The Empty Dossier in Youth Table Tennis Scouting: Data Discipline and the Trap of the Hollow Conclusion

On March 12, in my office in Shanghai, a 41-page dossier was placed on my desk. Inside were photographs from four angles, three phone-recorded video files, notes from three training sessions, a hand-filled technical scoresheet, and on the final page, a "Conclusion" field left blank. My 24-year-old assistant asked: "What do I write here?" I said: "Write that we don't know yet." He laughed. I did not. After 39 years in scouting, I have learned that the most honest answer is sometimes an empty one. Not out of laziness, but because that dossier genuinely did not contain enough verifiable facts to support any judgment at all. The interesting part came four days later. A news item about the very player in that dossier appeared on three sports pages, describing him as a new discovery of his age group. Three pages, four days, and not one of them carried a match-source statistics table. The sediment layer of talent never lies on the surface. It lies deep beneath the strata people are unwilling to dig. CONTEXT: A MARKET THAT LIVES ON HEADLINES Youth table tennis is an ecosystem with its own rhythm. The International Table Tennis Federation's competition system and the WTT Youth series are divided by age category, running year-round at high density. Every tournament is a data point. Every data point is a fragment in the growth curve of an athlete — what I call the stratigraphy of talent. In Vietnam, youth table tennis grows mainly through gifted schools, provincial and municipal training centres, national age-group championships, and the school competition system. It is a broad network, but the mesh is uneven. Some places record every training session in detail. Others keep only the final match result. The gap between those two recording styles is where empty dossiers are born. I live in China, where youth table tennis is organised into a clear ladder: local training centres, provincial sports academies, national youth teams, and a top tier of athletes who have entered the international competition system. Observation density here is far higher than what I once knew at home. The paradox is this: the more raw data exists, the more dossiers end up blank in their conclusion section. The reason lies in time pressure. A youth tournament lasts five days. A selection camp lasts three weeks. A coaching staff needs answers within seventy-two hours. A report writer needs a summary line within twenty-four hours. In that compressed window, the easiest section to fill is always the impression, and the hardest is always the statistics. I once sat through a selection meeting that ran four hours, of which three and a half were spent arguing about feel and the final thirty minutes on cross-checking the scoresheet. The final decision was made on those thirty minutes. That reveals a rule: when data is not prepared in advance, it will always be handled last, and once handled last, it becomes decoration for a decision already formed. Breaking news is a shallow pit. Talent is an underground current. The annual season adds another layer of pressure. There is no long off-season to reload data. Youth tournaments follow one another, training camps follow one another, reports follow one another. Scouts are pushed into a state of continuous reaction, and in that state the empty dossier becomes the most convenient solution: fill in a generic sentence, attach a potential label, and move to the next match. That is where I drew my first principle. An empty dossier must be marked as empty. It must not be filled with adjectives. CORE ANALYSIS: THREE YEARS OF TAPE My foundational rule is simple and very hard to follow: do not form a judgment about a young athlete on fewer than fourteen fully watched matches spread across at least three seasons. Fourteen is a number I derived empirically. Below that threshold, week-to-week form variance overwhelms the real technical signal. Above it, the trend line begins to separate from the noise. Three seasons is the minimum time span needed to observe at least one injury cycle, one physical growth cycle, and one tactical change cycle within the athlete himself. The crowd looks at the screen; I look at three years of tape. The screen shows the most beautiful rally. The tape shows the most repeated rally. The two often contradict each other, and what decides an athlete's career is the second. In table tennis, I operationalise this rule into four metric groups that must exist before any conclusion is written. The first group is serve and receive metrics. I record point-win rate on serve by spin type, point-win rate on receive by stance position, and direct-error rate within the first two exchanges. This is the most stable group at youth level, because it reflects basic technique more than psychology. The second group is rally metrics. I record rally-length distribution, point-win rate in short rallies under four exchanges, and point-win rate in long rallies over seven exchanges. A young player who only wins short rallies is a player without a physical base and without the ability to construct a point. The third group is pressure-point metrics. I isolate point-win rate at 9-9 and above, win rate at deuce, and win rate in deciding games. This group has the smallest sample and the greatest risk of misreading, so I always attach the absolute sample size alongside the percentage. The fourth group is error metrics. I classify unforced errors by cause: footwork error, rotation error, placement-selection error, decision error. These four have four entirely different coaching implications, and merging them into a single "errors" column is the most common mistake in the scouting reports I have read. Three decisive metrics, no more. This is a discipline I imposed on myself after years of failing to persuade coaching staff with long data tables. A database is for storage. A report is for action. A report offering twelve metrics will cause the reader to act on none of them. So for each athlete I select the three metrics with the highest discriminating power between the prospect group and the average group, and only those three appear at the top of the report. The rest sits in an appendix, colour-coded by risk, and is opened only when there is a dispute. ANATOMY OF AN EMPTY DOSSIER Over the years I have classified empty dossiers into four types. The classification matters because each type requires a different handling, and mishandling the type leads to a wrong conclusion. The first type is empty due to missing data. The dossier has full forms but the numeric fields are blank because the recorder was not trained to count. This is the most common type at provincial youth level, where nobody sits and counts every rally. The correct handling is to state "no sample" explicitly and recommend adding a recorder, not to speculate from memory. The second type is empty due to unreadable data. The dossier has numbers, but they are inconsistent: total points won do not match the sum of game points, total matches do not match total minutes played. This is the most dangerous type, because it creates a sense of completeness. The correct handling is to cross-check at least two sources before using any figure. The third type is empty due to mislabelling. The dossier is tagged with an age group or event category that does not match the reality of the matches watched. When the label is wrong, every comparison against a benchmark is wrong too. The correct handling is to verify the label before analysis, even if that delays progress by a day. The fourth type is genuinely empty. No data, no video, no notes. Only a name and an introduction. This is the type the media likes most, because it permits writing anything. It is also the type for which I refuse to write a conclusion, even under deadline pressure. Of the four types, only the fourth is truly "no information". The other three are "information not yet processed". Distinguishing those two states is the boundary between a scout and an advertiser. Two old cases shaped my method. In 2026, when football pages in Shanghai reported that a 16-year-old was being pursued by a major English club for a rumoured 20 million euros, I was the only editor who did not approve the story. I stayed back and watched all fourteen of his national U19 matches, counting 23 shots with only 1 goal, a 64 percent pass accuracy, and 11 turnovers in his own half. I wrote a rebuttal arguing that the fever lacked data. The outcome: no club signed him, the rumour collapsed, and the newsroom recognised my method. In 2026 I stood outside the fever. Those who laughed at me then have stopped laughing. In 2026, on the strength of that episode, I was invited as an analyst for a television channel during the World Cup. While most viewers praised Kylian Mbappe at 19, I pointed to data from the European U20 Championship: only 1 goal in 3 matches, and a physical-output drop of roughly 18 percent in the second half. I stated that no one should pay over 150 million euros for a player whose consistency was unproven. Mbappe went on to score 4 goals at the tournament, won the trophy, and I received a great deal of mockery. On the night of the final I sat down with the tape and asked what variable I had missed. The answer was not technical. It was that I had applied a model built on an average sample to an individual outside the distribution. Since then every report of mine carries a mandatory section titled "What the data does not measure", and I write "based on available data" rather than "certainly". Mbappe only comes once. But the process that finds him repeats forever. THE DATA VAULT AND THE 0.08 PERCENT In 2026, when competitions were suspended, I was 49 and lost my commentary work. I spent more than 300 days building a database of 320 young athletes from 2026 to 2026, covering endurance metrics, injury frequency, and month-by-month form variance. When the whole world turned off the lights, I sat in the data vault and listened to the future fall. The cross-check forced me to adjust my own scale. Among those 320 athletes, the group able to sustain peak form across three consecutive seasons accounted for roughly 0.08 percent. Mbappe was in that group. In other words, my 2026 error was not in the data but in the absence of an outlier group to compare against. This is the lesson I want to carry intact into table tennis. In youth table tennis, an outlier group also exists, but it is smaller and appears later than in football, because table tennis demands a longer technical accumulation period before physique and psychology become decisive variables. That means: at U15 level, labelling someone a "top prospect" carries far lower reliability than at U19 level. Since 2026 my reports include charts, risk colour-coding, and a dedicated chapter titled "Data suggests, but reality decides". I have also shifted my focus from introducing athletes to describing scouting methodology. The reason is simple: a list of names goes stale in two seasons. A process does not. An error in the 2026 transfer window reinforced this. In June that year, a Shanghai club asked me to assess a 19-year-old Brazilian midfielder. Across 48 matches watched, he recorded 12 assists but showed 9 clear lapses in concentration after the 75th minute, expressed as reduced involvement in decisive phases. I reported that the metric needed another season of tracking before valuation. That report was judged too cautious. But what I learned from it had nothing to do with being right or wrong about the player. I learned that when a declining metric appears at the same time marker across many different matches, it is a systemic signal, not a random one. And a systemic signal at minute 75 in football corresponds to a systemic signal in the fourth game of a table tennis match. APPLYING IT TO YOUTH TABLE TENNIS Moving the four football metric groups into table tennis required changing the unit of measurement. Football's unit is the minute. Table tennis's unit is the exchange and the game. A physical-decline metric in table tennis shows through three signs. The first is reduced accuracy of the footwork initiation in the fourth game onward. The second is a rising rate of wrong placement selection in long rallies. The third is a rising footwork-error rate while the decision-error rate stays flat. These three signs are countable. And because they are countable, they can enter a report as figures rather than as remarks. For Vietnamese youth table tennis, I believe there are three data gaps to close first. The first is serve and receive data by spin type, which is barely recorded at domestic youth level. The second is rally-length distribution, which reveals whether a player is winning through technique or through opponents' unforced errors. The third is per-game data, which allows early detection of physical and psychological issues before they become habits. In China, where I work, those three gaps have been partly closed, but the problem has moved elsewhere: there is plenty of data, yet it is not standardised across centres. A "serve point-win rate" at one centre may be defined differently at another. When data is not standardised, cross-athlete comparison becomes meaningless, and the scout is forced back onto feel. This is why I always ask for the definition before asking for the number. A number without a definition is a number that cannot be verified. In the March 12 dossier I mentioned at the start, the "Conclusion" field was left blank for three specific reasons. First, only four matches had been fully watched, below the threshold of fourteen. Second, no statistics table accompanied the video, so every technical remark lacked a source. Third, the age-group label on the dossier did not match the age recorded in the training notes. Those three reasons belong to three different empty types: missing data, unreadable data, and mislabelling. For a dossier like that, writing "prospect" is fabrication. Writing "no prospect" is also fabrication. The only honest option is to write that there is not yet sufficient basis. THE CONTRARIAN ANGLE: A BLANK FIELD IS NOT A VERDICT There is a misreading I encounter constantly in selection meetings. When a dossier is marked "insufficient information", people read it as "no risk". When a dossier is marked "high risk", they read it as "thoroughly checked". Both readings invert the label's true meaning. An empty dossier is the dossier with the highest degree of uncertainty in the entire set. It should not be grouped with dossiers that have been verified and returned a neutral result. Treating the two as equivalent is the most serious error in a decision process, because it makes the least-monitored athletes the easiest ones to select. There is a second, subtler paradox. In many systems, a document with a formal title, tables, and a table of contents will be read as a conclusion. When that document is only a structural framework not yet filled with content, readers still tend to believe it has been filled. The effect is strong enough to be dangerous: a structurally complete but hollow analysis can appear more credible than a short analysis containing real figures. I have imposed a rule on myself to counter that effect. If a document has more than one third of its sections marked "insufficient information", it must be titled "Insufficient-Data Notice" and must not be presented as an evaluation report. Correct titling does not solve the data problem, but it prevents the document from being misread. The third contrarian angle concerns the Vietnamese and Chinese contexts. I was born in Vietnam and have only four years working directly in the Chinese market. Four years is far too short to apply an old model to a new ecosystem. Before each judgment, I ask myself: what rhythm does the training ecosystem here operate on, and does that rhythm match the model I am using. The differences lie in three points. Competition density in China is higher, so samples accumulate faster. Internal competition within each age group is fiercer, so form variance is larger. And the lifecycle of a young athlete inside the system is shorter, so the cost of waiting one more season is greater. These three points mean my fourteen-match threshold must be adjusted per system rather than applied mechanically. At the same time, I must resist another habit of being 55: believing a validated model needs no updating. I schedule a quarterly cross-check of the old database against new development rhythms, and I record every instance where my model predicted wrongly. Those failures are kept in a separate section, never deleted. A model with no failure record is a model that cannot improve. "The next Mbappe" is a language trap. When people call a young athlete the next version of an established star, they replace an evaluation process with a comparison. A comparison does not generate data. It only generates expectation, and expectation cannot be measured by any metric. TAKEAWAY: WHAT I WILL WRITE IN THE BLANK FIELD In the March 12 dossier, the "Conclusion" field was eventually filled, but not with a prospect label. It was filled with three lines: the number of additional matches required, the three metrics to be added, and the expected completion date. Those three lines turned an empty dossier into a work plan. I believe this is a shift Vietnamese youth table tennis can make immediately, without waiting for major resources. Instead of asking "does this athlete have potential", ask "how many more matches and how many more metrics are needed to answer that question". The second question always has a concrete answer, and that concrete answer is the only thing that can be verified three years later. Data suggests, but reality decides. The scout's job is to make the suggestion strong enough that the decision no longer has to rely on luck.