it to the very edge of the stump.

Picture it: the 48 per cent ball that survives as not out today becomes out tomorrow. Someone will call that transparency. What arrives is a new convention with no route of appeal. Right now, at least a human error leaves room for an apology. When a model errs, who carries responsibility? The provider says it supplied the calculation. The committee says it followed the law. What remains is a silent number frozen on a screen.

At the 2026 World Cup in Qatar I spent three days working through semi-automated offside data, and the 1.88mm call taught me something. Millimetre-based officiating never becomes infallible; it manufactures the experience of infallibility. If cricket walks that road with ball-tracking, it must first learn to publish its own uncertainty — otherwise fans will get precision, not justice.

So the fault is in the premise. The premise is that ball-tracking is a measurement. The error was not in the pixel; it was in that assumption. If someone shows me the projection error holds consistently below a millimetre and frame interpolation plays no part in locating the bounce, I will change my position — and I am writing the condition down in advance so there is no room for excuses later.

What should happen next: publish the projection data, the confidence interval and the bounce-point clarity alongside the match centre for every review. Change the vocabulary, because umpire's call describes a lottery, and that concealment is the real instrument. Use a plain term such as projection uncertainty that blames nobody and admits the process's limits. If the law is going to rest on a model, the model's uncertainty should be part of the verdict, not a hidden clause. And give two separate answers in every decision: would it have hit, and how confidently. Run the numbers together and most of the calls that generate outrage turn out not to be uncertain at all. They are marginal. Margins can be measured; they cannot be asserted.

Which leaves the question: if the technology can return two different verdicts on the same delivery, and the only tie-breaker is who spoke first, is it umpiring the ball — or the umpire?

I trace the call backwards until I find an assumption wearing a badge.