Lessons from an Empty Dataset: Cricket Analytics, Missing Information, and the Immutable Ledger

At half past one last night I opened a spreadsheet on my laptop. The columns were there, the rows were there, the formatting was immaculate — but every single cell carried the same sentence: "insufficient information." Format — Test, ODI, T20; who batted, which venue, how the pitch played, whether there was dew, whether DLS intervened — the same silent answer everywhere. I dragged the cursor slowly across the empty cells and felt, in place of the thrill I had at twenty when I built Pedri's load dashboard, a kind of pitiless emptiness. The spreadsheet tells me nothing; instead it throws a question back at me — what are you going to do now?

The question sounds harmless, but it is the oldest trap in cricket analysis. When we lack information, we usually do one of two things: we honestly admit we don't know, or we fill the empty cell with a guess. The second is more comfortable, because a guess at least gives a story, and stories satisfy people. Yet my whole profession rests on the opposite principle — an empty cell is itself a datum; filling it with a guess means fabricating information. This is not merely an ethical question, it is a methodological one, because fabricated information later poisons the decision that trusts it.

Context: The game that records everything, yet trusts nothing

Cricket today is one of the most data-productive sports on earth. Hawk-Eye tracking for every ball, a second-by-second log of every run, field-placement maps, bowling loads, sprint counts, strike-rotation patterns. Broadcasts now run a one-percent win-probability meter continuously, and the viewer at home easily assumes the number descended from the sky. It did not. Every model is built by leaving something out — sometimes fatigue, sometimes travel, sometimes merely an insufficient sample. A model's strength lies not in its number but in its honest acknowledgement of the limits hidden inside its assumptions.

I remember 2026. A seventeen-year-old schoolboy in Singapore, I scraped the event data of all 64 Russia World Cup matches and built a simple xG model. Croatia became my test case: they scored 14 goals from 10.8 xG, and in the semifinal against England Luka Modric completed 89 percent of his passes while covering 10.4 kilometres. The eye-test said Croatia were merely lucky; the model said Modric's progressive passes were the real engine. From then on I dropped the word "luck" and began saying "unsustainable variance." The lesson was one: a claim must carry a number behind it, and every story must survive the model's test.

The same lesson sharpened in 2026. When COVID halted the whole game, I sat in my university's data lab working on the Bundesliga's Project Restart. I compared home win rates before and after empty stadiums: 43.3 percent fell to 33.3 percent. A regression model showed away teams gained 0.18 xG per match in the crowdless environment. Empty stadiums taught me that silence is a variable, not an absence. That paper won me an internship at a Singapore sports-data firm, and taught me that a crisis is also a kind of laboratory.

And the context I am writing in right now is a transfer window — a market of information in which rumour is priced higher than fact. Every day someone says this opener's release clause is such, that spinner's salary is such. Claims without sources, prices without proof. Here the analyst's job is not to add to the noise but to sift signal from noise. And to sift signal you need precisely what my empty spreadsheet lacked — a verifiable source.

In modern cricket analytics, every metric we use — strike rate, economy, strike rotation, dot-ball percentage, powerplay average — carries a blind spot. Strike rate does not say how difficult each ball was; economy does not say how much pressure each over carried. If one bowler bowls at the death and another in the powerplay, placing their economies side by side is to pay two different jobs the same wage. This is where context-adjusted models are needed, and the first condition of such a model is to admit what data is missing.

Core analysis: Emptiness can be measured, but fabrication gets caught

Let me return to the empty spreadsheet. The real problem here is not the absence of numbers but how we interpret that absence. In statistics two things are strictly distinct: "no data" and "data that shows nothing." The first is a dark room — who knows what is inside. The second is a lit lamp that revealed the room to be genuinely empty. Confuse the two and the analysis cuts its own feet, and the reader is deceived.

In nine years in sports analytics, what I see is this: the industry is not accustomed to handling the "no data" state. When a player's injury record is unavailable, we quietly assume he is fit. When a league does not publish audited revenue, we assume the biggest number, because a big number makes a good story. Yet missing information is often itself information — it tells us who is hiding, what is being hidden, and why now.

This is where blockchain becomes relevant, and I insist it is no fashion. An immutable, distributed ledger can partly answer the problem of data provenance. In cricket, data is today scattered across ten hands — broadcasters, boards, fantasy platforms, scouting agencies, betting markets. Who changed which number when, who said "economy 7.2 at the end of the over" while the record shows 7.4 — there is no central system keeping that account. A timestamped ledger can fill this gap: if every ball-by-ball record, every selection decision, every contract update is written once, it becomes nearly impossible to alter quietly. Data integrity then no longer depends on anyone's politeness; it stands on a mathematical rule.

When I look at a database like cricsultan.com, where "cross-checked" is written beside each statistic, I understand — that little word does the real work. Because in the cricket economy information is now a commodity; and where a commodity has a price, counterfeit goods appear. A verifiable ledger builds a wall in that market of counterfeits, one that matters not only for the analyst's sake but for the reader's.

Yet my own experience keeps me cautious, so that I do not see data only through a machine's eye. In 2026, tracking Pedri across the Euro and the Tokyo Olympics, I saw a season of 73 matches. At the Euro his pass accuracy was 92.3 percent; in Tokyo his high-intensity distance fell 11 percent in extra time. The model said — this is burnout risk. I built a load-management dashboard, it became my first paid analytics project, and it helped win me a junior sports data analyst job. But the interesting thing — data alone did not tell me why he was tired; sometimes the rising match count is the answer, sometimes the club's decision, sometimes his own mind. Numbers show direction, not cause.

This is why I refuse to see players as mere assets. Yes, I measure load in minutes and high-intensity distance; yes, I read form as mispricing and fatigue as depreciation. But if a defender sits on a €45m release clause, and a sixteen-year-old boy arrives in the same sentence, I stop — because a price can be written in a ledger, but a boy's body and mind cannot. Here an old conviction returns: former stars opening academies is mostly branding, while genuine grassroots coach education is chronically underfunded. With data we find talent, but data never builds talent — a trained coach and the right environment do.

And a large part of missing information is not accident but structure. Ball-by-ball data for associate-nation matches is often incomplete, the historical record of women's cricket is patchy, and domestic league scorecards go unaudited in many places. This emptiness is not neutral — it tells us whom we value and whom we do not see. A ledger can reduce this inequality, if it records every level of cricket to the same standard; otherwise it merely polishes the accounts of rich leagues.

Seen industrially, it is a chain: upstream is the grassroots and talent supply, midstream the national teams and leagues, downstream broadcast and commercial markets. When data integrity fails, the greatest damage is upstream — because there decisions are made on little information and with heavy responsibility. If a selector knows how many balls a teenager has bowled, how much rest he has had, he might save a career. This is why provenance is not merely a matter of technology; it is a matter of player welfare.

I see risk at three levels. The worst case: emptiness is filled with guesses, a wrong decision is taken, and the ledger records it immutably. The middle case: data exists but is unverified, so the decision is half-blind. The best case: emptiness is clearly flagged, the source verifiable, and the interpretation cautious. The first is the most dangerous, because there the error looks confident.

Contrarian angle: Immutable does not mean true

Here I want to rein in my own enthusiasm. Blockchain protects data integrity — but it cannot catch an interpretive error. If wrong information is written once into a ledger, it remains wrong immutably, forever. Immutability and truth are not the same thing; the chain remembers what was written, not whether the writing was right. Without grasping this difference, we grant the data machine a power it does not have.

There is another trap, my own favourite risk — mistaking correlation for causation. Home advantage fell in empty stadiums, that is true; but it is no proof that the crowd's roar was the sole cause. Travel load, neutral venues, bio-bubbles, shifting umpiring norms were likely alongside it. Watching matches year after year, I learned that the drama of a single match is never proof of a conclusion — it is a natural experiment that needs replication, and a clear admission that other variables were in play. The analyst who drops this caution turns his model into a costume for his own ego.

Finally, one point I consider the ledger's biggest gap: human consent and experience are not recorded on any chain. A player's fatigue is his own story — capturing it in numbers does not complete it. An analysis that measures only the workload but omits the player's testimony, his voice, his permission, cuts off a corner of its own data ethics. So I say: emptiness and testimony are both parts of the model.

Takeaway: The next signal

I did not close the empty spreadsheet. Instead, beside each "insufficient information" cell I added another column — "what remains to be verified." Emptiness stopped being a shame and became a to-do list. The signal I want to watch next season is not runs or wickets — it is the transparency of information provenance. Which platform can give the source of its numbers and which cannot — that gap will decide who is analysing and who is selling guesses.

The spreadsheet was my cloister; the World Cup was my first pilgrimage. Today I know the cloister door is sometimes empty — and learning to honour the empty cell may be the hardest data skill of all.