The Immutable Ledger: Cricket's Review Data and Its Unwritten Market

In a T20 last season I sat with a stopwatch. The on-field umpire raised his finger for a review; the decision came back on the big screen after 3 minutes 42 seconds. In those 222 seconds I counted 11 replays, heard the crowd roar twice, and watched how far the third umpire's shoulder moved. Someone whispered next to me that the umpire had got it right.

Maybe he had. But something else was happening in those 222 seconds that nobody talks about. While the ball-tracking, UltraEdge and Real-time Snicko feeds were being loaded into the vendor's server, a data packet left the ground — through the broadcast compound and on to a data centre. Who saw it, how many seconds before the rest of us, at what price — that ledger is not in my hands. The question is not whether the umpire was wrong. The question is whose property the record of his error happens to be.

I stopped trusting my eye a long time ago and built a calibration protocol instead, because an eye holds memories, not time. In this piece I take the machine apart and follow every wire to where it connects.

What the ledger says: 2026 to now

On 23 July 2026 in Colombo, the player review system was used in a Test for the first time, in the India–Sri Lanka series. The vocabulary then was review — the right to ask for help. Over the following years the vocabulary changed to challenge. A review is a polite request; a challenge is a limited resource. That single shift tells the whole story: a right of verification slowly became a tactical asset, saved for the last over.

In 2026 the soft signal arrived. For low catches, the on-field umpire would raise his finger gingerly and send his opinion upstairs. Seven years later, in 2026, the ICC abolished it. The technical logic was clean: with two layers of opinion in place, there is a place for responsibility to hide. Nobody answers, because the responsibility hangs between two people.

Ball-tracking sits inside all of this. A system like Hawk-Eye is built by a vendor, bought by the broadcaster, and operated by the match official. The company televising the match is the company paying for the machine through which the decision is checked. The third umpire does not own the technology; he reads its output. The entire power structure is hidden in that one line.

In the 2026-25 season the Kookaburra Smart Ball entered a home Test in Australia. A chip inside the ball measures bat contact and spin. The device now lives inside the ball itself. There is no longer any distance between the data and the toy.

One ball, two kinds of truth

Open the machine. For LBW, ball-tracking first plots the trajectory, then predicts the path towards the stumps. Where the prediction is certain, the decision changes. Where it is not, the original stands — the so-called umpire's call. In the language of the laws, an umpire's decision is overturned only when it is clearly wrong. In the language of technology, that clarity is defined as a percentage: if roughly half the ball is predicted inside the stump zone, it stays umpire's call.

Something strange happens here. A decision made by a human is not replaced; it is parked inside an algebraic grey zone. A moral instruction from the laws — clearly wrong — is converted into a vendor-defined confidence interval. Nobody asks why that interval is exactly 50 per cent.

Second component: the three-metre rule. If the point of impact on the pad is more than three metres from the stumps, the ball-tracking prediction is treated as unreliable and is not used. Which raises a question: if the same camera cannot trust its own prediction at three metres, where does that confidence come from at 2.9? The answer is structural, not technical. Confidence is a function of a human-set threshold, not of truth.

Third component: audio. In 2026, when the Bangladesh Premier League was abandoned and the stadiums emptied, I started isolating whistles and shouts from closed-door footage. That work taught me a sentence I still repeat before opening any file: when the stadium empties, the microphones start testifying. In cricket that testimony arrives along the edge of the bat, as an UltraEdge spike. A sound wave, within a millisecond, saves a dismissal or sinks it. Sound is now evidence.

Fourth component, the one nobody writes about: where the feed goes. A review is a rare moment when unknown-but-soon-to-be-known information is created inside the game. That 20-to-40-second window suspends the live market and reopens it at a new price. Whoever holds the frames first holds the time advantage. And the frames come off a vendor's rack — a single room inside the ground.

Last season I tagged review sequences myself — a few hundred. For each I logged three times: the moment the third umpire saw the first frame, the moment the decision appeared on screen, and the moment the vendor's data flag became visible on the system. The gap between the first two is dramatic — that is for television. The gap between the second and third is near zero — that is for nobody, and for everybody.

Write the boring hypothesis first

I force myself into one discipline: before writing the surprising theory, write the dull one. The dull hypothesis is that umpires have not got worse. Errors did not increase; recording did.

Now let the data try to beat it. In my own log there is no steep decline in on-field standards. But the centre of the argument has moved. Five years ago the argument was about why an umpire erred. Today it is about why the process is opaque. The problem is not the sport; it is the administration. Here the dull hypothesis loses, because while the error rate held steady, the publication, explanation and commercial use of those errors changed.

The real blind spot: we do not audit the technology, we audit the man. Who catches a spoofed output on the tracking interface? Who detects a frame-sync failure? Who audits the vendor's confidence interval? No player, no spectator, no journalist has the right to enter that room. I have worked five different roles in this industry, and I still admit I have not sat in the room where these thresholds are written. Better to name what the ledger cannot see before someone else names it for me.

One more thing about myself. Surviving in this industry requires keeping doors open, and open doors make it hard to write against the people holding them. My solution: access is not allegiance. The room that is generous to me is not my subject; what gets decided there is.

A hash-chained ledger

Now the proposal. My nightly ledger — the one I filed at 4 a.m. through 64 matches of the 2026 World Cup — was a paper ledger. If I found an error the next morning, I wrote the correction in my own hand and never erased the original. Digital systems need that principle, enforced harder.

Every review sequence could carry an immutable record: which frames were used, their timestamps, the confidence values in the vendor's output, how many seconds the decision took, and an access log of who viewed the feed and when. Each entry would carry the hash of the previous one, so altering an old entry would make the chain contradict itself. Putting a match fixture on a blockchain is not the point; this is documentation infrastructure. Budget will not answer the question. Intent will.

This does not fix the market. The asymmetry that anyone with money can buy the data is not corrected by a ledger; the silent distance between that room and the ground remains. But one thing becomes certain: nobody can unilaterally erase the record of a decision. And cricket's crisis of confidence is now a crisis of records.

Name what this ledger cannot see. Who writes the hash, and on whose server the root of the chain sits — neither question has an answer yet. No independent source supplies frames; the frames come from the same vendor at the centre of the dispute. The root of the chain would have to sit with someone outside both the broadcaster and the ICC, or the difference between an audit and an audit will exist only on paper.

I still do not trust my eye. I trust my ledger, because the nightly ledger did not lie; it just waited for me to read it again. Cricket's review data is waiting in the same way. The question now is small and sharp: when the immutable ledger is first published, will the board publish it, or will the vendor?