When the Frame Goes Missing, So Does the Verdict: Cricket's Evidence Chain and the Case for a Verifiable Decision Ledger

There is no mark on the Hot Spot monitor. The slip fielder at Trent Bridge is waiting, the wicketkeeper has gone up, the batsman is standing in his crease. July 2026, the first Ashes Test, England against Australia. Stuart Broad is on 37. He edges Ashton Agar, the ball travels towards slip — the eye saw it, the infrared camera did not. Australia's two reviews are already gone. The on-field umpire gives it not out, and under the review protocol that is the final decision. Broad goes on to make 65.

The question that moment created is not written on the scoreboard. It sits on the screen — can you deliver a verdict on a frame that does not exist?

Twelve years later, in the winter of 2026, my desk runs more than a dozen feeds: ultra-motion at 50 to 300 frames per second, Hawk-Eye ball tracking, UltraEdge audio waveforms, live ball speed. The failure mode has not changed. The systemic failure is no longer in the umpire's eye. It is in the habit of treating an empty field as a filled one. This piece is about that habit, and about its antidote.

Cricket's review system is not a gadget. It is a legal architecture. Since DRS arrived in the 2026 India-Sri Lanka series, what has been built is essentially a courtroom: who may appeal, what standard of proof applies, and whose side the doubt favours. Umpires do not merely apply the laws; they construct the interpretation of the laws.

The architecture has three layers, and they need separating. The first is fact detection: where the ball pitched, where it struck, whether the bat made contact. The second is prediction: the path the ball would have taken after bouncing. The third is threshold: whether that information amounts to a clear and obvious error.

If any of the three has a gap, the protocol says the on-field decision stands. The principle is simple. Its consequence is uncomfortable. Absence of evidence is not innocence; absence of evidence means the decision is unchanged.

After the 2026 Hot Spot crisis, cricket worked hard on these layers. Real-time Snickometer arrived, edge detection became more reliable, and in 2026 the ICC abolished the soft signal entirely — the on-field umpire's provisional call no longer anchors the third umpire. The 2026 MCC Code cut the timed-out window from three minutes to two. Those two changes look unrelated. They share one philosophy: reduce the load of assumption.

I learned that philosophy in football. In 2026, at 53, I joined the Premier League's VAR pilot at Stockley Park as one of four London-based analysts. I audited 38 matches, logged 47 checked incidents, recorded 12 overturns, and found three missed calls that later became training clips. After I built the seven-point review checklist — angle, offence, point of contact, speed, offside line, restart, communication — average review time fell from 84 seconds to 52. I briefed 20 referees at St George's Park, and the FA adopted the checklist for training.

That checklist got me hired by ITV as a remote VAR analyst for the 2026 Russia World Cup. The whole experience compresses into one line: “From Russia, with a remote desk and a five-second delay.” Across 64 matches I logged 29 penalties, 22 reviews and three overturned penalties. In the final I built a decision tree for Ivan Perisic's handball that explained the 3.5-minute review, and it later went into a 12,000-word tournament audit.

In 2026, during the pandemic hiatus, I joined the Premier League's behind-closed-doors task force. Analysing 92 empty-stadium matches, I found that the whistle-to-signal delay rose by 0.8 seconds without crowd noise. I wrote an audio-cue protocol — “check complete”, “on-field decision stands” — and tested a 12-page silence protocol across 14 matches. The league credited it with an 18 percent reduction in communication errors. That is when a stopwatch entered my writing. I began recording not only what happened but how long it took to decide. “Empty stadiums made the audio protocol louder than any crowd.”

UEFA hired me for Euro 2026 in 2026. Fifty-one matches, 18 penalties, five VAR overturns. I flagged the 2.1-minute penalty check in the Denmark-England semi-final separately. Then Tokyo, 32 matches, six VAR interventions. That is where the Referee's Eye column began — 24 weeks, 1.2 million readers. In it I built one habit: show the text of the law before offering the opinion.

In cricket, the same architecture is far more complex, because more kinds of decision arrive on the same screen — LBW, catches, stumpings, boundaries, timed out. Each has a different evidential basis. All of them end at one question: is the record complete?

What does the third umpire actually see? The viewer sees one screen. The third umpire sees six to eight layers, and every layer has its own failure mode.

Broadcast cameras typically run at 50 or 60 frames per second. The high-speed cameras used for reviews run at 100 to 300. The number matters, because frame rate is really a question of distance. At 150 kilometres per hour, the ball travels 41.7 metres per second. At 50 frames per second, two frames are 20 milliseconds apart — the ball advances 0.83 metres per frame. At 300 frames per second, the gap is 3.3 milliseconds and the ball advances 14 centimetres.

So the moment of contact is not a point. It is a window. At 50 frames that window is 83 centimetres wide; at 300 frames it is 14 centimetres. A review decided on a 50-frame feed can produce a different verdict on a 300-frame feed — same ball, same bat, same law, different frame rate.

That is why the oldest line in my notebook still holds: “The monitor does not lie; the angle does.” The monitor does not lie, because the monitor says nothing. Frame rate, camera position and interpolation assumptions combine to tell a story. The review umpire is giving evidence about that story.

One number about Hawk-Eye ball tracking should always be kept in mind: the company's own stated accuracy for ball position is 3.6 millimetres. But the tracking does not measure the whole path. It measures some distance past the bounce and predicts the rest mathematically. The further you get from the bounce point, the wider the prediction error.

This is where LBW's umpire's call comes in. A cricket ball's circumference is 22.4 to 22.9 centimetres, so its diameter is roughly 71 to 73 millimetres and its radius roughly 35.5 to 36.5 millimetres. The rule is that if the centre of the ball falls within half a ball's width of the outside of the stump, the decision belongs to the on-field umpire. That is not a close call. It is a declared uncertainty band of about 3.5 centimetres. What cricket has done here is honest: a measurement that does not know its own error margin does not overturn a verdict.

The second failure mode is subtler: treating the absence of evidence as evidence against. Hot Spot uses infrared cameras to detect friction heat. But protective tape or a sticker on the edge of the bat masks the thermal signature. If the ball brushes the back of the bat, the camera angle is no help either. Trent Bridge in 2026 was exactly this: no mark, therefore no contact — a conclusion that is logically invalid and procedurally correct.

UltraEdge and Snicko fall into the same trap. A spike on an audio waveform is not automatically the edge of a bat. Bat on pad, ball on helmet, footsteps beneath the camera, even noise from the stands can produce a spike. The 2026 empty-stadium data taught me this lesson: without a crowd, the delay from the umpire's whistle to the signal rose by 0.8 seconds. Change the acoustic environment and the same incident sounds different.

Against all of that stands one decision that is the cleanest review in modern cricket. 6 November 2026, Delhi. Bangladesh against Sri Lanka, World Cup. Angelo Mathews was walking out to bat when his helmet strap broke. He was not ready to play within the allotted time. On the Bangladesh captain's appeal, the umpires gave him timed out — the first in the history of international cricket.

That decision required no ball tracking, no edge detection, no frame-rate arithmetic. It required a clock and a law. The 2026 MCC Code had cut the limit from three minutes to two, and the limit is mechanical. A broken helmet strap is equipment failure, but the law allows no exemption for it. The decision carrying the least interpretive load is the one most robustly defensible. That is the central theorem of my 2026 checklist.

14 July 2026, Lord's. The overthrow in the World Cup final. The ball ran to the boundary, the runs were added, but the real question was different: had the batsmen crossed for the second run at the moment of the throw? That is not measurable. That is judgement. No camera can state a batsman's position and intention to the millimetre. That single decision changed the outcome of a World Cup final, and nobody is certain today whether the process was right — because nobody wrote the process down.

Another 2026 change supports the same theorem. The soft signal was the on-field umpire's provisional call, which the third umpire could only overturn on conclusive evidence. For catches this frequently produced the wrong result: the camera showed something clearly, yet a weak prior kept pulling the decision back. In Bayesian terms, a low-confidence prior drags the posterior towards itself. The ICC removed the soft signal and gave evidence priority.

A limit on evidence nonetheless survived, and keeping it was strategically intelligent. Umpire's call was not abolished. The reason is clear to me: close to the stump line, ball-tracking prediction is not accurate enough to justify disregarding the on-field eye. A system that knows its own limits is credible.

Now to the real gap, the one nobody wants to write about. Cricket announces review verdicts but does not publish review records. Which angle was used, at which frame rate, what proportion of the ball's path was measured and what proportion predicted, how long the decision took, what words the third umpire used — there is no standardised, machine-readable, publicly accessible log of any of it.

Football is the same. The seventh item on the checklist I built at Stockley Park was communication, because a decision that cannot be explained is a bigger problem than a decision that is wrong. I wrote a 12,000-word audit of Russia 2026 for precisely this reason: the match ended, the verdict was announced, and the process was recorded nowhere.

This is where my proposal sits, and it is where the underlying idea behind blockchain applies to cricket — a hash-chained, append-only, timestamped and verifiable decision ledger. A Review Record for every match: the minute of the incident, the type of review, the angles and sensors used, the frame rate used, the law cited, the decision, the duration, the confidence level. Each entry's hash is chained to the previous one, so nobody can later swap a frame out of the middle of the record.

This is not speculative. In 2026, when I standardised phrases such as “check complete” and “on-field decision stands”, fixing the words alone cut communication errors by 18 percent. Standardising language is the first step of a process. Standardising the log is the second. The real contribution of blockchain technology is not cryptocurrency — it is the idea that a record you can alter is no longer a record. In officiating, that is an ethical question rather than a technical one.

There is another dimension that keeps surprising a football man in cricket. The transfer market has offside lines too; you just cannot see them. In football a club spends a hundred million pounds on a young player because his data model shines, while dressing-room chemistry fits into no model at all. Cricket inverts this at selection: two innings from one series displace a ten-year career record. The problem is identical in both — what can be measured is believed with indecent enthusiasm.

I know who will dislike this piece. The crowd wants a verdict, and the crowd is never pleased by insufficient evidence. The commentary box is under the same pressure: every contentious moment needs a clear call, or the audience changes the channel.

But clear and obvious error is a threshold, not a feeling. A threshold can be measured; a feeling cannot. The argument is not really about the verdict. It is about the process — who decided, in how long, on what information.

There is an uncomfortable truth here. More technology has not reduced controversy; it has relocated it. The question used to be, did the umpire get it wrong? Now it is, what was the frame rate? Where did that Snicko spike come from? What percentage of the ball tracking was prediction? Controversy has moved from the field to the screen, and from the umpire to the analyst.

That second move exposes the blind spot in my own trade. Just as the sticker stopped Hot Spot from showing a mark in 2026, an analyst fills an empty field with his own inference. When a feed drops, a sentence begins with probably, and ten minutes later the probably has become a headline. When the evidence chain breaks, the biggest risk is not falsehood. The biggest risk is confidence.

One more thing belongs here, and almost nobody writes it. Year after year, the marginal calls in matches involving smaller teams go unaudited. The crowd and the sponsors watch the big fixtures. Yet those weaker sides lose or gain one or two reviews every series, and those two reviews often decide whether a home series is won or lost. Those small, uncomfortable calculations at the start of the year are the real cost.

Over the next five years I expect three changes, and all three are measurable. One, review records will be published, because otherwise the trust deficit widens. Two, sensor metadata — frame rate, angle, the measured-to-predicted ratio in ball tracking — will become a mandatory log field. Three, independent audit for match officials, testing not the correctness of the verdict but the integrity of the process.

In a cricket world that cannot reconstruct a decision from its own record, every controversy eventually becomes a contest of memory. Who saw what, who remembers what, who said it loudest.

So the question for me is no longer whether the verdict was right. The question is whether the verdict can prove itself from its own evidence.