The Empty Analysis: Nine Data Layers of Table Tennis and the Value of a 'Not Enough Information' Verdict

THE EMPTY ANALYSIS: NINE DATA LAYERS OF TABLE TENNIS AND THE VALUE OF A 'NOT ENOUGH INFORMATION' VERDICT

A result file with nothing to read

At 22:40 on August 13, 2026, in a nineteenth-floor apartment in Jing'an District, Shanghai, I ran the last analytical pass of the day. The machine returned exactly the format I had asked for: a title, nine sections, tables, bullet points, an inferred-information block, a risk-flag block.

In twenty-three of twenty-five cells, the same sentence repeated: not enough information, cannot be assessed.

No player to dissect along an age curve. No tournament to position inside an Olympic cycle. No serve-rule change to trace. No foreign-match win rate, no deciding-game record, no head-to-head table. The information-points field from the first-stage deconstruction was completely empty, and the entities field was empty along with it, because that field is designed to be drawn from the very list that was missing.

My usual editor messaged three words: anything yet? I answered with a ratio. Twenty-three out of twenty-five. He sent back a question mark. I replied: there is nothing yet. And I will not write otherwise.

In this profession there is one kind of draft nobody wants to file: the draft that says nothing can yet be said. It has no appealing headline, no character, no climax. It is also the only honest draft the input dataset permitted. The rest of this piece explains why.

Why table tennis is hard to read through numbers

Modern professional table tennis runs on a dense year-round calendar, a tiered event system, and a rolling points mechanism built on a fifty-two-week window. Every week, old points fall out of the window and must be replaced with fresh results. That mechanism produces what I call points-defence pressure: a player who loses nothing at all can still slide down the ranking, simply because the points he was defending have expired.

That is the first structural difference from team sports. In football, a thirty-eight-round season produces a sample large enough that most metrics stabilise. In table tennis, a match ends after three to seven games, each game to eleven points. An eleven-point game is a very small sample. Two balls travelling the wrong way at 9-9 can reverse the entire outcome, and the full data record of that match will then describe something that does not fully correspond to the true form of either player.

My job is to separate those two things. Based on my experience tracking hundreds of matches across different event tiers, I hold one rule: the smaller the sample, the more the number must be read with an interval attached, and the more cautious one must be about conclusions that sound extremely coherent.

There is a further difficulty that is rarely discussed. Table tennis is a sport whose technical variables are far harder to measure than they appear. Spin has no universal unit. Racket-head speed at the contact point is not published consistently. Ball height at the bounce, trajectory length, dwell time on the rubber — all of these are valuable data, but they exist only inside a few capture systems, and those are usually closed to the public. The result is that most published table-tennis analysis must reason from coarse indicators: scorelines, game counts, point counts, win streaks.

Reasoning from coarse indicators is legitimate, provided the writer states clearly what he is reasoning from. It becomes irresponsible when the writer drapes it in the tone of a precise measurement.

The nine layers any table-tennis analysis must pass through

The framework I use has nine layers. They are not nine sequential steps but nine cross-sections of the same object. A good piece usually touches only three or four of them, but the writer must know what the other seven are saying — or refusing to say.

The technical, tactical and equipment layer. This is where a match is actually decided. The basic unit of analysis is the first three shots: serve, receive, and third-ball attack. The whole shape of a game is usually set within the first two seconds of each point. To assess a player at this layer I need the point-win rate on the first three shots, the win rate when serving short versus long, and the rate at which he is attacked first after receiving. Without those figures, every remark about a 'good serve' is only an impression.

The rule prohibiting hidden serves has been in force since 2026. That is an important marker when tracing technical history: it pushed the value of the serve away from concealment and towards readable variation in spin and placement. Any analysis of the generations before and after that marker which ignores the shift has a foundational error.

Equipment is also a variable, and the most underrated one. Rubbers, blades, sponge thickness, and the regulations around adhesives have all changed across eras. When a player switches equipment configuration, an adaptation period is required. It usually lasts weeks to months, and during that window every form comparison against that player's earlier self is distorted. An analysis without an equipment-change timeline is an analysis reading the wrong variable.

The player data and head-to-head layer. Here I need four groups of numbers. The first is ranking position and the structure of points being defended: how many points expire within the next eight weeks, and what pressure that places on the competition schedule. The second is the foreign-match win rate — matches against opponents from other national associations. This is the core indicator for measuring a player's true strength on the international stage, because it filters out the large block of domestic matches whose difficulty is on a completely different level.

The third group is the record across the three biggest events: the Olympic Games, the World Championships and the World Cup. When a player wins men's or women's singles at all three, it is called a Grand Slam. The term exists because those three events carry a pressure structure unlike the annual circuit: higher density of strong opponents, shorter recovery windows, and amplified psychological error. The fourth group is performance in deciding games and at key scores. A player may win sixty-five percent of his matches but only forty-five percent of his seventh games. That gap is the whole story about him.

Head-to-head records must be split by period, never merged. The last two years say one thing; the record at the three majors says another; and if a specific opponent keeps winning despite a lower ranking, we have a datum that must be named correctly: a nemesis.

One more indicator is rarely written correctly: the age curve. A table-tennis player's peak does not coincide with peak physical capacity, nor with peak experience. It sits where reflexes are still fast enough and reading of the rally is sharp enough. Misplacing a player on that curve is the most common cause of forecasts that go wrong within an eighteen-month horizon.

The event system and points rules layer. Each event carries a different value, and that value shifts depending on its position in the four-year cycle. A title at the highest tier brings ranking points, prize money, and — most importantly — a seeding position at later events. Seeding determines draw difficulty. Draw difficulty determines the number of games a player must play. Games determine physical depletion. That causal chain is long, and every link in it is measurable.

Draw structure deserves more serious analysis than it usually receives. The same player, in the same form, can reach a semi-final via three light matches or via three five-game matches. The difference shows up most clearly in the later rounds, when the legs are no longer as quick as they were in the first round.

Prize-money distribution also deserves its own reading. When prize money is heavily concentrated in the final rounds, financial pressure forces players in the middle group to enter more events to maintain income, and entering more events raises injury risk. That is a traceable causal chain, not a complaint.

The competitive landscape layer. To describe the balance between one dominant table-tennis nation and the rest of the world, three minimum indicators are required: seats inside the top ten, titles at the majors across the last five editions, and the depth of the under-twenty-one cohort. The third is the least mentioned but carries the greatest long-term weight. A nation can be dominating at the senior level while its junior pipeline has already run dry. The lag between those two facts is usually one Olympic cycle.

The threat level of an opponent must be described with three parameters: who threatens, how they threaten, and how long the threat window lasts. An eighteen-year-old improving fast is a threat across an eight-year window. A thirty-one-year-old at his peak is a threat across an eighteen-month window. The two demand entirely different responses.

There is also a regional variable at this layer. National associations do not progress at the same speed, and that speed depends on domestic league density, the number of youth training centres, and how open the competition market is. Comparing two associations while ignoring those three factors is comparing two sets of numbers taken with two different rulers.

The rules and governance layer. This is the layer fans care about most and where public data is thinnest. Every competition-rule change creates beneficiaries and losers, and identifying those two groups is a solvable problem given sufficient historical data. Reforms to the number of points per game, to the interval between points, to the number of tactical time-outs, to qualifying formats — each shifts the distribution of advantage in a determinate direction.

In selection matters specifically, controversy tends to erupt at the intersection of quantitative standards and human discretion. When a major-event place is awarded on internal ranking, the public accepts it. When it is awarded on the coaching staff's assessment, the public questions it. The correct handling is not to choose one over the other, but to publish the weight of each criterion in advance.

The coaching staff and talent pipeline layer. A national team must answer three questions: does the head coach have enough authority to make hard decisions, does the personal coach fit the player's technical profile, and is the staff stable enough to see through a four-year cycle. None of the three is measurable through public indicators, yet all three leave traces in competition data: the number of changes in leadership, the frequency of equipment reconfiguration, the degree of playing-style change between two consecutive seasons.

The pipeline is more measurable. The central question is conversion efficiency: how many players from the junior group step into the senior group, and how long it takes. If a nation keeps promoting newcomers early while their retention rate stays low, that is the signature of a pipeline that is wide but shallow.

The risk surface layer. The risk list of a professional player is longer than outsiders imagine. There is injury risk, particularly to the shoulder, wrist, lower back and knee. There is risk from an unfinished technical overhaul. There is risk of being countered by a specific opponent archetype. There is risk from playing too many events inside a short window. There is selection risk. There is generational-gap risk. There is governance risk and public-opinion risk. There is systemic risk, when an entire generation is trained to one template and neutralised by another.

In that night's output, the entire list sat empty. There was no subject against which to screen risk. The only identifiable risk was at the process layer: the input was empty, so any downstream conclusion would be a product of imagination rather than inference.

The public narrative and expectation layer. Every player, at every moment, has a story being told about him. That story can be measured on three parameters: whether it has a data foundation, how large the sample behind it is, and how long it can plausibly last. A story built on three matches has a short life. A story built on two seasons lasts longer and deserves a place in the model.

In this sport, public opinion tends to polarise fast. After a big win, a player is described with adjectives that cannot be measured. After a loss, the same player is described with adjectives that also cannot be measured, only with the opposite sign. Both directions generate noise, and noise always has a cost: it pulls the expected price away from the true value, causing anyone deciding on that expectation to misprice the probability.

The gap between market expectation and objective assessment is where informational value appears. When the crowd expects an outcome at seventy percent while the model returns forty percent, that thirty-point gap is the entire value a data journalist can deliver.

The industry transmission layer. Finally there is a long chain from upstream to downstream: equipment and youth development at one end, the event system and associations in the middle, broadcasting, commerce and derivative markets at the other. A change upstream takes years to surface downstream. A change downstream can feed back upstream within months, usually through capital flows and through the decisions of parents weighing a professional path for their children.

None of those links appeared in that night's dataset. No equipment brand, no event, no capital flow, no policy.

The counter-intuitive point: an empty verdict is still a verdict

The default industry reaction to an empty dataset is to fill it. The pressure is real: pages need copy, readers need content, and a piece stating that no conclusion is yet possible generates no pageviews. Our profession has cultivated an entire ecosystem that manufactures content out of fragments, and most of the time it works smoothly.

But there is one class of error that ecosystem can never correct on its own. When a writer fills a blank with speculation, then another reader cites that speculation as a fact, and then a third builds a model on the distorted fact, the error has entered the structure. It is no longer a wrong article; it is a false datum that has been granted citizenship.

I have seen the opposite happen, and that is why I keep this discipline. In 2026, before a group-stage match at a football World Cup, I built a model on the retreat speed of the defensive line and the number of sprints above twenty-five kilometres per hour. The model produced an expected-goals figure leaning heavily one way, but the probability of that side losing reached twenty-two percent because the centre-backs were pushing too high. I wrote the piece and was mocked. The final result stunned the world. When the naked eye sleeps, the data stays awake — and it had already seen.

That shock, in the end, was no shock. It was simply the first time the number was listened to.

Yet precisely because I have been right that way before, I must be more careful about the next time I am right. A correct forecast produces a dopamine hit that is hard to resist, and that dopamine is the analyst's enemy. The test I set myself after every occasion the model matches reality is to rerun the scenario assuming the ball rolls the other way. If the conclusion holds, I wrote it correctly. If it flips, I was only narrating the result after learning it.

The same standard applies to that night's empty analysis. Had I filled the nine layers with speculation, I would have had a three-thousand-word piece that read very professionally and contained not a single defensible fact. I chose the opposite.

One further principle must be stated, because it is the most common error in table-tennis analysis. Correlation and causation are different things, and in a small-sample sport like table tennis they are almost always mixed. A player changes rubber and then wins five matches in a row: that is a correlation, not yet a causal relation. A team wins more when one individual sits out: that is a correlation until opponent quality, schedule and physical condition are controlled for. The writer's job is to label correctly, not to tell the most coherent story.

I write drily, so that the game we love is not buried by sentiment.

Signals to track in the next round

An empty verdict is not an endpoint. It is a waiting state, and a waiting state always has exit conditions.

The first condition is the existence of a complete information-point list, in which each item can be cited by number so that every conclusion attaches to a specific source. This is the strictest requirement and also the easiest to skip.

The second condition is the identification of at least one named entity: a player, a national association, or an event. No entity, no analysis.

The third condition is an assessment of source quality and time sensitivity. These two fields are usually treated as procedural, but they are the reliability anchor for everything else. Without them, every conclusion floats.

Once those three conditions are met, the nine layers open automatically. The technical layer will have first-three-shot data. The player layer will have points-defence pressure and foreign-match win rate. The event layer will have draw structure. The competitive layer will have junior depth. And the industry transmission layer will have capital flows to trace.

What I want to leave behind is not a complaint about an empty data file. It is an observation about value: in an information market where anyone can say anything about any player, the capacity to say 'not enough data yet' is a competitive capability. It generates no pageviews today. It generates credibility over the next ten years.

The value of a conclusion in table tennis lies not in how compelling it sounds, but in how long it holds before new data arrives. And if the new data has not arrived, the only way to hold is to wait.