Half a Ball Wide: The Review Economy, Umpire's Call, and the Auction Window's Invisible Ledger

MCC Law 36, frame 47. On the ball-tracking screen the final projected point stopped just inside the outside edge of the stump, a hair inside half the ball's width. The on-field umpire had raised his finger. The question hanging in the third umpire's room was not whether the ball hit, but whether the contact sat inside the tolerance band. Two words appeared on the screen: Umpire's Call. The decision did not change.

It was November 2026 in Rajshahi. In the canteen of a coaching centre, a spreadsheet sat open next to a plastic cup of tea. That BPL season was the first to run DRS across the whole tournament, and the Rajshahi Kings campaign was swinging on three reviews in four matches. I decided that evening to stop offering verdicts and start keeping records. Forty-one reviews: ball-tracking frame, whether umpire's call applied, seconds between appeal and signal, and which umpire was overruled. By the final I had posted the table to a page I called Third Umpire Notes. Three thousand followers arrived.

The spreadsheet was my first whistle.

In this auction window everyone is talking about retention lists, release clauses and purse value. Fine, that is natural. But this piece is about a small, almost invisible column in that ledger: what a review costs, who burns it, and why cricket's most argued decision is not actually taken by technology. It is taken by a tolerance band written into a protocol.

Start with the law. MCC Law 36 defines when a batter is leg before wicket: the ball must have pitched, impact must be on the pad or person, the impact line must be parallel with the stumps, and the ball must have gone on to hit the stumps. All four are matters of the on-field umpire's eye. DRS added a fifth layer to that eye: measurement. But measurement carries its own error, and the ICC playing conditions gave that error a legal language.

DRS was first used in an international match in July 2026, in the Sri Lanka-India Test at Colombo. The technology stack then was three parts: Hawk-Eye ball tracking, Snicko/UltraEdge for sound, and Hot Spot for thermal contact. Hot Spot is effectively retired, UltraEdge now runs in real time, and ball tracking reads the trajectory at far higher frame rates. The hardware changed. The architecture of the protocol barely did.

Three thresholds matter. First, pitching and impact line: if the ball pitches outside leg, or impact is outside the line, it is simply not out, with no umpire's call available. Second, distance of impact: if the point of contact with the pad is more than three metres from the stumps, the on-field decision stands. Third, the most argued of all: projection. If the ball would hit the stumps, but by less than half the ball's width inside the line, it is umpire's call.

That third one is the story.

The review budget matters too. Tests allow two reviews per innings, topped up after 80 overs. ODIs and T20Is allow two. A successful review is retained; a failed one is lost. A review is therefore not a renewable resource. It is a card that runs out. A captain who does not understand that is gambling with a finite asset without knowing the odds.

Franchise cricket adopted DRS fully from the 2026 IPL, and the 2026 BPL had already run it across an entire season. Bangladesh's domestic T20 league was not far behind the international circuit in absorbing the technology. Yet discussion of review data here was close to zero. Runs went on the scoreboard, bad luck went into commentary, and frame numbers went nowhere. Nine years later the gap has not closed; it has widened, because there is more talk about reviews and less about the data underneath them.

June 2026. In a hostel room, one semester behind schedule, I watched the World Cup with a stopwatch running. That was football's first VAR World Cup: 29 penalties and 20 on-field reviews across 64 matches. Over 31 nights in four-hour sleep blocks I logged incident type, review duration and final call. Within 48 hours of the final a Bangla breakdown ran on a Dhaka sports site, drew 60,000 reads, and turned into a small monthly column.

I logged the World Cup from a hostel room, one frame at a time, and that became my rule. Protocol heading first: what the law says, what the official did, what the technology saw. Verdict after. Never the other way round.

Now the actual work. An LBW review is not one decision. It is a staircase of four separate decisions, and each step carries its own error bar. Anyone who reduces a review to 'wrong or right' has collapsed four steps into a single sentence, and that sentence is the great simplification of cricket commentary.

Step one: pitching. Where the ball hit the surface. The tracking cameras follow the trajectory, but the exact point of contact with the ground does not always land on a captured frame. When the bounce falls between two frames, the system interpolates. Interpolation is an estimate. An estimate is a probability. A probability is never certain.

Step two: impact. Whether the ball hit pad or bat, and when. UltraEdge catches the sound spike, but the audio spike and the visual contact do not always land on the same frame. When bat and pad are close together, the sound of the bat can mask the sound of the pad. This is where the famous 'bat first' arguments are born, with two sides telling two stories from one fragment of a spike, and both stories looking credible.

Step three: impact line. Whether the point of contact sits in line with the stumps. This is geometry, comparatively simple, and consequently the least disputed step.

Step four: projection. Extending the trajectory to the stumps to see whether the ball would hit, and by how much. This step is less geometry than probability.

In my 2026 ledger, projection decisions were the largest single category among the 41 reviews, and they produced the highest share of umpire's call outcomes. Batters were almost certain about pitching and impact. On projection their arithmetic went soft. That is natural: you can see pitching and impact. You cannot see a projection. You are challenging an estimate, not an image.

Which leads to the first working conclusion: a review's success depends not on the appellant's certainty, but on his confidence in the one step he cannot see.

Now duration. Through 2026 I timed every review from appeal to the third umpire's signal. A pattern kept returning. Reviews that ended in an overturned decision had a longer median duration. Simple reviews, a big no-ball or a clean catch, resolve quickly. Complex reviews, especially those with occlusion in the ball tracking or a conflict between tracking and UltraEdge, run long.

But duration and accuracy are not a straight line. Many long reviews ended in umpire's call with the decision unchanged, and some very fast reviews were proven wrong. Duration measures the complexity of a decision, not its quality. Conflating the two is the most common habit in commentary, and it produces lines like 'the umpire is hesitating'.

In 2026 the stadiums emptied. The Bangabandhu BPL final, Rajshahi Royals lifting the trophy in January, was the last live sport I saw, and by May my freelance income was close to zero. I wrote a plan in one night: 90 days, three blocks a day, no exceptions. From behind-closed-doors Bundesliga and K-League audio I built a library of 300 clips of whistle tones, crowd gaps and referee-mic bleed. By August I was volunteering with the Bangladesh Football Federation's referees' committee, logging decisions in resumed domestic leagues for free.

A 90-day plan is just a referee. It measures time. It does not measure emotion.

An empty stadium taught me that silence has a data trail. With no crowd, you can hear how many seconds passed from appeal to whistle, and from whistle to signal. That gap in sound is most audible in cricket around umpire's call, because the whole ground is holding its breath at once, and the length of that silence is itself a measure of how hard the decision was.

Now the arithmetic in money terms. The expected value of a review needs three inputs: the probability of an overturn, the runs or wicket gained if it is overturned, and the cost of the review lost if it is not. In T20, a wicket is worth roughly five to seven percent of a par innings score, shifting with the phase of the innings. The second review is worth more than the first, because losing it removes the option entirely for the rest of the innings. That is the foundation of review burn rate.

Which is why the most expensive asset in a T20 innings is not a yorker at the death, but a wasted review in the tenth over.

And here the auction window connects. Every transfer rumour needs a timestamp and a source. Every franchise should look at one column before retention: how many reviews a player burns per innings, and how many of those succeed. From the reviews I have logged, batters split clearly. Some will not review unless they are near certain. Some review on the emotion of the appeal. Likewise, some bowlers push the captain's shoulder on every LBW shout, and some read their own trajectory and stay quiet.

Nobody reads that column in a retention meeting. They read strike rate, economy, match-ups, powerplay splits. Yet a match is often decided by two badly used reviews. That gap is the biggest inefficiency I see in the auction.

There is another thing. Franchise data departments are absorbing real money, but most of it goes into ball tracking and match-up models. Very few teams work on review discipline. That data is the cheapest available, because the source is public: the ICC and franchise broadcasters keep every review clip, and often post them publicly. The table is free. Nobody is building it.

In esports, the replay is the referee. In cricket the replay is not a referee. It is a witness, and a witness's testimony is meaningless without a protocol.

Now the argument this piece is really about. The popular view is that umpire's call is a weakness. If the technology shows the ball hitting the stumps, it should be out. Many commentators, many fans, many former players make this case, and it sounds simple and fair.

I see it differently. Umpire's call is not a compromise. It is an admission: ball tracking carries an error bar, and inside that error bar the human eye is being given more weight than the machine. Consider the sources of that uncertainty: camera count, frame rate, seam movement, partial occlusion. Several millimetres survive all of them. Treat that uncertainty as zero and you are passing off a measurement as truth when it contains a margin for error. Umpire's call acknowledges that margin. That is not defeat. That is honesty.

Second argument: reviews do not reduce controversy, they relocate it. The argument used to be about the umpire's eye, where he stood, how much his view was blocked. Now the argument is about the tolerance band, why half a ball and not a third. The controversy has not shrunk. It has moved address. Technology does not remove doubt. It changes the language of doubt. Anyone who thinks technology ends the argument is standing one step into the process, not at the end of it.

Third, and most uncomfortable: the broadcast economy has an interest in review controversy. A disputed umpire's call produces three minutes of slow motion, eight angles, a graphic and eight panel discussions the next day. That content has value, and that value is now built into rights deals. Nobody wants the argument to shrink. The same broadcaster keeps the frame-level data of every review in its archive and never publishes it. The information exists. The transparency does not. That is the crack between the rights bubble and audience trust.

Fourth, the auction argument. Franchises spend heavily on the word 'clutch'. But what is clutch? A six in the last over? Or a review burned pointlessly in the tenth? My logs suggest the second matters as much as the first. Nobody pays for the second, because the second does not show up on a scorecard.

There is an uncomfortable truth here too. The biggest enemy of review discipline is emotion, and the biggest source of emotion is the captain himself. Under pressure a captain has two options: listen to his own doubt, or listen to the noise. The noise is always loud. The doubt is always quiet. Those who listen to the doubt convert more reviews, and their teams burn fewer across a long season.

One more point nobody writes. If you measure a review only by overturn or no overturn, you lose half the accounting. Many reviews fail and are still valuable, because they show the team where this bowler's trajectory is going on this pitch and how much the ball is stopping. That information feeds the field for the next over. A review is not only a decision tool. It is a data collection tool. Teams that understand this get two jobs out of one review.

Looking forward, two trends are clear. First, automation. Just as semi-automated offside took the line decision away from humans in football, no-ball and ball-tracking calls are steadily leaving the umpire's hands in cricket. Smart balls, automated ball-stump contact, even bat and ball sensors are no longer experiments. They are plans. The question is no longer when, but who actually decides.

Second, transparency. If a tolerance band determines outcomes, the data behind the band belongs in public. Publish a review audit after every series: how many reviews, how many ended in umpire's call, which step carried the most uncertainty, and how often each umpire was overturned. Players who know this will change how they appeal. Fans who know it will change the language of their argument.

The question is not whether technology is better than an umpire. The question is whether, having handed technology the power to decide, you are willing to publish the limits of its accuracy. If you are not, the half-ball margin will stay exactly where it has stood for nine years, whispering through every disputed dismissal: the uncertainty never left. It only changed address.