Table Tennis and the Fight Against Ownerless Numbers

Hook

On the morning of August 13, my inbox received a three-page statistical table about an international table tennis tournament that had just ended. The layout was clean, the figures neatly rounded, each cell colour-coded by importance. But when I traced the source note at the bottom of the page, I found only blank space. No tournament name, no collecting body, no timeframe. Three pages of data, not a single anchor to verify.

For anyone who works in analysis, that is a familiar moment. A table beautiful enough can make us forget the first question: where did it come from. In table tennis, where advanced statistical systems are far thinner than in football, that question becomes a matter of survival. When the source is blurred, every conclusion built on it is as fragile as the second floor of a house without foundations.

Context

I began following table tennis seriously in 2026, when I helped produce coverage for several major events, including the Table Tennis World Cup. The work forced me to record every rally, every point, then cross-check against the organisers' electronic board. That period taught me one thing: table tennis is a sport that generates data constantly, yet records it haphazardly.

A singles match lasts 40 to 50 minutes on average and produces hundreds of points. Each point is a combination of speed, spin, placement and tactical choice. Counting only total points won discards almost the entire story. The problem is that, unlike football with its expected-goals metric, table tennis has no widely shared statistical standard among federations. Each event collects data its own way, each broadcaster draws its charts in its own style. The result is a scatter of figures with little capacity for comparison.

According to data published by the International Table Tennis Federation in its most recent annual report, events with detailed point-level statistics account for fewer than one third of all events in the system. The rest provide only match and set scores. That number tells us that most of the table tennis story remains blank at the data layer.

Amid that situation, a reporter easily falls into two traps. One is trusting absolutely whatever figure is handed over. The other is inventing a figure to fill the page. Both end the same way: readers receive information that looks precise but cannot be traced.

Core

My first principle when handling table tennis data is to establish a chain of evidence. A figure deserves trust only when I can answer four questions: who collected it, how, over what period, and whether it has been independently verified. Miss one of the four and the figure should appear only as a hypothesis, not a fact. Numbers never lie; only the people reading them deceive themselves.

Take the point-win rate on serve. It is the most basic metric in table tennis, yet its definition varies. Some count points where the opponent misreturns. Others count only points won directly from the serve. The two methods produce figures that differ by 10 to 15 percentage points. Without a clear definition, any comparison between two players is meaningless.

My experience watching matches live reveals a paradox: the flashiest metrics are often the least valuable. A player winning 11-4, 11-5, 11-3 looks dominant on the scoreboard, but if the opponent self-destructed on 20 of 30 lost points, that win says little about class. Conversely, a 12-10, 13-11 win after saving four match points carries information about nerve that no summary sheet records.

So when I analyse, I always separate two layers of data. The first is the results layer: point scores, set scores, match scores. This is the easiest to verify because organisers publish it openly. The second is the process layer: point structure, point sequences, the moments a player loses rhythm. This is harder to verify, but it holds the real analytical value. A figure in the second layer should enter an article only when I have counted it myself or have footage to cross-check.

Once I tested a tracking sheet of my own for a domestic event. I recorded every point of 24 matches, classifying how each point ended: service winner, forehand loop, backhand loop, counter-loop, opponent error. The result was surprising. Among younger players, the share of points won from opponent errors was markedly higher than among veterans. In other words, the younger players won more through their opponents' mistakes than through actively creating points. This is the kind of information a pure scoreboard never reveals.

But to obtain it, I paid in time. Twenty-four matches at roughly 45 minutes each, plus recording and cross-checking, consumed nearly a month. That is why most table tennis coverage stops at the results layer. No one has the resources to recount every point for every event. And that very gap creates the opening for ownerless numbers to multiply. Data is the confession of those who once trusted feeling.

When I must use an unverified figure, I choose to state its confidence level clearly. A metric labelled estimated from footage is entirely different from one labelled published by the organisers. That labelling does not make a piece less engaging; it makes it more trustworthy. Readers have the right to know what ground they are standing on.

Contrarian

Here a counter-intuitive view appears. Many assume that advancing technology will automatically solve the data problem. High-speed cameras, sensor systems, automatic point-scoring artificial intelligence. It sounds reasonable. But technology solves only the collection stage, not the verification stage. A sensor system can misrecord, a recognition algorithm can confuse one stroke with another. If readers have no way to trace it, a machine-generated figure is as ownerless as a figure a person invented.

The deeper issue lies in psychology. When an article has an information gap, the pressure to fill it is enormous. Writers fear their piece looks thin. Readers are used to every claim carrying a figure. So the gap gets filled with numbers that sound plausible but have no basis. No one deliberately lies. It is simply that no one is willing to leave a blank space alone.

I have been in that position. Once, wanting my piece to look rigorous, I nearly cited a metric from an unclear source simply because it matched my argument. I stopped at the last minute. The article was published with a sentence admitting the necessary data had not been fully verified. Its readership was far lower than others. But that was the moment I understood that staying honest with data matters more than keeping up a flawless surface. Reputation can vanish in a night, but a data model endures.

Takeaway

Vietnamese table tennis is at a stage where deep data can create a genuine competitive edge, both on the court and in the press. But that edge is worth something only if it is built on traceable figures. The question practitioners must ask is not how much data we have, but how many blanks we dare to leave when verification is impossible. A table tennis scene mature in analysis will be one that dares to admit its own limits.