Twelve Empty Fields and the Word N/A: When a Martial Arts Analysis System Has Nothing to Read

Three in the morning in Bangkok. I opened the analysis file that had just been pushed back to me and the first thing I did was count the empty fields. Twelve. Article title: N/A. Source: N/A. Article type: unclassified. Core viewpoints: not a single line. Information points: zero. Entity list: a sentence pointing at itself — identify from the information points above. I sat in front of the screen for a long while, not out of surprise, but because the scene was uncomfortably familiar. It was identical to the back row of the press tribune at Rajamangala in 2026, when I raised my hand to ask the home head coach about a 3-5-2 with a tilted left flank and got a sneer instead of an answer. I did not argue. I took notes. But this time the thing sitting silent was not an older man. It was a machine built to read articles and return conclusions. That report was the second tier of a two-tier process. Tier one deconstructs: take a raw article, extract title, source, article type, viewpoints, information points, entity list, time sensitivity. Tier two receives that package and applies the professional frame — eight analytical dimensions, from technical-tactical matchups to market transmission. The whole architecture rests on one contract: every conclusion at tier two must trace back to a numbered information point at tier one. No number, no conclusion. That night, tier one returned an empty package. Structurally valid, fully fielded, with not a single grain of information inside. What made me stop was not the technical failure. It was how the report handled that failure. It did not guess. It wrote insufficient information, cannot assess across all eight dimensions, then locked itself shut with one sentence: any fighter name, record, matchup or business figure produced at this point would be fabrication. A system had just refused to work. In my profession, that is rare behaviour. The only domain label tier one left behind was martial arts. A parent category. It cannot separate at least three analytical tracks that cannot share a single set of criteria. Modern competitive combat sports — MMA, boxing, kickboxing, Muay Thai, grappling — run on win-loss records, finish rates, style matchups and pay-per-view economics. Sanda runs on its own scoring system: hits, throws, ring-outs, weighted quite differently from professional kickboxing because throws and caught kicks are permitted. And taolu — choreographed forms — has no concept of winning or losing, no knockout, only a difficulty score and a performance-quality score. Applying knockout rates to a taolu piece is a category error, and it is the kind of error a system never admits to. It will return a beautiful table instead. I have watched this species of category error for thirty-three years. In 2026, when the pandemic shut every stadium, the desk asked me to write about football without crowds while nobody had any data. I went back through Opta figures for ten leading European sides before and after the shutdown and built my own quantitative frame: pressing indicators, completed passes, possession duration. The result: home advantage shed 11.3 percent of its shots on target when the stands were empty. No specialist outlet had published that figure at the time. But to get it, I had to accept something many colleagues will not: state the limits of the method before stating the conclusion. The empty report did exactly that. Which is why it is worth more of my time than the hundreds of name-stuffed analyses that cross my desk every week. One more thing caught me. Of the eight dimensions, the health and career-risk dimension was left entirely blank, and the report attached a note: this is a data-absence status, not a finding of no risk. To an outsider that reads as pedantry. To anyone who works in combat sports it is a life-and-death boundary. Cumulative brain injury and weight-cut collapse are the two most persistent acute hazards in this discipline, and both are purely data-dependent variables: number of bouts, number of knockout losses, between-fight weight, history of missed weight. Without a fighter's name there is nothing to screen. A blank field in dimension six must never be read as a health certificate. Three further dimensions sat in the same state. No organisation was named, so no operating tier could be assigned — UFC, ONE, Bellator, PFL, the four major boxing sanctioning bodies, Glory, K-1, ADCC or the international wushu federation — and therefore neither the governance architecture nor the revenue flow could be identified. Not a single monetary figure appeared: no transfer fee, no contract value, no pay-per-view buy rate, no gate. No betting line, no media prediction, no fan poll, so the expectation-gap table — the core defence against promoter-driven overrating — could not be built. Three dimensions, three voids, and not one of those voids was permitted to become a conclusion. Then comes the part I consider most serious, and it sits in the recommendations rather than the analysis: silent null propagation. Tier one emitted a shell marked deconstruction successful with an information-point count of zero. No hard gate tripped. Tier two took that shell, ran all eight dimensions anyway, and printed a document that looked complete. If the end reader is an editor under output pressure, that document becomes a bylined article very easily. The most dangerous article in the world is not the empty one. It is the one stuffed with names that do not exist. I have a professional habit dating from 2026, after I mispronounced Luka Modrić three times in the first half of a World Cup group game and was mocked online all night. Instead of deleting the post, I spent a month reviewing footage of all 64 matches and built a pronunciation table of 512 players with nationalities, stress marks and dialect variants. It became an internal document at the broadcaster and was still in use at the 2026 World Cup. The lesson was not be careful with names. It was this: every mispronunciation is a system trying to say something. I misread one name, but the system had misread all of us. The empty report said the same thing in a different language. It was not wrong. It simply showed that the article-reading machine had broken at ingestion, and that the entire chain behind it was preparing to legitimise emptiness as a professional surface. This is where I want to go against the room. The usual reaction to an empty report is to blame the tool: weak model, poor input, immature process. That reading is convenient and partly true. But it misses a paradox: in sports media, the most severe data shortage is usually not mechanical. It is human. I have sat in meetings where nobody dared say I have no numbers. I have read smooth previews of matchups whose author had never watched a single round of one of the two fighters. The machine, when it had nothing, said nothing. People, when they have nothing, usually write another three hundred words. Second paradox: the over-broad martial arts label is nobody's personal mistake, it is how an entire industry categorises content in order to sell advertising. A taolu piece and a piece about an MMA title fight sit in the same section of the homepage because they draw the same reader pool. But they cannot share one evaluation standard. That ambiguity is not a technical detail — it is the place where a system ranks human beings without knowing what it is ranking. Third paradox, and this is the part I want to leave open. My industry is building ever more sophisticated pipelines to read articles, while the skill we are losing is the skill of saying I do not know. An empty season does not lack spectators — it lacks data about the heart. A report that knows when to stay silent is a report with dignity. The trouble is that dignity does not appear in any metrics table, so nobody rewards it. One detail I have not resolved. Every blank field in that report said N/A rather than being truly blank. That means the extractor ran, reached the document, and returned nulls — a fetch or parse failure, not a title-less document. Somewhere out there is a real article, with a real author, possibly with a real fighter waiting to be read correctly. The empty seat never lies — it only exposes what we do not want to hear. And what it exposed this time was an article our system dropped before anyone had a chance to read it. The minimum viable input the report set for a re-run is embarrassingly short: title, outlet, publication date; three to five numbered information points; at least one named entity — fighter, coach, promoter, federation; one specific sub-domain tag; and if the piece concerns a bout, that bout plus every record, ranking or result quoted. Five lines. Five lines and all eight dimensions switch back on. What you measure on the screen is not the draw — it is what you hear in the silence. That night, in the silence, I heard exactly one thing: the system did not fail because it did not know. It failed because it was nearly allowed to pretend that it did. The question I carried into the next morning was not how to fix the machine, but how to teach an entire newsroom that saying insufficient data is a professional act rather than a failure to be hidden.