The Zero-Information-Point Report: A Case Study of Silence in Cricket Analytics

Rajshahi, 7:30 in the evening. The window is open against the May heat, the fan is on its second setting, and there is a file open on my screen — eight chapters, each heading immaculately placed: format and match analysis; player technique and data; team landscape and ranking; league and commercial ecosystem; rules and governance; risk matrix; public narrative; industry transmission. I read all eight in twelve minutes. There is not one sentence containing a team, a player, a venue, a date or a score. Every cell carries the same line: “N/A — insufficient information.” Exactly one field in the entire document is populated: Domain label: cricket_world.

I have watched the game for forty-eight years and have written from a scorecard since 2026, when I started on Prothom Alo’s Wills Cup coverage in Dhaka. I have never seen a sheet like this. It is not a blank scorecard. A blank scorecard at least carries the team names and the overs column; someone in a tea shop will argue over it. This is different: a sheet with the name of a game on it, no score, and beside every column a polite note that the information does not exist. At sixty-four, I still trust the anomaly more than the average. This file is an anomaly, because the failure here is tidy, courteous, and spread across eight chapters.

Scoring and analysis are two trades that cricket keeps confusing. The pipeline behind this file runs in two passes. The first pass decomposes the article: title, information points, author stance, entities, time sensitivity, source quality. The second pass builds eight analytical dimensions on top of that decomposition. The first pass is the scorer; the second is the commentator. When the scorer’s book is blank, the commentator’s report is fiction no matter how elegant the prose.

I know both passes from my own work. In 2026 my World Cup accreditation arrived through The Half-Space, not through twenty-four years of print. I carried one notebook through sixty-four matches in Kazan and hand-coded 1,200 pressing sequences. On 30 June in Kazan, France beat Argentina 4-3; my notebook held France’s 4-2-3-1 with Blaise Matuidi pinned to the left touchline as a defensive winger, Argentina’s midfield screen dissolving, and three French goals inside eleven minutes — 57’, 64’, 68’. The Matuidi piece ran 6,000 words; my editor cut it to 900 and paid me for 900. That cut was not an injustice. It was a gate. The coding was pass one; the column was pass two. Thin coding, short column.

Cricket showed me the same gate in August 2026 at Sher-e-Bangla National Stadium in Mirpur. Thirty-four degrees Celsius, 81 per cent humidity, Bangladesh beating Australia by 20 runs, Shakib Al Hasan taking ten wickets in the match — 5/68 and 5/85. I filed the match report for the daily, then wrote a separate 4,200-word piece on how Bangladesh’s bowlers survived 88 overs of thermal load, and published it on The Half-Space from my desk in Rajshahi. The daily’s editor never ran it. Nine hundred subscribers arrived in eleven days. Ten wickets in Mirpur taught me that a newsletter nobody asked for can still be a control group.

In 2026 the control group got a second edition. When sport stopped, freelance income fell sixty per cent and I retreated into film and data. The Bangabandhu T20 Cup ran from 24 November to 18 December 2026, entirely at Mirpur, with zero spectators. I coded all 33 matches and 4,112 balls. Death-over wickets for the designated home side fell from 38 per cent to 24 per cent. The 2026 silence was not an absence; it was a variable with a pulse. The paper was rejected by two journals and read by 40,000 people.

That variable is useful again tonight, because the silence in the file on my desk is not a mere absence either. It has a shape, a fingerprint, a cause and a threshold. The question is not what the article contained. The question is why the article was never read.

The first problem is format anchoring. Cricket metrics do not travel between formats. A strike rate of 130 in a Test and a strike rate of 130 in a T20 are different animals; an economy of eight runs an over is par in a T20 and a catastrophe in a Test. The report’s first dimension refuses to guess the format, and that refusal is correct. The cost of it is that the seven dimensions beneath also go dark, because every one of them hangs from that anchor. A format tag is not a small piece of information; it is the frame inside which every other number acquires meaning.

I learned the unit problem in Kazan by hand. I thought I was collecting pressing data; in fact I was choosing a unit — “pressing sequence” — and that unit decided which patterns would be visible to me and which would not. Matuidi pinned to the left touchline existed as a finding only because my sheet had a column for a defensive winger’s average lateral position. Without that column, three goals in eleven minutes would have remained three goals. The Stage-1 extraction is exactly such a sheet. The unit you choose decides which patterns are invisible to you.

The second problem is role dependence. Opener, anchor, finisher, pace, spin, keeper — the metric set changes with the role, and one player’s sin becomes another’s virtue. A finisher’s dot-ball percentage and an anchor’s dot-ball percentage are not the same offence. Without a name and a role, metric selection is not difficult; it is illegitimate. The 38-to-24 figure from 2026 meant something because I had nailed down four anchors in advance: phase (death overs), role (bowler and batter), format (T20), venue (one ground). Remove any one and the number drops from analysis to decoration.

The third thing is the fingerprint. Label populated, body empty. There are two explanations. One, the article genuinely carries no information — rare in cricket, where a page without a team, a player or a date is close to a contradiction. Two, something broke between fetch and extraction: a paywall, a consent wall, a JavaScript-rendered page, a redirect, a video-led gallery. That class of failure produces precisely this signature: metadata survives, body dies. Which is why the most valuable line in the report is its admission that the dominant risk here is informational, not sporting.

Cricket already lives with label-without-body scorecards. The wicket is recorded; the dropped catch at long-on in the previous over is not. The DRS projection is recorded; the projection’s margin of error is not. The A-tour match has a result and no ball-by-ball. Data that is not recorded does not vanish neutrally; it vanishes selectively.

That selectivity is the real risk, because the failure is not random. Paywalled, JavaScript-rendered and video-led sources will fail disproportionately, and the corpus will drift toward text-heavy outlets and away from broadcasters. Structurally this is the twin of home-ground bias: count only the grounds with good camera coverage and you will systematically overrate the players who play there. The error belongs to the boundary of the count, not to any player. The report’s recommendation to track the empty-output rate per source domain against a 5 per cent threshold is the right method for exactly this reason: missing data clusters, it does not scatter.

The report also lists what each empty dimension would have required, and that list is really a quality-audit design: a format tag and innings state; a name, a role and a sample basis; a team, a ranking and a home-away split; a league, a rights figure or an auction price; a governance decision, a rule or a DRS-DLS controversy; a workload or injury record; a stated expectation to measure against; a rights deal or a capital flow. None of these exist. All eight dimensions are silent, but the silences are of different kinds, and telling the kinds apart is the work.

Before the commercial layer, one discipline needs restating, because thin data is precisely when everyone breaks it. A price is a market’s opinion; an average is a bowler’s record; they are different instruments and cannot be read on the same scale. A player’s auction price rising does not mean his international strength rose; it means squad quotas and demand gaps moved. With little information, people merge the two, and the merger produces the “expensive but ineffective” narrative. There is no auction figure here, so no premium can be judged — but the emptiness is itself a warning: in an empty cell we reach first for the price, because the price is easy to obtain and performance is hard.

The final recommendation is engineering-precise: when information points equal zero, do not run pass two; return INSUFFICIENT_INPUT. I recognise that gate, because my own life contains a version of it. In 2026 the daily’s editor never ran the 4,200-word piece. That too was a gate — editorial rather than technical — and its output was identical to INSUFFICIENT_INPUT: a valid finding removed from the corpus without explanation. Nine hundred people found it anyway in eleven days. The lesson for the pipeline is plain: a gate that can only say “insufficient” tells you nothing about what was gated, why, or how often.

A tactic is a hypothesis; the match is peer review. So is a gate a hypothesis, and the re-fetch is the review. Here that means specific chores: pull the source again with a rendering-capable client, log the HTTP status, log the final URL, log the body length. Only when the body returns at a non-trivial length should pass one run again. Without those logs, whatever follows is guesswork wearing a lab coat.

Here I part company with the report, which is rather pleased with its own honesty. “I will not manufacture teams, players or formats” is the right sentence, but it is a format, not a virtue. Anyone can type N/A into an empty cell; the hard work is finding out where the empty cell came from.

Notice also that the report has already issued a rating on a null: sporting value, one star out of five. That is a verdict on the quality of an article nobody read. Cricket does this daily — we settle a player’s future on twelve balls, declare a series decided on a DLS par score, fix a batting order on one innings of strike rate. The most dangerous number is the one computed on an empty input.

The second gap is quieter. The gate the report recommends conceals the incident it is meant to catch. A silent INSUFFICIENT_INPUT is calmer than a wrong answer, and calm failures are never fixed, because nobody complains about them. If any downstream consumer reads an empty pass one as “no finding,” the pipeline will systematically under-report precisely the articles it failed to parse.

The third gap is the most uncomfortable. The domain label survived. If cricket_world was assigned from the URL or the metadata rather than the body, then the one living field is not evidence of anything — it is a routing hint dressed as a discovery. When a label is the only surviving piece of information, it is not information; it is habit.

So what do I watch next? Three things. The empty-output rate of pass one, segmented by source domain; anything above roughly 5 per cent on a single domain is a systemic fetch fault, not bad luck. The provenance of the label — content-derived or metadata-derived — because that single answer decides whether the label can be used analytically at all. And the result of the re-fetch, with status, final URL and body length logged. Without those three logs, the same file will return in the same silence next month.

I have a duty on my own side too, and I am writing it down now so that I cannot deny it later. From this week the notebook carries a second column: the matches I could not parse. The 1,200 sequences from Kazan taught me that hand-coding scales; the 4,112 balls of 2026 taught me that silence is codable. A third lesson is arriving — failure to read must be coded too, or the failure rate stays permanently unknown.

The next verification is a simple question. When the pipeline returns another empty book, will anyone ask who the scorer was? In cricket we never ask the scorer anything, though every number we quote is his handwriting. At sixty-four, I still trust the anomaly more than the average — but only when the anomaly has been written down.