Empty Data, Full Conclusions: The Quiet Disease of Vietnamese Sports Analytics

A 42-page report landed with a V.League coaching staff. The technical assessment table had twelve rows, and all twelve read "insufficient information". The risk matrix had seven cells; all seven were blank. The head-to-head section contained no numbers at all. Yet the final page — the conclusions — carried seven tactical recommendations, bolded and numbered, reading as smoothly as if a mountain of evidence sat beneath them.

I kept that report. Not because it was useful. I kept it because it is the cleanest piece of evidence for an occupational disease: analysts fear blank space more than they fear being wrong.

Context: a profession without infrastructure

Over the past decade, the number of Vietnamese people working in sports data analytics has grown faster than the number of data sources they can reach. V.League 1 has 14 clubs and seven matches per round, but second-by-second player-tracking data is neither fully collected nor published. Domestic competitions usually offer only goals, cards, pass counts and possession share — metrics whose meaning was stripped out long ago. Anything deeper has to be coded by hand from video.

Badminton runs on the same rails. The World Badminton Federation publishes results, head-to-head records and rankings, but does not release landing-point data, movement distribution or rally duration. Names like Nguyen Thuy Linh, Le Duc Phat or Nguyen Tien Minh have enough data attached to build a head-to-head profile, yet not enough to answer the hardest question: why does a player win more often in a deciding game than in the first?

That gap is fertile ground for a peculiar product: reports with no data but plenty of conclusions. When the transfer window opens, that product sells extremely well.

Core analysis: four layers of a blank cell

Layer one: missing data is not the same as zero data. A cell marked "insufficient information" is an honest admission. A cell marked "0" is not. Zero completed passes is entirely different from nobody having measured completed passes. But when a spreadsheet runs its formulas, both produce the same arithmetic result. That is why the single most important technical skill in this trade is telling the two kinds of blank apart.

Layer two: the pressure to fill. A report with seven blank cells will not be signed off. Nobody writes in the covering email, "we have no data". They write: "Your team does not press aggressively enough." It sounds reasonable, it is not logically false, and it cannot be verified. It is safe.

I once built an xG model for a match in which the home side generated 3.4 expected goals against 0.8, then lost 0-2 through two individual errors. I wrote that the home side had played better. The online crowd called me a numbers-blind fool. That night I retreated into 200 historical matches and rebuilt the model on cumulative xG sequences instead of single results. The lesson was not "xG is wrong". The lesson was: one match is not one sample.

Layer three: substitute metrics. When positional data is unavailable, analysts reach for PPDA — the number of passes an opponent is allowed per defensive action. It can be computed from broadcast footage without a chip in the shirt. It is cheap, feasible, and says more than possession share. At the 2026 World Cup, Germany allowed South Korea 11.4 passes per defensive action, against their own group-stage average of 9.2. That number explains the 0-2 defeat better than any commentary about desire. PPDA has limits too: it measures intensity, not the quality of the final decision.

Layer four: transfer-window noise. During a transfer window, demand for numbers spikes because money moves through contracts. Agents need a metric to sell a player; clubs need a metric to justify a fee. The verifiable items — release clauses, wage bills, actual minutes played, matches missed through injury — get pushed aside in favour of the word "form". Form is a word with no unit of measurement.

Contrarian angle: a beautiful number is the most suspect number

There is a paradox in this trade. The thinner the data, the more ornate the presentation. A chart full of colour, arrows and percentages is often a sign that the author had to compensate with a great deal of inference. A chart with three rows and two cells marked "no data" is usually more trustworthy.

A beautiful number is the most suspect number. Not because beautiful numbers are always wrong, but because they usually emerge from a process designed to produce them. If a metric matches the hypothesis perfectly on the very first run, the correct response is not celebration but an audit of whether the input data was filtered in that direction.

I have seen Vietnamese transfer reports claim a player has "the highest chance-creation rate in the league" on the basis of seven matches. Seven matches is not a season. Seven matches is a random sequence that can reverse within three more rounds. And the person who produced that number is usually not lying. They simply did not disclose the sample size.

The problem is not confined within borders either. The same metric can be defined differently across two league systems. A 25-metre pass may count as a key pass in one league and a safe pass in another, depending on how the data provider classifies it. When an analysis imports an entire metric set from abroad and applies it to a V.League match, the result is often elegant in form and distorted in substance.

A beautiful number is the most suspect number — true at both ends: when it appears in someone else's report, and when it appears in your own.

Vietnamese badminton sits exactly at this intersection. Players such as Nguyen Thuy Linh or Le Duc Phat compete on the international circuit where data is relatively complete, yet most domestic analysis still stops at "won or lost", "went deep or exited early". The difference between winning a deciding game 21-19 and losing it 19-21 is not spirit. It is who preserved shot quality on the fifth stroke of a fourteen-second rally. Knowing that requires landing-point distribution data. Without it, every conclusion about "character" is just interpretation.

The reverse test

Before releasing any analysis table, I run a single question: if this number is wrong, does my conclusion still stand? If the answer is "it still stands", the number is not really carrying the conclusion — it is decoration. If the answer is "it collapses", the next questions are: who collected this data, by what method, and is there any incentive bending the error in one direction?

Applied to a report made entirely of blank cells, the test gives a clean result: the conclusion has nothing to stand on. It is standing on air.

What would change if this holds

If most analytical reports across Vietnamese leagues carry high blank-cell rates yet still produce conclusions, the problem does not lie in analysts' ability. It lies in the contract. Nobody signs a contract under which the deliverable is permitted to be a single page reading "no data yet". So the thing that needs fixing is the definition of the deliverable, not the skill of the person producing it.

Once that definition changes, the transfer market benefits first. A player assessment that states plainly, "we hold data for 26 of the last 34 matches; the remaining eight lack reliability", is worth far more than a polished assessment declaring the player "consistent". Scouts pay for clarity about uncertainty, not for manufactured certainty.

And if that holds, the next question is not how to analyse better. It is: who will be the first to pay for a report that is allowed to be empty?