Stratigraphy of an Empty Cell: In Cricket's Data Pipeline, 'No Data' Is Not 'No Risk'

That morning there was a spreadsheet in front of me, and inside it, one empty cell. The header read 'ball recoveries per over'. Directly beneath sat a formula that was treating the blank as zero and averaging accordingly. The software had not made a mistake. The person who wrote the formula had already decided that absence means zero. The truth was different: the ball-by-ball feed for that match never arrived. I opened the notebook before the legend was written, because the empty cell was itself a document. It said this — where there is no information, there is no analysis; and where there is no analysis, every decision taken is standing on top of that empty cell.

In cricket this error is the most expensive, because the decisions are made off the field — in the dressing room, at the auction table, in board meetings, in a sports ministry's file. A wrong average means a seamer's overs are misallocated. A wrong blank means a young batter's career. A wrong provenance means a fan token worth a fortune whose foundation is nothing at all.

Context: a two-stage pipeline and cricket's uneven data literacy

Modern cricket analysis runs on a two-stage pipeline. The first stage decomposes raw material — match events, figures, quotes, sources. The second builds deep analysis on those fragments: format, player, team, league, rules, risk, public sentiment, industry transmission. Between the two stages sits a narrow bridge called an 'information point'. When the bridge collapses, the second stage stands empty-handed. Then two roads open: either admit the empty hand, or fill it with guesswork. The second road is always more comfortable, and always more dangerous.

Cricket is unusually exposed to this crisis because the game's data literacy is brutally uneven. At the top level, every ball's pace, bounce, line, length and even spin revolutions get recorded. One layer down — bilateral series among associate members, women's domestic leagues, under-19 regional tournaments, A-team tours, warm-up matches — often not a single number beyond the scorecard is preserved. Where there is no broadcast, there is no Hawk-Eye; where there is no Hawk-Eye, the height of that delivery, the release point, the swing will never be written anywhere.

The empty stadium still had strata to read. During those months of 2026 I did exactly this work — when Brazilian under-20 leagues returned to empty stands, I coded eleven matches from a single camera angle and surfaced a defensive midfielder nobody was naming. The number was 8.3 ball recoveries per ninety. It was not a statistic; it was a dig site. Every ball-by-ball sheet from an associate-level or under-19 regional cricket match is the same kind of dig site, if somebody preserves it.

This is where the blockchain question enters, and it should. Over recent years cricket boards have moved toward fan tokens, digital collectibles and tokenised memberships. The entire value of that model rests on one premise — that match data is clean, verifiable and tamper-evident. If the data layer is broken, the token is only a handsome wrapper around an empty cell. Tokenising before provenance is verified means gilding a guess.

Core analysis: the stratigraphy of an empty cell

The first-level error begins when someone treats the word 'null' as a single thing. Missing information comes in three kinds, and each demands entirely different treatment.

Structural null — the question was never asked. Example: nobody has ever measured a leg-spinner's flight trajectory in a bilateral ODI among associate members, because the measurement infrastructure does not exist there. This is not a lack of information; it is a lack of a measurement decision. Filling this null requires new match feeds, not new formulas.

Extraction null — the information was born, but was lost in the pipeline. A parsing error, an empty payload, a broken ingestion script — any one of these can reduce an entire match's analysis to zero. The danger is here: this null looks exactly like a structural null. Both are empty cells. But one means 'we do not know', the other means 'we had it, and we lost it'. If a downstream system merges the two, lost information permanently drops out of decision-making.

Genuine null — the event never happened. A bowler did not send down a single wide in that match; the question of an average does not arise. This is the only null that can safely be read as zero.

Not separating these three is the central error of cricket analysis. And that error has a specific social cause: an empty cell raises a question, and a question demands an explanation from the analyst — which no editor, executive or coach actually wants to hear. So the empty cell gets filled with inference, and the inference matures months later into a selection, an injury, or a lost auction.

Cricket's null-rich strata can be mapped. Beneath top-level international cricket sit at least five layers where information is generated but never stored: (a) bilateral series among associate and emerging members, especially the unbroadcast ones; (b) women's domestic and A-level matches; (c) group stages of under-16 and under-19 regional tournaments; (d) team preparation and intra-squad games; (e) franchise trial matches unaffiliated with a main squad. The best prospects hide in the sediment of untelevised games. An analyst who scouts only broadcast matches is not building a talent map; he is building a popularity list.

Workload excavation: overs, flights and recovery windows

In the summer of 2026 I counted 64 competitive matches played by one footballer — Barcelona, then the Euros, then Tokyo. Building a model from minutes, high-intensity sprints and recovery days, I wrote that a soft-tissue injury risk in the following club season was severe. In September, the quadriceps injury arrived. That model belonged to football, but its architecture is universal — minutes, travel, rest. A footballer's minutes were not a statistic; they were a dig site. In cricket the equivalent dig site is more complex, because three different formats make claims on the same body.

A load model is a stratigraphy of a career. The upper layer shows numbers — overs, runs, averages. The layer beneath holds travel: Dubai to Melbourne, Melbourne to Lahore, two different climates inside three days. Below that sits the administrative layer: NOCs for franchise leagues, board clearances, insurance clauses, injury-reporting obligations. And at the very bottom lies the layer nobody ever writes down — the recovery window, the body's actual rest between two spells.

In a tournament cycle these layers erode fastest, because time is compressed here. An international tournament's schedule is built on broadcast contracts and logistics, not on physiology. So the question that should be asked is — 'if this seamer bowls four matches in a row, by what percentage does his expected pace drop across the next two spells?' The question that actually gets asked is — 'he's fit, right?' The first question needs a model. The second needs a nodding head. And inside that nod sits zero information, which nobody admits is zero.

Bangladesh's pace unit is a familiar example of this question, though the team is chosen for its structure, not its numbers: a limited pool of fast bowlers, the same faces across three formats, and franchise-league demand for a large part of the year. In that structure an injury means not just one player's absence — it means rebuilding the entire spell plan. Yet this risk is usually measured by 'did he play or not', not by 'how much bowling load is he carrying'.

Base rates versus signal: the apophenia trap

The greatest risk in my profession is not failing to find talent. It is finding talent in the wrong place. An eighteen-year-old averages forty across six innings. The media will call it a 'discovery'. But the real question is — at this age, at this level, of all the young batters who have touched that number, what proportion held it over the next two years? That answer lives in base rates, and base rates live in long-term scorecard archives — which are frequently not archived at all.

Here one methodological rule is sacred to me: declare the threshold first, then look. That means writing down at the start of the analysis what would make me call something a 'signal' and what would make me call it 'noise'. Take, for instance, a spinner with an economy rate of 3.9 at an under-19 regional tournament. That number says nothing on its own. It speaks only when a threshold was declared beforehand — that a minimum of one strangling over per four-over spell (a dot-ball ratio close to a maiden) must be cleared before he enters the 'high-potential' list. Without a pre-declared threshold, what happens is that the number is seen first and the story is built afterwards — and the story always makes the number look reasonable.

This rule matters even more in the blockchain era, because once a wrong threshold is placed into a smart contract or an automated scouting feed, it quietly keeps selecting the wrong players for years. Code does not forgive; code only executes.

Provenance: a rumour is an artefact

Every transfer rumour is an artefact until its provenance is checked. In cricket this is even truer, because the auction cycle, the NOC cycle and the selection cycle are all periods of informational instability. During these windows, the 'source' reporting an injury before a match is often an agent, sometimes a rival franchise, sometimes simply an account reposting an old photograph.

Here blockchain has a genuine use — but not where the conversation usually sits. The use is not in the token; it is in the ledger. If a match's ball-by-ball data is written so that nobody can quietly alter it later, the analyst's question can shift from 'who gave us this number' toward 'what is this number saying'. A tamper-evident record does not reduce the number of false stories; it reduces their lifespan.

A warning is necessary here. The market value of a fan token rests on engagement, and engagement rests on narrative. But a player's development, injury risk and selection probability are not narrative — they are measurement. If someone mints a token without cleaning the ledger, they are essentially marketing an empty cell. Pricing before provenance is verified means gold leaf on a guess.

Confidence tiers: field notes first, full report later

My own past taught me a lesson I keep forgetting and keep having to relearn. In 2026, after the Russia World Cup, I wrote a scouting note — roughly 1,200 words, with footnotes, tables and diagrams. The note was right. But I published it three weeks late, because I wanted the footnotes perfect. In those three weeks my first 400 followers arrived from someone else's writing. Perfectionism ate my momentum.

The solution was not compromise but tiering. Now I publish field notes first — short, dated, with the level of inference clearly flagged. Then the full report follows slowly. Every field note carries three confidence tiers: high confidence (multiple independent sources, the number verified), medium confidence (one source, the number internally consistent), low confidence (a signal exists, unverified). This tiering is a contract with the reader — I do not know what I do not know, and I will not hide it.

In professional cricket this contract is almost absent. A club's internal report does not state its confidence tier; it states only a recommendation. A board's note records only 'suitable' or 'not suitable'. So when a forecast later proves wrong, nobody is accountable — because nobody ever wrote down that their confidence was medium rather than low. An independent audit trail is not merely an ethical question; it is the only way to improve future forecasts.

Contrarian angle: 'more data' is not the solution, it is a multiplier of the problem

The industry's reflex is — collect more data. I think that reflex aims at the wrong target. More data creates more nulls. Adding a new feed means new parsing errors, new empty fields, new ingestion failures. For an organisation that has not fixed its null handling, twice the data means twice the confidence with the same amount of error.

A second contrarian observation is more uncomfortable. Executives demand clarity, not uncertainty. A coach does not want to hear 'I don't know'; he wants a number he can act on. That demand creates pressure on the analyst to fill the empty cell with inference — and that inference later acquires the status of a decision. In cricket the big mistakes are not usually born of ignorance; they are born of confident misinformation.

Third, the wave of fan tokens and digital assets creates a reverse risk. The speed of tokenisation exceeds the speed of data cleanliness. The market arrives first; the ledger arrives later. Inverting that order produces a market where asset value is set by broadcast intensity and chatter rather than verifiable information. This is the same bubble architecture visible in the young-player market — where a valuation of one hundred million euros lands before fifty top-flight appearances. Price built on incomplete information is not an investment; it is a wager.

This is my central disagreement: the person the industry calls the 'best scout' is not the one who finds the hidden gem. The best analyst is the one who can say, with evidence — 'I do not know yet, and here are the three reasons why.' Because the reward for finding a hidden gem is uncertain, but the reward for preventing bad information is guaranteed.

Takeaway: the next competitive edge is null management

In cricket the next frontier is not more data collection. It is null discipline — the capacity to separate three kinds of absence, to pre-register thresholds, to declare confidence tiers, and to keep an audit trail someone can open five years later. Analysis departments doing this work today will hold the advantage tomorrow; those that do not will repeat the same mistakes with double the confidence.

The question now stands before every analyst facing an empty cell: when you see the blank in your next report, will you fill it with a zero — or admit that it is a dig site?