Nine Dimensions of Table Tennis Analysis and the Value of Refusing to Conclude

On a Tuesday morning, the analysis file arrived from the collection stage. Nine dimensions, correct template, every heading in place, every table cell present. The first line read: insufficient information. The second line said the same. By the ninth line I closed the laptop and understood that I had just received a complete report about something that never existed. No event name, no athlete name, no match date, not a single information point. Only one field remained populated: the domain label, table tennis. And that single label was precisely the thing I did not dare to trust.

Outsiders tend to assume an empty analysis is harmless. The opposite is true. A blank page is harder to handle than a wrong page, because it does not denounce itself. It still looks formal enough to be forwarded, still carries enough tables to be cited, still has nine sections so that someone skimming it concludes that a human being has reviewed it. The greatest danger of an empty report is not that it contains nothing, but that it looks as though it contains everything.

Nine layers every table tennis match must pass through

The framework I use for table tennis has nine dimensions. Technique, tactics and equipment: the topspin loop, the first three shots, rubber and sponge, the adaptation window after a weapon change. Player data and head-to-head records: ranking, points-defence pressure, win rate against associations outside one's own territory, performance in deciding games. Event system and points rules: event tier, the 52-week rolling deduction, entry deadlines. The competitive landscape between one dominant table tennis nation and the rest of the world. Rules and governance, including selection mechanisms. Coaching staff and the talent pipeline. The risk surface. Public narrative and expectations. The ninth layer opens outward into the industry: equipment, youth development, the commercial value of the event.

These nine layers share one easily missed trait: every one of them needs a named entity to attach to. Without an athlete's name, the second dimension collapses first, because the win rate against opponents from other associations — the metric that separates media ranking from competitive reality most sharply — cannot be computed without knowing who played whom. Without an event name, the third dimension loses its temporal anchor entirely. An analysis of Ma Long would need to know where he sits on the age curve; an analysis of Wang Chuqin would need to know which months carry his points-defence pressure. Without a name, both are empty words.

Gaps have structure too

What made me stop was not the number of empty cells but their position. A genuinely thin but real article still leaves traces: a title, a source, a classified article type. Here, even the fields a machine fills automatically were blank. When the mechanical half of a pipeline dies together with the content half, the fault almost certainly sits at the document-retrieval step, not at the analysis step. Three possibilities, ranked by probability: the source page sits behind a paywall or a bot block; a field-mapping error between the two stages; or the document was assigned the wrong domain label.

I sit in front of a screen to see what nobody in the stadium notices. Here, what I saw was a domain label standing alone with no entities attached to it. In data-pipeline operations, that is the signature of a fallback label, not of a genuine classification decision. A label without entities is an unverified label.

The real risk sits on the reader's side

The domain risk list in the report was left blank, and that is the correct handling. Listing injuries, selection pressure, generational gaps, governance risk — all real for table tennis in general, none of them tied to the document that is missing. Inserting them would manufacture something more dangerous than emptiness: false specificity.

Only one row in the risk matrix was rated high, and it belongs to no sport at all. It is process risk: a report with nine complete sections can be read by a downstream system as a finished analysis, then trigger alerts, rankings or summaries built on top of it. The flaw is not in the system; it is in the belief that the system is right.

The industry-transmission dimension was handled the same way. No equipment brand, no event commercial data, no sponsorship contract appeared in the document, so every arrow in the transmission map stayed blank. Writing "neutral" for each link would assume that the channel had been examined and found unaffected — an assumption with no basis whatsoever.

What is worth doing, in order of priority

The first priority is recovering the raw text and re-running the collection stage. That single action converts the entire report from zero into substance, which gives it a higher return on effort than anything else on the list. If the source document is an image, a video or a PDF, ordinary text extraction will keep failing; enabling optical character recognition and speech recognition at the retrieval step is the fix with the highest chance of success.

Once content exists, the most valuable dimension remains the decomposition of win rates against associations outside one's own territory. Based on my experience following matches, this is where media narrative and competitive reality diverge furthest, and also where fans are most often led astray by a single ranking number.

Four signals worth tracking

The null-record rate in each incoming batch: above five percent is a systemic problem, not a one-off glitch. The confidence of the domain label: a label present with no entities extracted is a sign of fallback labelling. The availability of the source field: a null source while other fields are populated shows the pipeline broke at the content stage. And finally, an audit of whether any downstream system is acting on template-only reports.

The other side of silence

This profession carries a familiar pressure: file within 24 hours of the match. I have refused that pressure many times, and each time I lost a little traffic. But there is a distinction readers are rarely given: "examined and found neutral" and "there was nothing to examine" are entirely different states. Writing "neutral" for a dimension that was never checked is the most sophisticated lie an analysis table can tell, because it turns ignorance into a conclusion.

The seventh camera angle shows that truth is a relative concept. In the France–Australia match in 2026, the entire studio insisted the referee was wrong, until the angle from behind the goal was switched on. That lesson has a reverse side: with no camera angle at all, declaring the referee right or wrong is fabrication either way, including declaring that he got it right. A good referee is not one who never errs, but one who knows where he erred. A decent analyst is the same: aware of what is not yet known, and saying so.

The database of 1,400 decisions did not find justice, but it found patterns. The first pattern I drew from it, after six months of rebuilding, was a pattern about missing data: the decisions argued over most fiercely tend to be the ones with the fewest camera angles, not the ones that are most clearly wrong.

What remains after the blank page

For Vietnamese table tennis, this story is closer to home than its technical surface suggests. At domestic events, serve data, point-win rates in long rallies, and the exact moment a player loses rhythm are rarely recorded systematically. Writers are forced to choose: reconstruct from stadium memory, or state plainly that there is not enough data. The second option does not sell much advertising, but it is the foundation for everything built afterwards.

If domestic organisers agree to record just four fields per match next year — who served first, who won the long rallies, who won the deciding game, and who won against opponents from outside the territory — then within three seasons we will have something we currently lack: a domestic predictive model instead of commentary built on instinct. No permission is required. Only the decision to start recording.