The Empty Dossier and Combat Sports’ Data Thirst

In a press room at an arena east of Bangkok, a young colleague opens his laptop, clicks into a folder named after a fight code, and waits. The screen returns a blank frame. No title. No fighter’s name. No weight class. No source. All that remains is a single line of label — a generic category tag the system assigned itself, enough to fill the gap but not enough to establish anything. Ten minutes later, he finishes typing his conclusion and hits send. I read that draft twice. The first time to check for errors. The second time, I realized the thing neither of us had said out loud: if I weren’t sitting here tonight, that blank frame could have become a fully formed analysis — smooth, with tables, with risk warnings — and wrong from root to tip. That is where the story begins. An empty dossier, a complete output, and a trap the entire combat sports industry is walking into without anyone naming it. Combat sports media has drifted far from the way I learned the trade. Fifteen years ago, when I was still an athlete turning toward journalism, ringside reporters went down to the bench to read the rhythm of a fight. They sat in the corner of the arena, eyes fixed on how a fighter rose between rounds, on the tension of the hand wraps, on the breath escaping at the corner stool. I have a habit of hearing a fight before looking at the scorecard: the sigh after a missed strike, the sound of a foot switching stance, a fighter talking to himself across two seconds of recovery. I watch fights with my ears, and I hear the passes nobody ever throws. Today most of that work has been handed to data pipelines. An original article runs through several layers: text extraction, category assignment, entity extraction, time-sensitivity scoring, source grading. At the final layer, a model turns that raw material into judgments. The process works well when the input is dense. It has one lethal blind spot: when the input is empty, the system never stops itself. It still has to return an output, because the output format has already been locked in place. That is how a new product was born: the empty analysis. On the surface it looks complete — sections, tables, conclusions, risk warnings — but it contains not one fact. And the most frightening part is not that it is empty. The frightening part is that nobody checks whether the input was full. Based on my own experience of watching matches, the micro-rhythm reading I learned on the training ground cannot be fully digitized — yet it is the only thing that keeps an analysis from drifting. To understand why an empty dossier is more dangerous in combat sports than in any other discipline, you have to start with the question of classification. Tagging something as “martial arts” sounds like an administrative formality, but it determines the entire value of the analysis. That label blurs three entirely different things: professional combat sports — MMA, boxing, kickboxing, Muay Thai, grappling; traditional martial arts performed as routine — forms scored on difficulty and performance quality; and sanda — a hybrid combat branch permitting punches, kicks and throws. These three groups require three different measuring sticks. For an MMA bout, finish rate, takedown defence and control time are valid instruments. For a scored routine, those instruments are meaningless, simply because the concept of a “finish” does not exist. Scoring a routine with the win-loss logic of a professional ring manufactures a systematic false conclusion — the kind of error more dangerous than having no conclusion at all. An empty dossier that cannot be classified has nothing to analyse. Yet the system still returns an output. The most serious gap sits here: weight cutting is the single variable with the highest predictive power for mortality in this field, and it cannot be screened at all without a fighter’s name. No walk-around weight, no weigh-in weight, no missed-weight history, no short-notice replacement. In a genuine news report, this is precisely where the value concentrates — and precisely what promotional framing cuts first. Fighters make weight by dehydrating; the consequences are kidney failure, rhabdomyolysis, collapse at the scale. Those events are worth something for a few hours and nothing afterwards. A pipeline running a day late has already lost the most expensive part of the value. Brain health is the same. Screening for CTE — the degenerative brain disease tied to cumulative head impacts — requires total strikes absorbed, knockout count, concussion history and return intervals. Without a fighter’s name, that set of variables does not exist. It must be said clearly: an empty result here may never be read as “the fighter is safe.” Nothing has been confirmed, and nothing has been cleared unfairly. Next comes the governance vacuum. The applicable ruleset cannot be identified when it is unclear whether the subject is a state athletic commission, a belt organization, or an international federation. The Unified Rules of MMA, boxing rules, K-1 and Glory rules, Muay Thai rules, sanda rules and routine-scoring rules differ on fundamentals — including whether a “finish” is even a concept. Without a known ruleset, no penalty scenario can be simulated: you need a jurisdiction, a rule version, and a specific alleged infraction. Fabricating the scenario produces fabricated probabilities. Combat sports has one more layer few people mention: when the subject is undefined, every transmission model becomes void. Economic models — fighter revenue share, gate structure, pay-per-view tiers — need an originating shock to propagate from. No shock, no transmission path. Of course, general industry reference points can still be cited. Fighter revenue share at major MMA organizations typically sits in the high teens to around twenty percent. For top boxers, that share can exceed fifty percent — close to the split seen in team sports leagues. Entry-level purses for a low-tier fighter are often only tens of thousands of dollars per bout, not enough to cover training costs. But this is general industry knowledge, not data from the dossier at hand. As for why the dossier came up empty, it can be narrowed by elimination. The pipeline ran as far as category assignment — it returned a label — but produced no content. Three common causes: the source file was image-only, such as a scan or a PDF with no text layer; the source was video or a podcast with no transcript; or the source document was extremely short, nearly empty. A fourth possibility deserves attention: the system used a generic classifier instead of a combat-sports-specific one, so it skipped the very step that mattered most without flagging an error. There is another pressure that cannot be ignored: the weight of the output template. When a system is designed to always return a “conclusion,” a “risk warning,” a “signal to track,” the emptiness of the input becomes a formatting fault rather than a fact. The machine does not say “I do not know.” It says “risk at an undetermined level” — a line that sounds cautious but has in fact written an unverified attribute into the record. In medicine, that is a prohibited error. In sports analysis, it is published every day. But stopping here would make this piece nothing more than a bug report. The counterintuitive angle I want to push lies elsewhere: an empty stadium is the greatest mirror of a team’s identity. So is an empty dossier. It exposes what a full dossier conceals. That blank frame exposes more than a lack of data. It shows how accustomed we have become to filling gaps with prose. An entire generation of combat sports analysis is trained to produce text before it produces fact. Faced with that, emptiness becomes a professional ethics test: do you dare send back an empty result, or do you force it full? Here is a paradox I am certain of. The most decisive data of a fight is the hardest data to extract. It is not in the stat sheet. It is in the breath of a fighter as he reaches the corner stool, in how he plants his rear foot after taking a shot, in the silence just before the bell. We talk about finishes, but the real memory is the silence just before the roar. Pipelines are good at extracting numbers. They are weak exactly there. When everyone believes one truth, I start believing in error. When the whole industry believes a full output is the sign of a good process, I start looking at the input. And what I find is usually a blank frame. The final layer of the problem is public narrative. Without a dossier, no story can be classified — from the coronation of a new star, the continuation of a dynasty, to a rematch script or a legend’s farewell. And the paradox of scepticism is this: scepticism needs a claim to push against. With no claim, scepticism itself becomes the only data left. The next signal to track is not on the scorecard. It is in the source folder — where an article can be empty and nobody notices. If this industry wants to keep a shred of integrity, it will have to install a hard gate at the extraction layer: no entities, no conclusion; no fighter’s name, no analysis table. Beyond that, let the blank frame speak for itself.