The Wrong Framework: How a Missing Classification Makes Every Martial Arts Conclusion Meaningless

Someone sent me a dataset about a martial arts event with one short instruction: analyse it. The data was dense — number of bouts, win percentage, number of finishes before the time limit, average minutes per round, even a distribution chart of strikes by round. Detailed enough that I could reconstruct each fighter's breathing rhythm on paper. But it was missing exactly one line, and that was the most important line of all: which kind of martial art is this. Which school, which ruleset, which scoring system. The person who sent the data assumed those were administrative details outside the scope of analysis. I read it the other way around. Without that line, the entire dataset above it collapses into a pile of numbers with no unit of measurement. An 82 percent win rate in professional boxing tells one story; the same rate in an unsanctioned challenge bout tells a completely different one. When data starts to talk back, tactics finally speak — but only if we know what sport we are asking the data about. More than forty years of watching sport taught me a suspicious habit of the trade: people lump everything under one word, 'martial arts', and then analyse it with a single toolkit. That default toolkit is MMA or boxing. Wins and losses, knockout rate, finishing ability, first-round stoppage count. It is convenient, and convenience itself is why it has survived so long. The problem is that 'martial arts' is a label far too broad, wide enough to hold at least three classes of subject that differ in nature. The first class is modern competitive combat sports: MMA, boxing, kickboxing, Muay Thai, wrestling, grappling. There, opponents exist, wins and losses exist, finishes exist, and every statistic means something. The second class is traditional and performance martial arts, typified by taolu — forms and weapons routines. There, no opponent exists. There is no one to defeat. The third class is sanda, a contact discipline with its own ruleset, in which a throw or a push that sends a fighter off the platform scores on par with a clean strike. These three classes require three different systems of logic. Blend them together and the result does not land in the zone of reduced accuracy. It lands in the zone of systematic error. Try building the framework for the first class. In modern combat sports, technical questions always open with the style matchup — when two approaches meet, who gets exposed first. When a grappler meets a striker, the issue is distance, timing of entries, and who controls the space before the fight has time to take shape. Next comes the quality of the record, and this is where data gets abused most. Whom did you beat, at what stage of that opponent's career, in which weight class, after how many rapid weight cuts. A clean win rate against opponents past their peak is worth less than an ugly win rate against opponents in their prime. People pull out the win-loss number and stop there, while the real answer lies in the composition of that run of bouts. Then there is the conditioning variable. Age comes with fight mileage. Fight mileage comes with wear. But the greatest pressure on a modern combat athlete usually comes from the weight cut. Someone who forces a body from its natural weight down to a competition division in forty-eight hours and then rehydrates in twelve is trading away something the dataset never records: neural performance. With no weigh-in result and no rehydration history, every judgement about form is guesswork. The next opponent may not be the biggest risk. The sauna is. Move to the second class and the toolkit collapses entirely. In taolu there is no win rate. There is no knockout. The athlete walks out alone and performs the routine. The scoring system is divided into panels: one for technical difficulty based on a pre-registered movement list, one for execution quality, one for overall impression. Points are deducted for errors and added for difficulty. The athlete competes against a standard, not against a human being. Bringing knockout rate into a taolu analysis sounds impressive but is meaningless. Applying taolu's deduction scale to MMA is meaningless in the opposite direction: a takedown is not a 'deduction', it decides the fight. Health risk between the two classes also differs in kind, not merely in degree. An MMA fighter accumulates brain injury through cumulative impacts, and the analyst needs knockout data, concussion counts and long-layoff history. A taolu athlete accumulates musculoskeletal injury from repeated landings, from lowering the centre of gravity, from rotating the torso at maximum range thousands of times a year. Both carry the martial arts label. But those two young people are wearing down through two different mechanisms and need two different intervention regimes. The third class, sanda, is the hardest because it is hybrid. There is contact, there are wins and losses, there are knockouts as in combat sports; equally there are throws, pushes off the platform, and a scoring system in which a clean takedown equals a clean kick. An analyst trained on MMA will misread the rhythm of a sanda bout: praising a neat hand combination while the fight was already decided in the distance battles along the floor. Extend outward and every remaining layer demands its own data. Institutionally, professional combat sports operate under state or national athletic commissions plus bodies that sanction title bouts. Traditional and performance martial arts operate inside national federation systems tied to multi-sport games with a limited number of entry slots. An informal challenge between two different martial groups often has no regulator at all. Three legal worlds, three levels of transparency, three different understandings of the same word, 'fairness'. Revenue follows the same pattern. Professional combat sports live on ticket sales, broadcast packages, and at the top tier, pay-per-view income. Taolu athletes live on national squads, equipment sponsorship and games quotas. The cost of a management error on the two sides cannot be compared: one side loses a contract, the other loses a four-year competition slot. This is where I have to pull the story toward what I enjoy most, because it shows the framework error reaches beyond academic debate. Between the pitch and the esports arena there is an invisible bridge, and I earn my living proving it is wobbling. In esports, rules on betting transparency trail the market's growth by several years. The result is that competitive integrity erodes far faster than in traditional sport. Apply that principle to martial arts: wherever the scoring system is opaque and the regulator is weak, integrity risk arrives earlier, not later. The transfer market is teaching the same lesson in a different language. There, noise drowns out signal, and most of that noise is generated by intermediaries with a direct interest in inflated valuations. In martial arts, the equivalent intermediary layer is the promoters and the match brokers. They do not lie in the sense of inventing events. They simply select the most favourable framing for their client and let the audience fill in the rest. Now I have to argue against myself, and I do it publicly because I have been wrong before and once wrote five thousand words dissecting that error. The argument above sounds tight, but it can still be an elaborate way of dodging. When we declare 'insufficient information to classify', we may be evading the responsibility to make a judgement. I once simulated the roar of the crowd for an empty stadium, and realised the loudest applause came from the data. That experience taught me that saying 'unknown' is a judgement, not an evasion — provided it is stated plainly instead of wrapped in impressive phrasing. The counter-intuitive point sits here: a wrong framework is worse than reduced accuracy. An under-informed analysis can be repaired once data arrives. A mis-framed analysis moves further from the truth with every additional data point. That is why I fear thick reports more than thin ones. And there is a professional habit worse than mis-framing: reading silence as safety. The absence of a risk section in a report gets understood as the absence of risk. In combat sports this is the costliest mistake of all, because what goes unrecorded is often precisely what is grinding a fighter down backstage. Not rating a person's risk does not mean that person carries no risk. It only means the analyst has not done the work. Cannot be assessed is different from neutral. The correct status of an empty file is 'unknown', and 'unknown' in combat sports is a warning, not a blank space to be filled with belief. Next time you read an analysis of any bout, try reading the first line and asking yourself: has the writer identified the discipline, or merely borrowed the handiest toolkit? If it is unidentified, the rest of the piece, however long, is decoration. One generation plays games, one generation watches football, and whoever stands between them sees them crying over the same thing — and all of it begins by naming correctly what we are looking at.