The 68% Dossier With No Footprints: Lessons From an Untraceable Statistic

On September 14, 2026, an eleven-page dossier on a leading Vietnamese women's singles player landed in my inbox in Jakarta. Page three carried one bolded line: “Win rate in rallies over 20 shots: 68%.” Over the following three weeks that line appeared in at least nine news items, two internal sponsor dashboards and one transfer proposal sent to a Japanese club. I reopened footage of the last 22 matches, took out my notebook and started counting rallies one by one. I needed to know at which minute that 68% was born, at which rally, against which opponent, and who pressed the stopwatch. After four days of counting, I found its footprints. There were none.

I work as an auditor of player valuation dossiers in the transfer market. Over the past four years the flow of Southeast Asian women's badminton has visibly changed direction: clubs in Japan, Korea and Indonesia recruit more Vietnamese players than before, and every contract comes with a dossier prepared by the representation side. Nguyen Thuy Linh and Le Duc Phat are the two Vietnamese names that appear most often in the region's tracking sheets. My job has three steps: find the denominator, find the time window, find the source.

A number only means something when you know how many rallies it was calculated on, across how many minutes, and in which phase of the match. In football, movement data is standardised to the point where every pass carries coordinates. In badminton, most statistics still live at the broadcast layer: a graphic appears between games, nobody records the denominator, and it disappears when the match ends.

The path of that 68% makes the point. It began with a television graphic during a match in July 2026, was screenshotted, posted to a fan aggregation page, then picked up by a news outlet, then placed in an agent's dossier. Across four rounds of copying, the denominator vanished. What survived was 68%.

What matters is that contract value was negotiated on top of what survived. The buying club asked me one question: is this number real. I answered that I could not yet conclude, and in my line of work that answer costs more than any assertion.

I rebuilt everything by counting by hand. Across the player's last 22 matches, covering domestic competition and three international events, I logged 1,914 rallies, an average of 87 per match. Each rally was tagged with shot count, server, point outcome and match minute. I split every match into three windows — game 1, game 2, game 3 — and within each game isolated the stretch from 15 points onward.

Here is the recount. Rallies of 20 shots or more: 203 out of 1,914, or 10.6%. The appendix attached to the dossier said 19%. Win rate inside those 203 rallies: 51.2%. The 68% figure could not be reproduced under any split I tried, including when I isolated only matches against top-40 opponents. Average rally length: 6.9 shots, against 8.4 in the dossier. From 15-15 onward in game 3: 118 points, 57 won, or 48.3%; of those 57 winning points, 35 came from rallies under 7 shots.

I stopped at that last line. It runs against the story the dossier wanted to tell. The dossier was selling an endurance player who wins long rallies and holds her nerve late. The footage shows her decisive points are mostly created very early in the rally: serve, return, one finishing shot. Without a denominator, a number loses its value. The transfer market is still paying for numbers that have already lost theirs.

Every figure I publish has footprints. And I can point you to them: rally 41 of game 2, minute 23, against an opponent ranked more than 30 places higher. Data does not carry the roar of the crowd. It carries the truth.

I should state my own limits. Twenty-two matches is a small sample. The opponent pool is uneven: nine matches against top-40 opposition, the other thirteen mostly regional. Arena temperature, shuttle type and indoor drift all affect rally length, and I have not standardised all of it. So I do not conclude that this player is weak in long rallies. I conclude only that the 68% figure does not exist in the original data.

There is a professional reflex I try to avoid. Having found an error, people jump to the opposite conclusion and sell it as a discovery. I do not do that.

The correlation between winning long rallies and winning matches says nothing about direction. Does she win long rallies because she is ahead and choosing to extend them, or does she win matches because she takes the long rallies? With 203 rallies, those two possibilities cannot be separated. Some things look like luck, and are in fact an equation. Here the equation does not yet have enough variables to solve.

The real blind spot sits elsewhere. The dossier is not short of data because nobody watches Vietnamese women's badminton. It is short because nobody records the denominator. Big clubs buy from news items, where a handsome number gets passed hand to hand; smaller teams buy from footage, where the number has to be counted again. The paradox is that smaller teams often get the better contract, because they are not paying for the storytelling.

And if that 68% was once accurate in another season, against another opponent pool, it still has to be proven again from scratch. Being right once does not create an attribute.

The signal to track next is not long-rally win rate. It is win rate inside rallies under 7 shots from 15-15 onward, measured against the same opponent pool and the same arena conditions. When that dataset is published, I will count it again. If it survives the second count, it deserves a place in a valuation dossier.