The Third Umpire's Ninety Seconds: The Gap Dhaka's DRS Log Never Shows

The entry in my notebook for that evening reads: Match 14, second innings, 17.3 overs, a catch behind the wicket, the on-field umpire's verdict out. The decision came down from the third umpire's room after two minutes and fifty-one seconds. Six replays, three angles. The one moment the whole decision hinged on — whether the ball grazed the edge — arrived from a fifty-frame-per-second recorder, with the word ultra-motion glowing in the corner of the screen. Twelve thousand people were drumming their hands against their chests, and none of them knew that two other cameras had captured the same ball on angles the broadcast never showed.

Between the roar of the ground and the silence of the third umpire's room sits an invisible wall. On one side lives emotion, on the other protocol. I have been taking notes from both sides of that wall for about six years — sometimes from the stands, sometimes at two in the morning, rewinding a broadcast feed on a laptop. The question was never whether the third umpire got it wrong. The question was who builds the information he rules on, who trims it, and who answers for the gap.

I started the VAR-Log because Dhaka's newsrooms receive controversy fast and evidence slowly, and when evidence is late the argument settles on the third umpire's name instead of on the system. So for every referral I log three things: the timestamp when the check begins, the timestamp of the final decision, and the number of frames the broadcast actually showed. In years of doing this, those three columns have never lied to me.

The Decision Review System is not a single device; it is a chain. From its first Test use between Sri Lanka and India in July 2026, and its fuller rollout from 2026, it has rested on three layers — the on-field umpire's eye, ball-tracking technology, and the broadcast feed of audio-visual evidence. If any layer is thin, the final decision is only as strong as its weakest link. We forget this because the big screen shows us only the last layer: the third umpire's finger.

The ICC's minimum DRS requirements are strict: a defined number of cameras for ball-tracking, a high-speed camera behind the stumps, audio-based edge detection (UltraEdge/Snicko) for catches, and a fifteen-second timer for a player to request a review. That timer tightened around 2026 and changed the psychology of the game, because the call now has to be made fast, on emotion, often in the voice of the least data-aware player on the field.

In June 2026 the ICC made a major change, abolishing the soft signal for catches. Previously an on-field umpire offered a preliminary call, and if evidence was inconclusive that call stood. Under the new system the burden falls directly on the third umpire. At the same time Umpire's Call was retained — if ball-tracking shows the ball hitting the stumps but within the margin, the field decision stands. Together these two choices create a strange condition: the machine now has far more power, while its limits remain hidden inside the protocol.

In Bangladesh the weakest layer is the broadcast feed. The BPL adopted DRS in the late 2010s, but its camera budget never matched the big three formats. The feed viewers watch and the feed the third umpire rules on often come from the same camera set. That is where a quiet conflict is born: broadcasting wants a spectacle, justice wants data. The two are not always the same.

A DRS decision is not a single moment; it is an assembly of four variables — frame rate, audio calibration, ball-tracking geometry, and the review timer. Pointing at any one of them does not mean the decision was wrong; it means questioning a specific layer of evidence.

The first variable is frame rate. The third umpire's room typically receives fifty to three hundred frames per second, and the ultra-motion label often shows interpolated frames built from a fifty-frame recording. Interpolation means the machine inserts an estimated image between two real ones. A ball arriving at one hundred and forty kilometres per hour covers roughly eighty centimetres between two genuine frames. If the contact happened in that gap, what we see on screen is an estimate, not a measurement. My log has at least eleven cases where the slowest frame suggested bat touched ball while the actual frames showed a clear gap.

The second variable is camera geometry. Two or three cameras usually cover the blind arc behind the keeper, and the bat-pad-ball triangle is often captured only from that arc. But what looks most cinematic on broadcast is the side angle. The angle most vital to the decision is the least discussed; the angle that catches the eye is the most aired. In my log of caught-behind referrals, the heavily used side angle showed the decision in two dimensions in three of four incidents, while the genuinely three-dimensional information sat in the camera behind.

The third variable is audio — Snicko and UltraEdge. This technology reads the sound of contact, but the sound comes through the broadcast audio channel, which carries crowd claps, horns, sledging, the keeper's gloves, even a spectator shouting nearby. The sound is never clean; it is a cloud of probability, not a fixed point. My notebook has cases where Snicko showed a faint spike in a full stadium, yet the same ball's slow motion showed clear daylight between bat and ball. Which is true — the sound or the image? The protocol does not answer clearly, and that gap sits quietly in the third umpire's room.

The fourth variable is ball-tracking and the geometry of Umpire's Call. Ball-tracking extrapolates the path and makes a prediction near the stumps — that is a model, not a measurement, and a model carries an uncertainty margin. If the ball's centre grazes the inner edge, the model says it is hitting, but the protocol says this is inside the margin, so the field decision stands. A strange logic follows: when the machine is confident, the decision is the machine's; when the machine sits at its boundary, the decision is the human's. Uncertainty is handed down from the machine to the person.

I have logged more than two hundred third-umpire referrals across three BPL seasons, each with a timestamp. Average review time in my log was over a minute and a half, and roughly one in three decisions was overturned. What startled me more than the overturn rate was this: the referrals that took the longest were the caught-behind cases, where audio and frame rate work together. The length of time is itself information: the longer a decision takes, the weaker its evidence.

The second half of the fourth variable is the timer and team psychology. A fifteen-second timer means a captain decides almost blindly, often on the keeper's shout, often mid-over, in the most tired phase of an innings. In my log, reviews are frequently taken when the top order's attention is weakest. DRS is no longer only a tool of justice; it is a resource — hoarded by some, squandered by others. And because the timer is tied to the broadcast clock, the broadcaster holds indirect power: the rhythm of which replay airs when can shape a captain's call.

Change the format and the balance shifts. Tests offer more time and more reviews, so teams can think. T20 offers no time, high emotion, and reviews that become a gamble. Yet the protocol frame barely changes — the same timer, the same Umpire's Call margin, the same absence of a soft signal. That structural sameness is the real problem. In a format where the game turns over eight balls a second, ruling with a fifty-frame recorder means using a machine beyond its capacity.

This is where I follow a personal rule that has sped up my work: three decisive frames, one rule citation, then stop. The reason is simple. Online DRS arguments are often endless — one side says no contact, the other says contact, and neither stops, because neither holds the full feed. But when I pick exactly three frames and say this frame shows it, this frame does not, the argument becomes measurable. A lack of evidence cannot be filled with attitude; limiting the frame count lets us say where we know the truth and where we are guessing.

Here I reach an uncomfortable conclusion, and it is not in the third umpire's favour. When the crowd curses the third umpire by name, it points at the wrong target. He is the last person in the chain, asked to make the hardest call on the weakest information. In my log I have never found a case where a third umpire had clear evidence and ruled against it. What I found was inconclusive evidence, and responsibility for that inconclusiveness sits upstream — in protocol design and feed control.

Abolishing the soft signal in 2026 deepened that inconclusiveness. Previously, when evidence was grey, the on-field signal balanced things: the machine offered evidence, the human decided. Now the evidence is the machine's and the decision is the machine's room, yet the blame remains human. The machine has been given power without being brought under accountability — that asymmetry is the central flaw of modern DRS.

The accountability question grows harder because it is not always clear who owns the feed. Match officials use the host broadcaster's feed. The broadcaster selects viewer-friendly angles, and what reaches the officials is a subset of that same set. Nowhere is it written which angle was dropped, why, or who decided. Every time I searched for that list from Dhaka, I found a blank page.

Someone might call this excessive suspicion of data. My answer is that I hold the same suspicion about match statistics, and not without reason. A heatmap makes us think we understand a player's role, yet it is really a machine's imprint, often inversely related to that player's duty inside the system. The same applies to DRS frame rates, camera angles, and Snicko spikes — they are machine testimony, and machine testimony deserves cross-examination just like human testimony. When we trust a machine without reason, we are accepting a model as truth.

So who is responsible? Three institutions own the three layers. The ICC writes the playing conditions — the timer, the margin, the existence of the soft signal. The match referee files a post-match report evaluating officials' decisions, but that report never becomes public. The host broadcaster builds and trims the feed, the sole raw material of the decision. None of the three fully owns the blame, and that gap is what I record in my log as the accountability gap.

The cheapest way to close it is to publish the third umpire's full feed set — every angle, every frame rate — shortly after a match ends. No viewer will watch all of it, but its existence makes evasion hard. From my log I can say with confidence: referrals where I could see multiple angles ended their argument on the field; referrals where I saw one angle stayed trending for three days.

The next step belongs to the ICC Cricket Committee. It needs a simple ruling: set a minimum frame-rate standard and deny DRS permission to any tournament below it. This will be hard for events like the BPL, and that is true. But the question is whether we want partial technology under the name of DRS, or a game without full evidence. The trouble with partial technology is that it inspires confidence exactly where it does not deserve it.

What I want is not romantic but dry and arguable: a public log for every referral — the check's start timestamp, the number of angles used, the frame rate, the reason for the decision, and the name of the decision-maker. That is not mere process; it is a document of accountability. In any season where such a log exists, one argument stops: was the decision wrong, or was the information incomplete? Separate those two and the third umpire stops being hero or villain — he becomes a man doing a job inside a system whose limits the protocol sets and whose blame institutions carry.

That entry from Match 14 is still in my notebook, with a small question mark beside it. Whether the catch carried, I do not know and never will — because three of the five angles are not in my hands. But from the frames I did see, I can say this for certain: the decision was slow, the evidence incomplete, and the blame belonged to no one. If those two hidden angles surface on a broadcast even once next BPL season, perhaps I can close this log. Until then my notebook stays open, and beside every third-umpire referral I will write three questions — how many seconds the check ran, how many angles were used, and whose room that feed stayed in. In cricket's history, umpires have always been under pressure, but that pressure once came from the stands. Now it comes from the silence of a room where two images beyond the six replays went unseen — and perhaps deliberately so.