Nine Sections of Analysis, Not a Single Line of Data

At 2:47 in the morning, a file titled “stage-2-deep-analysis — table_tennis” arrives in the inbox. Nine sections. Nine tables. A seven-row risk matrix. A five-tier information-value rating. A glossary of technical terms at the end, with a note explaining that it was deliberately left empty. A one-line disclaimer. I scroll down. Every cell looks like every other cell: “N/A – insufficient information.” Advancement: N/A. Execution effectiveness: N/A. World ranking: N/A. Event tier: N/A. Source tier: N/A. Roster structure: N/A. Governance risk level: N/A. Four of the five information-value categories receive one star out of five. The fourth receives two stars, with an explanation attached: the report’s only value is that it documents a reproducible null-input case. One cell has words in it. The domain label: table_tennis. I started in this trade at the fact-checking desk of Sports Illustrated in 2026. The job then was simple to the point of being crude: read the draft, find what might be wrong, make the phone call, mark the margin in red pencil. An editor taught me a line I still use. If a reporter writes “this team has won eight of its last ten,” I do not need to know whether they won. I need to know whether there were ten. Thirty years later that desk has become a pipeline. I sit in Beijing, covering table tennis for the Chinese market, travelling with a research group of twenty-seven people and carrying tables that nobody in Vietnam, back when I was learning to write, would have thought to build. The pipeline has just handed me a report with nothing in it to check. The pipeline has two stages. Stage One reads the source article and strips out the driest material available: title, source, article type, core viewpoints, the list of information points, the list of entities, time sensitivity, source quality. Stage Two takes those fragments and runs them through nine domain dimensions built for table tennis: technique and equipment; player data and head-to-head records; event systems and points rules; the China-versus-the-rest landscape; rules and governance; coaching staff and the talent pipeline; the risk surface; public narrative; and industry transmission. Stage Two’s governing principle is severe: every conclusion must be anchored to a Stage One information point. No anchor, no conclusion. The rule exists precisely to override the writer’s instinct — the instinct to fill a gap with what one already knows. Stage One returned a blank page. Title: none. Source: none. Article type: unclassified. Core viewpoints: blank, not a single summary sentence. Information points: empty in the literal sense, not one atomic fact. Entities: the field instructs “identify from the information points above,” while above there is nothing to identify from. Time sensitivity: “not assessed in Stage One.” Source quality: “judge from the source fields,” while the source field reads N/A. The report did exactly one thing, and did it thoroughly: it refused. No player was assigned a metric. No event was ranked. No tactic was dissected. There is not a single name in it. Not Ma Long, not Fan Zhendong, not Wang Chuqin — and nobody else either. All nine dimensions stood still, and every place where they stood still was written down in words, with reasons attached. The author even recorded that this was deliberate abstention, not oversight. I read that passage three times. In my trade, courage is not saying what you think. Courage is leaving blank the cell you cannot fill, and then answering for the blank. A page of N/A looks uniform at first glance. Read closely and the blanks have different shapes, and the shapes are the evidence. The report isolates one detail worth more than the other eight sections combined. If the source article were real but merely thin on table-tennis content, Stage One would still have returned at least a title, a source, and an article type. Those are fields the extractor fills mechanically, without judgement. Here, even the mechanical fields are empty. Title N/A. Source N/A. Article type “unclassified.” The conclusion follows: the failure sits in retrieval or in document-type detection, before the machine ever reached the content. Four causes are ranked by probability. First, the fetch step returned an empty document — a paywall, a bot block, a JavaScript-rendered page, or a PDF or image the parser could not read. Second, a field-mapping or serialisation error between the two stages. Third, the article was routed to the wrong domain label. Fourth, a transient failure upstream. The only populated cell is the domain label, and the report handles it the way I wish every newsroom would: it does not trust it. The reasoning is technical. The label is assigned in Stage One, and Stage One failed across every other field. A label that survives when everything around it dies may be a genuine routing decision, or it may be the default value the system writes when it does not know what to write. I saw this in newsrooms in the 2000s, when tagging was done by keyword. One photograph of a table-tennis table at a sports conference was enough to push an entire boxing piece into the table-tennis section. The system did not lie. The system guessed, and then recorded the guess as though it were knowledge. The practical consequence is concrete: when the source document is recovered and the pipeline is re-run, do not carry the old label forward. Verify it independently first. This is where I want to linger, because it reaches well beyond one broken report. An N/A cell does not mean “neutral.” It does not mean “normal.” It does not mean “no problem here.” It means one thing only: nobody measured, and nobody has grounds to speak. Those three states get blended together in almost every piece of sports analysis I read. The report sets governance risk to N/A rather than “low,” and explains why with unusual clarity: writing “low” would assert that a governance situation had been examined and found benign. Nothing was examined. Writing “low” is inventing a conclusion out of nothing. In table tennis this kind of invention has a familiar shape. The foreign-match win rate — the win rate against associations other than China — is the metric that separates world ranking from real strength most sharply. It is also the hardest to compute, because it requires unpicking the calendar, the opponents, and the rounds. When nobody computes it, the default inside a fan’s head becomes “fine.” That word “fine” has no source. It is a gap plastered over with cement. My experience of watching matches taught me something that runs against the popular belief: a heat map tells you where the ball landed, not why the player was standing there. It is a new form of astrology, well designed and absolutely trusted. What it conceals is a player’s real role inside a tactical system — something visible only across a whole match, never in a count of landing points. The first dimension asks about technique and equipment: how far the playing style has advanced, how effectively it executes, whether the physical profile fits, whether the rubber and blade have changed and whether the adaptation period is over. Without a player’s name, all four questions are meaningless. The second dimension asks about player data: ranking, the points-defence pressure of the rolling 52-week cycle, head-to-head results over the last two years, records at the three majors, performance in deciding games. The report notes plainly that this is the dimension most dependent on named persons, and therefore the first to collapse when