When the Spreadsheet Is Empty: The Discipline of a Table Tennis Analyst

On the third night of a WTT Contender event in Beijing, I opened a forty-two-page spreadsheet the organisers had sent over. Every cell in the point-by-point sheet was blank. No serve-win rate, no efficiency figures after a forehand loop from mid-distance, no reaction time after a short receive. I checked the file three times, checked the server, called the colleague in charge of collection. The result did not change. After years in this trade, it was the first time an empty file left me more relieved than a file stuffed with hastily patched numbers.

People picture table tennis analysis as a craft of curving charts. In reality, most of my hours go into verifying whether the data is real. A single WTT-level match generates thousands of data points: serve timing, spin type, placement, number of direction changes, each player's standing distance after the third ball. All of it has to reconcile across at least two independent sources, the umpire's record and the semi-automated video log, before I allow myself to draw a conclusion. When one link breaks, I do not fill the gap with memory. A spectator's memory is a bad editor: it keeps the beautiful loop at 9-9 and forgets the twenty botched rallies before it. Empty data is an honest answer; data filled with guesswork is a polite lie.

The data architecture of modern table tennis flows through three layers. Upstream sits equipment and officiating: vibration sensors on the table, high-speed cameras, electronic scoreboards. The middle layer is the WTT system and national associations, where raw data is normalised into ranking points, head-to-head records, performance indices. The downstream layer is where I sit: broadcasters, news sites, coaching staff, independent analysts. When the upstream layer breaks, the entire chain behind it loses its anchor. Without a point-by-point log, I cannot compute an expected-value index for a serve, cannot measure the performance drop in the seventh game, cannot compare two playing styles. All I have left is memory, which cannot be verified.

The history of table tennis shows that every reform arrives with an argument about data. In 2026, the ball grew from 38 to 40 millimetres, cutting speed and spin and forcing fast-attack players to rebuild their entire rhythm. In 2026, matches moved from 21 points to 11 points per game, shortening each rally and raising the value of the opening points. In 2026, the hidden-serve rule stripped an advantage from a generation of heavy-spin servers. In 2026, speed glue was banned, and in 2026, celluloid gave way to plastic. Each time, the central questions were the same: who gains, who loses, and what consequence shows up in the match data.

Every reform leaves its own trace on the chart. When the ball grew larger, the number of rallies lasting more than five strokes rose, the win rate of far-table defensive play crept up, and the value of the loop dipped slightly. When games shrank to 11 points, the win rate of a player leading 5-1 became more dominant, because the remaining points were not enough to reverse it. When hidden serves were banned, the rate of points won directly from serve fell sharply, and the receive became the skill that separated players. None of these conclusions can be born from a spectator's feeling. They emerge only when you hold a continuous data series, long enough, spanning many seasons.

That is why I treat complacency lightly. If the organisers send me an empty file, the only thing I am allowed to do is stop and state clearly that I do not yet have enough evidence. There is no version of events in which I read a short press release, hear a commentator say player X is declining, and then build an analysis of that decline. What I protect is not my reputation, but the reliability of every number standing behind it. If I compromise once, every other figure I publish loses value.

There is a gap I meet constantly: the style label and the actual execution. Media brands a player as a fast attacker, yet the point-by-point log shows that player's loop rate unusually high on deciding points. The label exists because it is easy to remember, not because it is right. To strip it off, I need data by point, by spin type, by standing zone. Without that, I am merely repeating a prejudice that already existed.

I follow many events and write every rally into a notebook of my own. Those lines are not pretty, they carry no charts, only strings of numbers and abbreviations. But when an electronic file is empty, that notebook reminds me that clean data does not generate itself. It is made by someone who sat long enough to see things repeat, then recorded them exactly as they happened.

Deeper down, a data void exposes a temptation inside this craft. When evidence is missing, people readily turn correlation into causation. A player wins three straight matches after changing rubber, and a story of miraculous transformation is born. But three matches is a far too small sample. Weaker opponents, a softer schedule, a minor injury on the other side — there are countless other variables enough to explain the result without invoking the blade. Correlation does not judge; it simply waits for us to assign it a convenient cause.

In the case of that empty file, my instinctive reaction was to go looking for a story. The temptation was strong enough that I had to remind myself that an analysis made entirely of blank cells still holds more value than an analysis full of words with no root. Readers do not need another piece asserting what everyone already guesses. They need to know what is certain and what is not. The line between those two halves is the whole substance of the trade.

That honesty is not free. It costs me quick articles, eye-catching headlines, confident endings. But it keeps my data series usable for many seasons to come, instead of burning out within a week.

I still keep that empty file in a folder of its own. It reminds me that an analyst's credibility does not rest on always having something to say, but on knowing when to stay silent and wait for the data to arrive.

The next round will open, and the data will flow again. The question I carry is not who will win the title, but what the next file will show me that the naked eye skips over, and whether I have the patience to verify it against two sources, rather than once again telling a story that merely sounds plausible.