An Empty Field Is Also a Data Point: Cricket Analytics' Provenance Crisis and the Blockchain Audit Trail

I opened the file on a rain-grey Manchester morning, before the coffee went cold. A Stage-1 deconstruction report. No title, no source, no core viewpoint, no entity, no time-sensitivity. Every field blank, and beside every blank field a single line: insufficient information.

In cricket we know this scene. The scoreboard is up, the floodlights are on, the umpire stands at the bowler's end, and the batter has not walked to the crease. That moment is not a match; it is the emptiness before a match. You cannot write about emptiness unless you learn to read the emptiness itself as a data point.

The first xG model I built did not predict football; it predicted my patience. Standing here today, I see the same rule applies to a pipeline. However elegant the model, an empty input makes the output mere decoration.

The background needs stating. Born in Dhaka, based in Manchester, working in data journalism. In 2026, while a statistics student at the University of Manchester, I built an xG model from 380 Premier League matches. I tested Manchester City's 18-game winning run: 56 goals, but only 44.3 xG — a +11.7 overperformance. The table was clean and reproducible, and I published the raw code and data so anyone could verify it.

Then came Kazan in 2026. Germany: 74% possession, 26 shots, 2.7 xG; South Korea: 5 shots, 0.9 xG — yet the scoreline read 0-2. I built the shot map and the PPDA chart: Germany's PPDA 7.2, South Korea's 24.6. The autopsy was published within twelve hours and shared twelve thousand times. The lesson was plain: penetration, not possession; quality, not volume.

In 2026 the Bundesliga returned behind closed doors. I watched the first five rounds; the home win rate fell from 43.2% to 21.1%, home goals per game from 1.65 to 1.08. Every time the method was the same — establish a baseline first, then measure the deviation. If the baseline itself is absent, you can still build a deviation story, but that is not journalism, it is fiction.

Now the baseline itself is gone. The Stage-1 report is empty, everything N/A. This is where today's argument begins — how to practise the journalism of absent data.

A data pipeline has four stages: feed, label, standardisation, model. If any one is empty, the whole chain is unreliable. An empty field does not mean "no signal"; the empty field is itself a signal. In statistics we call it missing data, and it has two types — Missing At Random and Missing Not At Random. The difference is vast. In the first case imputation may work; in the second, imputation means fabricating a falsehood, the most dangerous habit in cricket writing.

Why are the fields empty here? Three likely causes. One, the upstream feed broke — the source article was either not retrieved or the parser failed. Two, the deconstruction stage ran correctly but the input was zero, so the output is zero — a tautology, but a healthy failure, because it invented nothing. Three, a standardisation gap between the labelling schema and the content domain — the "cricket_world" tag is applied, but there is nothing inside the tag.

Analytics without provenance is a dressed-up cartoon. I have placed Dhaka and London feeds side by side for years. The same delivery, two different labels in two places — one calls it "line", the other "good length"; one tracks speed by gun, the other by radar. The number is the same, but its birth certificate is not. An analyst who does not know this gap is measuring the same data twice and believing it is a double check.

But provenance does not only look backward. It also works forward. If every tournament's data sits on a verifiable ledger, then season-to-season, format-to-format, even country-to-country feed comparisons all become traceable. Much of the standardisation crisis cricket suffers today is not a shortage of information but a shortage of certificates for that information.

This is where blockchain becomes relevant. In cricket analytics, blockchain's real use is not the crypto story but the audit trail. With an immutable record of every data point's origin, timestamp, transformation and correction, answering "where did this 2.7 xG come from" takes seconds, not twelve hours. The ledger would record: on this date, at this venue, this shot event came from this ball-tracking feed; who labelled it, who corrected it, when. Then no empty field remains a mystery; it is either a clear flag or a clear error.

My years of watching matches tell me the commentary box's words "momentum" and "temperament" are often born in a place where someone filled an empty field with a story. The eye test is a witness; the data is the cross-examination. Blockchain is the court recorder of that cross-examination — on one condition: the input must be honest.

This is where my doubt begins. As data people we rush to fill every empty cell, because an empty table looks bad. And trusting blockchain, some believe that writing to a ledger purifies the data. Blockchain does not purify bad data; it makes the error immortal. Garbage in, garbage on-chain — only this time the garbage can never be deleted.

The second trap is mechanism-hunting. Seeing an empty field, the mind wants to invent a cause — "the parser failed", "the source was down". But I did not pre-specify the mechanism, and I ran no placebo test. So it is a guess, not evidence. The difference between a pre-specified mechanism and a post-hoc story is the difference between possession and penetration.

Third, baseline worship. If the baseline itself is wrong — wrong era, wrong competition, wrong pitch, wrong source — then measuring a deviation means stating a wrong answer quickly. In a report that says N/A, the only honest output is: "I do not know, and I know why I do not know." That is not humility; it is method.

The next-round signal is clear. The audit does not begin by filling the empty field; it begins by seating the empty field — fetch the source article, populate Stage-1, then build the table. I do not chase narratives; I build a table and wait for them to arrive. The question now is this: when your pipeline comes back empty, do you write a story, or do you raise a flag?