From Ball-Tracking to Blockchain: Who Keeps Cricket's Decision Archive — and Who Deletes It

FIFA handed reporters three numbers at the close of the 2026 World Cup: 455 VAR checks, 29 on-field reviews, 20 overturned decisions. Before the tournament ended, anyone who wanted to know how often the technology intervened could find out. In that same year, across an entire cricket season, nobody can tell you how many third-umpire reviews took place. Not the ICC. Not the member boards. The broadcaster's hard drives hold fragments.

My Dhaka VAR-Log is a habit borrowed from football, and it turned uncomfortable the moment I applied it to cricket. At Mirpur I sat with a stopwatch through one review: two minutes and fourteen seconds from check-start to verdict. The 3D ball-tracking animation appeared on the big screen. What never appeared was which frame caught the impact, when the cameras were calibrated, or how wide the projected path's margin of error was. In the highlights package the review got five seconds. Six days later I could not find the clip.

The archive keeps the verdict and discards the evidence.

The rule was born in 2026, in Colombo, with the India versus Sri Lanka series. The idea was simple: if the on-field umpire errs, the players should be able to correct it. Eighteen years later the system has many layers, and its bookkeeping is still dark. ICC playing conditions give each team two reviews per Test innings, topped back up after 80 overs; one each in ODIs and T20Is, retained if successful. The signal requesting a review must come within fifteen seconds. A fielding captain, himself holding a fielding position, has that much time.

Ball-tracking decisions contain a chamber called umpire's call. If the ball strikes the stumps less than half a ball's width inside the outside edge, the on-field decision stands no matter how clear the projection looks. The existence of that chamber is an admission that the measurement itself carries error. What nobody admits is the size of the error. Hawk-Eye and Virtual Eye bury their accuracy figures inside commercial confidentiality. The viewer sees a flawless yellow line; the line's own uncertainty is never printed on screen.

In Bangladesh's domestic circuit this is more muffled still. The third umpire watches a host broadcaster's feed. The electronic record of the decision enters the match referee's report as yes or no. Which angle was used, how uncertain the drawing was, what the camera frame rate was — none of it exists in that log. I started the Dhaka VAR-Log precisely because what nobody reports has to be written down by someone.

Since 2026 my notebook has collected 307 reviews across the Bangladesh Premier League, domestic limited-overs matches and a scattering of ICC fixtures. The timings are my own stopwatch and the on-screen timestamps, not the vendor's claims. The average on-screen time came to 81 seconds. That average is the biggest lie in the file, because the distribution bulges at both ends. Boundary catches finish in 40 to 50 seconds. LBW projections drag on for a minute and a half to two minutes, sometimes longer. The viewer's patience is tested exactly where the question is hardest.

More than two in five reviews in my log landed on umpire's call — 41 percent. The system's most common output is not a verdict; it is a stalemate. Of the remaining 59 percent, some are clearly correct calls and some clearly wrong. Blur those three categories together and every argument about the system's performance becomes meaningless.

I separate the two teams' success rates. Where the sample allows, home sides have won 38 percent of their reviews and touring sides 29 percent. The sample is small and the confounders are endless — pitch character, the toss, the number of cameras, even the light. The question still stands, because before the VAR era I saw the same pattern in football: with stadiums emptied, the home penalty rate fell from 0.31 to 0.22 per match. Technology reduces error; it does not erase bias.

Then there are the three frames. An LBW review is decided in three moments. First frame: where the ball struck the pad, the line of impact. Second: where the ball would have been as it passed the stumps, the projected trajectory. Third: how much of the ball would have entered the wicket zone. The first is usually transparent. The second and third are pure model inference, and the uncertainty of that inference is never published.

The vendor's model draws a probable path for every delivery, and the confidence interval around that path never reaches the screen. Yet that number is needed precisely where it is missing. If the ball clips leg stump with a 35-millimetre margin and the model's own error is 20 millimetres, the decision isn't technology; it's a coin toss. The viewer doesn't know, because the viewer is shown a perfect line and a perfect ball.

At Euro 2026 I logged 142 VAR checks across 51 matches, 18 of them overturns. The armpit-offside image that caused the loudest argument wasn't really about the line. It was about the line's thickness, the camera frame rate and the tracking software's interpolation. Cricket's ball-tracking suffers from the same family of problems. One difference: football changed its rules, raised frame rates and handed tracking data to authorities. Cricket has not taken that step.

Then comes blockchain. Since 2026 it has flooded cricket's commercial layer — fan tokens, NFT tickets, crypto-exchange shirt sponsorship. Bangladesh's franchise league has not been exempt; digital-asset branding has entered the jersey economy. None of it touches accountability. This is monetisation, not transparency.

One proposal does fit my log, though. Imagine every raw input of a review — camera IDs, calibration timestamps, the vendor's output, the match referee's final ruling, stadium audio — written as a hash into a permissioned ledger. Nobody could quietly re-cut the clip afterwards, because the hash would change. A spectator could open a public register and see what the uncertainty margin was on a given review in a given match.

It sounds appealing. The problem is at the input.

If the ball-tracking vendor is the only data provider, the blockchain engraves that one vendor's version in stone forever. Garbage in, immutable garbage out. And who runs the ledger? The ICC, the member boards, the host broadcaster — the same institutions that dodge the question today would run the nodes tomorrow. That is not a blockchain. That is a distributed database wearing familiar names.

The real issue sits elsewhere. Umpire's call is not a bug; it is a deliberate error budget. The design concedes that human eyes and machine measurement are both limited, so borderline cases defer to the on-field call. Fans cannot accept it because the uncertainty is never shown to them — only the announcement is played. If the screen read that the probability of hitting the stumps was 73 percent with a model error of 18 millimetres, the argument at the tea stall would sound very different.

That is why starting with blockchain is the wrong order. First comes ordinary publication — review counts, uncertainty figures, latency. Hashing evidence that nobody can see accomplishes nothing.

And there is a gap nobody notices. The argument is always about the line. The weakest part is the clock. A fielding captain must decide within fifteen seconds, often with no information, sometimes on a single gesture from a teammate. At least ten reviews in my log involved a late signal, and that delay is recorded nowhere.

So name the mechanism. The ICC Cricket Committee recommends, the Chief Executives' Committee approves, and a new clause enters the playing conditions: a mandatory Decision Register published within thirty days of every series — total reviews, umpire's call percentage, average latency, overturn count, plus an independent calibration certificate from the tracking vendor. It is not an ambitious ask. Football did much of it in 2026.

Between the confidentiality clauses in vendor contracts and the commercial arithmetic of broadcast rights sits one plain question. When a machine overturns the umpire on the field, to whom is the new decision accountable? If the answer is that nobody is, then blockchain or Hawk-Eye, the yellow line on the screen will remain exactly where it is — at the point where the match ends and the accounting never begins.