The Blank Scoresheet: The Hardest Lesson in Table Tennis Analysis

The A4 sheet landed on the table at 8:47 p.m., after the evening session of a table tennis club in District 7, Ho Chi Minh City. On it was a scouting file on the opponent's last three matches: 72.4% of service points won, 48.6% of points won on the opponent's serve, 61.2% third-ball attack efficiency. The figures were presented so neatly that nobody bothered to ask how many points produced them. I counted. Eighteen service points behind a percentage carried to the decimal.

With eighteen observations, the margin of error is wider than the gap between two playing styles. A serve that clips the edge, a rally lost to a phone ringing outside the court, one questionable umpiring call — each of those shifts the percentage by nearly three points. The file speaks loudly, but what it says loudest is what it lacks: an adequate sample. The number knows how to hold its breath, and I wait for it to exhale.

This situation returns every season. The coaching staff needs a conclusion before tomorrow morning's session. The analyst needs a sample thick enough to conclude. Those two needs rarely meet, and most of my working life is spent standing between them, keeping the sheet from being filled with things that merely sound reasonable.

Two layers of record-keeping in one sport

Table tennis has an obvious paradox for anyone who works with data. At the administrative layer, the sport is measured rigorously. The International Table Tennis Federation (ITTF) calculates rankings from a player's best eight results over the most recent 52 weeks, with points expiring weekly. The WTT system, running since 2026, tiers events into Grand Smash, Champions, Star Contender, Contender and Finals, generating an enormous stream of administrative data: points, seedings, schedules, prize money, entry slots. Anyone can look it up, and because anyone can look it up, people assume they understand the match.

At the rally layer, the gap is far wider. A top-level match lasts under an hour and contains roughly seventy to a hundred points. Each point is a chain of decisions made in two or three seconds: the toss, the placement, the spin, the tempo, the read on the opponent's footwork. Recording that chain requires multiple camera angles, high-speed capture and a human coding every rally. At most domestic Vietnamese events, the venue has one camera and a scoresheet. The distance from a scoreline to technical behaviour is enormous, and that distance is where unsupported conclusions live.

Based on my experience watching domestic matches, a national-level match usually ends with fewer than two hundred usable rows of raw data — if anyone was recording. With nobody recording, the number is zero. And when the number is zero, the market generates something else to fill the space. Rumour is always in stock.

The transfer window is the clearest example. Every season, clubs add players, change rubbers, replace coaches, and the news travels far faster than anyone can verify it. Did a player move for money or for a guaranteed starting slot? Did a club sign a young athlete for potential or to fill a team-event registration? Without rally-level data, people have to guess with their gut, and the gut is always available. Vietnam's leading players, such as Đinh Quang Linh or Nguyễn Anh Tú, still compete at events where, after the final applause, nobody knows how they won a single point.

Nine layers of a file, and what happens when every cell is blank

My work starts with a nine-layer sheet I impose on myself so I do not jump straight to conclusions: technique and equipment; player data and head-to-head records; event system and ranking rules; competitive landscape; rules and governance; coaching staff and talent pipeline; risk surface; public narrative and expectations; and industry transmission.

Technique and equipment is the layer most sensitive to error. Normally I read serve-placement structure, rally-length distribution, third-ball attack rate, off-table defensive rate. These indicators are equipment-sensitive: since 2026, ABS plastic balls of 40mm and above have replaced celluloid, reducing spin, lengthening rallies, and stripping direct comparability from every earlier dataset. When the cell is blank, I cannot separate a miss caused by technique, by fatigue, by a freshly glued rubber that has not settled, or by an opponent who has read the intention. Three causes, three different remedies. One blank cell, and every remedy might be wrong.

Player data and head-to-head records is the most abused layer. Head-to-head records are quoted like prophecy, even though two players rarely meet more than a handful of times. A three-from-four record says nothing about the fifth meeting, especially after one side has changed style or changed rubbers. Rankings behave the same way. ITTF points expire weekly, so a player dropping a few places may simply be losing old points rather than form. Reading it correctly means separating expiring points from recent results, and knowing whether the player competed at all in that window.

The event system and ranking rules determine match density; density determines accumulated fatigue; accumulated fatigue determines rally quality at the closing points. One macro anchor for scale: from Seoul 2026 to Paris 2026, Olympic table tennis awarded 42 gold medals, of which China took 37, spanning successive generations from Ma Long to Fan Zhendong. That figure helps nobody preparing a group-stage match at a domestic event. It only reminds us that every macro conclusion is built on thousands of rallies, not on the last three matches.

The competitive landscape in Asia shows most clearly in WTT qualifying, where the number of players per country who survive qualifying is a fairly stable indicator across seasons. When the data is blank, the landscape gets ranked by feeling: whoever won recently is remembered, whoever appears rarely is undervalued.

Rules and governance is the layer that fractures data series. Table tennis has been through several rule changes: the 40mm ball, 11-point games replacing 21, the 16cm toss and no-hiding rule, then the plastic ball. Each change opens an adaptation period in which old numbers cannot be compared with new ones. An analyst has to mark the fracture date on the sheet, or end up comparing two things measured with different rulers.

Coaching staff and talent pipeline is the layer most easily fooled by a single moment. A junior title does not mean the pipeline is healthy. It means one individual was excellent in one short event. The right indicator is conversion rate: over three consecutive years, how many juniors reached the national top eight, and how many of them stayed there. Without that data, every claim about a golden generation is a photograph of one moment enlarged into a trend.

Table tennis has three familiar injury clusters: wrist and shoulder from spin volume, knee and ankle from lateral movement, lower back from the bent posture. Add the risk of being read tactically and the risk of unsettled equipment. With injuries, the return timetable is controlled by the team, so a line about waiting until the weekend usually means the injury has not healed. An analyst has no licence to diagnose, but does have the right to check the announced return date against the points actually played afterwards.

Public narrative and expectations is the layer that runs faster than data. After a win, a narrative forms within 24 hours and outlives every spreadsheet. A player who wins three matches is called inspired, when the data may show all three opponents ranked outside the world's top hundred. The crowd always needs a story before it needs a number.

Industry transmission is the last layer, and the one that pays for every gap above it: the equipment market, training bases, a player's commercial value, capital flowing into events. Without rally-level data, a player's commercial value is set by relationships and instinct. That is why the biggest sponsorship deals go to whoever won a major event, not to whoever posted the most stable numbers across seasons.

When all nine cells are blank, the only honest output is a short sentence: not enough data. Saying that in front of a coaching staff waiting for a conclusion is the hardest part of the job. My data café is busiest when the stadium is empty. Not because people are curious about numbers, but because when there is nothing to watch, they finally sit down to hear why no conclusion can be drawn yet.

The counter-intuitive angle: the problem is not missing data

Most people in the industry believe the answer is more collection. I read it differently. The real danger is not the gap, but the gap filled with guesswork. An empty cell harms nobody. An empty cell filled with inference does harm, because it wears the shape of evidence, and nobody checks it again.

In 2026, I used expected goals to show that a team's winning run was luck: actual goals exceeded expected goals by nearly five. The piece was widely shared, and I began to trust my own reflexes. A year later I went on live television and mispronounced a Russia striker's name three times in the first half. I did not lack data that day. I lacked verification, and I filled the gap with excitement. Old footage is a mirror, and only those who dare look see themselves.

That lesson changed how I work. Before every conclusion I ask three questions: about sample size, about the comparison group, and about the date the data was collected. In table tennis I set myself a threshold: I will not quote a percentage computed from fewer than fifty points. A rule that simple eliminates most of the files circulating in domestic competitions, including the best-presented ones. The crowd watches the scoreline; I watch the serve that got forgotten.

The counter-intuitive point is this: data discipline is not a technology skill, it is the discipline of refusal. Refusing conclusions, refusing pretty numbers, refusing the pressure to answer before training. I once feared the microphone; now I let the data speak for me. And when the data has nothing to say, my job is to keep quiet on its behalf.

What to watch in the next cycle

The signal for the next cycle is not on the ranking list. It is in the recording infrastructure. The club that publishes point-by-point data with its sample size is building a foundation. The club that publishes only scorelines is still buying stories. Over the next few seasons, the advantage in Vietnamese table tennis will not belong to the team with the fastest player, but to the team that asks better questions of a sheet still full of blank cells. What is missing is not new software, but a generation willing to say "not enough data" in front of an audience waiting for a number.