Blank Pages in the Analysis Room: When Vietnamese Football Mistakes Silence for Safety

7:42 on a Tuesday morning. On the monitor in the analysis room of a V-League club, a pre-match report runs to 47 rows. Thirty-one of those rows are empty. The metric column has names; the value column has nothing. No exclamation marks, no red cells, no warning lines. The report looks immaculate.

The head coach skims it in four minutes, nods, and signs the starting XI. By 19:00 that same day his team has lost 0-2 at home. Both goals come from the left flank, exactly the zone where that morning's report held not a single figure.

I was sitting in row seven of stand B that afternoon. Only midway through the second half did I realise the problem was not the back four. The problem was the sheet of paper. The club's analysis unit had sent out a perfectly formed false signal: a void presented as cleanliness.

That is the worst kind of failure in this profession, because it makes no sound.

Context: data-entry infrastructure has not kept pace with software spending

Over the past half-decade, Vietnamese football has changed fast at the top layer. Clubs pay for video analysis software, international data packages, multi-angle camera systems, and short overseas training placements for analysts. Domestic workshops on performance analysis for professional club staff have been run as well. V-League introduced VAR in selected matches from the 2026 season. SEA Games 31, hosted by Vietnam in 2026, marked the first time many domestic table tennis events were filmed to international standards, with electronic scoreboards and point-by-point data published.

The middle layer has received almost no investment. Purchased software assumes raw data will flow into the correct cell, in the correct format, at the correct time. When that does not happen, the software is not wrong. It returns exactly what it was designed to return: an empty cell.

The gap between "no data" and "data equal to zero" sounds small. In this trade it is the gap between a map missing pieces and a map showing the wrong road. On paper, both look identical.

Across seven V-League seasons tracked with my own model, most of the error in Vietnamese internal reports does not come from algorithms. It comes from unlabelled empty columns. An expensive model running on incomplete data still produces numbers that look highly professional, and that is the most dangerous part of it.

A three-layer pipeline, and the silent death at the first layer

A modern analysis report runs through three layers. The capture layer turns events on the pitch into records: passes, shots, positions, timestamps, actors. The processing layer turns records into metrics. The presentation layer turns metrics into a story a coaching staff can read in fifteen minutes before the team meeting.

Failures in the capture layer never raise their own alarm. A camera loses one angle, an event logger misses an entire second half because the battery died, a supplier file arrives three hours late, a match ID loses one character in a spreadsheet. Each of those ends the same way: an empty cell sitting among populated ones.

The processing layer treats an empty cell as a missing value. It does not infer, does not guess, does not warn. It simply skips it. And the presentation layer, where the human reads, receives a table with no formatting errors. Nothing turns red, because software is programmed to flag abnormal values, not absent ones.

I once rebuilt such a report for a match on matchday seven a season ago. The original carried 41 metrics; 22 were blank. Among the blanks were PPDA, the pressing-intensity measure, and entries into the opposition box. The coaching staff read it and concluded the opponent did not press meaningfully. They read half a truth correctly: there was no record of pressing, not an absence of pressing.

No red flag does not mean no risk

In software interfaces, a red flag means danger. No red flag means safety. That convention was designed for complete data, and it collapses entirely when data is incomplete. An empty table still obeys the convention perfectly: no abnormal values, therefore no alerts.

This is the point Vietnamese data readers skip most often. We are taught to ask what a number says, but almost nobody is taught to ask whether a cell has data at all. The second question matters more than the first, because it determines whether an answer to the first question exists.

In discussions with analysts at several V-League clubs, I noticed a repeating behaviour. When a report lacks data on a player, people drift into sensory description: "he's doing fine", "he's quick". Those lines fill the hole and look harmless. But they convert a data gap into a conclusion about ability, and nobody notes that the conclusion has no basis.

The match I described earlier followed exactly that script. The report lacked data on the opposing full-back's movement capacity. Nobody filled the gap. The staff assumed he was average. They attacked that flank for the whole first half and lost the ball fourteen times in that zone.

VAR: when slow data becomes part of the law

IFAB approved VAR in March 2026; the World Cup in Russia that same year was the first edition of the game's biggest tournament to use it. Since then, every VAR-enabled match generates an extra data layer: number of interventions, review duration, frame cut points, final decision. In Vietnam, the V-League began using VAR in selected matches from the 2026 season.

That layer has a property few analysts notice: it records not only events, but time. And time here is a genuine tactical metric.

In my own match logs, a goal left hanging for two minutes is not a footnote. Two minutes is enough for a team's pressing rhythm to cool, enough for defenders to reset their positions, enough for a winger such as Nguyen Quang Hai to lose the acceleration of a run he had to abort while waiting. What happens after the goal is confirmed no longer resembles what happened before.

This produces a data error that is very hard to spot. Count chances in the ten minutes after each VAR intervention and teams appear to create less. Conclude that VAR reduces attacking output and you have ignored the mediator: dead time. Technology does not make teams worse at attacking. Waiting does.

For clubs in the V-League, VAR data is missing something more important still: an entry log. Nobody records how long the VAR team took to agree, how long the feed took to stabilise, or whether footage ever failed to arrive in time. Without those records, any analysis of VAR effectiveness in Vietnam is running on half the data.

Another intervention type deserves more attention than it gets: the times VAR stays silent. Silence is logged as "no intervention", yet in many cases silence actually means "insufficient data to intervene". The two states are merged into one code in the statistics table. That is the clearest example of empty data being read as a decision.

The transfer window: where rumour fills the contract's empty cells

A football-free summer is when truth surfaces, with no media smoke left to hide behind. But truth only surfaces for those who hold data, and transfer data in Vietnam is almost always missing exactly where it matters.

The primary source for any transfer is the contract: length, salary, signing bonus, release clause, sell-on clause, agent fees, payment timing. Only a fraction of that is ever published. The rest becomes a gap, and gaps are always filled by rumour, because rumour behaves remarkably like real data: it has numbers, dates, and names.

When nobody knows the structure of a player's release clause, the market manufactures a figure. That figure does not come from a contract; it comes from the demand to have a figure. Once it exists, it gets cited again, then cited second-hand, until it stands as established fact.

This is why I sort transfer information into three tiers. Tier one comes from contracts or club documents. Tier two comes from agents or intermediaries with a stake in the outcome. Tier three is everything else. The line between tier two and tier three is routinely erased, and that is where most readers are misled.

One rarely discussed detail: the wage bill. In the V-League, most clubs do not publish their seasonal wage structure. So when a report says a club has signed a player, readers have no way to judge whether the deal breaks the wage ceiling. Without that data, every claim about next season's squad strength is missing a variable capable of reversing the conclusion.

I once built a simple model to test this. It ranked clubs on two axes: current squad quality and financial headroom. Feeding in the missing-data cases changed the output markedly. The problem was not that the model was wrong. The problem was that the model had to assign a value to empty cells, and every choice carries the bias of the person building it.

Table tennis: where the data gap runs longest

WTT was established by the ITTF and has run a new event system and ranking structure since 2026, bringing a more standardised data architecture to international competition. For Vietnamese table tennis that is genuine progress. But it only reaches the international tier.

At national championship and club level, point-by-point data usually exists only on the electronic scoreboard and vanishes after the last applause. Players such as Nguyen Anh Tu or Tran Mai Ngoc appear on many leaderboards, yet few of those boards show win rates at deciding points, points won on serve, or scoring distribution across service sequences.

This gap has concrete consequences. A player can win four domestic titles on the strength of his serve, then fail internationally because opponents read his third ball. Without point-level data, nobody detects when that strength stopped paying.

I follow table tennis as a data analyst, not a technical coach. From that angle, the most worrying thing about Vietnamese table tennis is not a shortage of good players. It is a shortage of records for each good player. A career not documented in granular data will always be judged by memory, and memory is selective.

The counter-intuitive angle: bad data beats empty data

When people discuss data quality, the reflex is to discard bad data. A wrong number, delete it. A metric measured differently by two sources, pick the better source. It sounds reasonable, yet in the daily operation of an analysis unit, bad data is more useful than empty data.

A wrong number leaves traces. It deviates from related metrics, it creates outliers on a chart, it forces the reader to stop and check. A 100% pass completion rate across a full match is a bad number, and precisely because it is bad it gets caught. Empty data does not. It sits quietly, raises no objection, creates no outlier.

This is the largest blind spot in any analytics pipeline. Quality-control systems are built to find errors inside existing data, not to find the absence of data. An inspector counts the doors that were opened, not the doors that were never installed.

Numbers never lie; only the people reading them lie to themselves. But there is a form of self-deception that line does not quite cover: deceiving yourself by reading a blank page as a statement.

Correlation is not causation, and in Vietnam we have a more dangerous variant: correlation between buying expensive software and believing you now own data. Those two things do not travel together. They merely appear side by side often enough for people to assign them a causal relationship.

Data is the confession of those who once trusted their instincts. It can only confess if it exists. An empty table has nothing to confess.

What the next cycle needs

Three weeks after that defeat, the club's analysis unit added a column to its pre-match report. The column holds no metric and no judgement. It records one of two values: data present, or data absent. I was told the fifteen-minute pre-match meeting now runs four minutes longer, and those four minutes are the most useful part of it.

That is the signal I want to track through this transfer window. When a report describes a transfer, ask which tier the source sits in. When a report says a player performed well, ask whether the data cell for that player exists. When a table looks immaculately clean, look for the empty rows before you look for the numbers.

What decides the quality of a football nation over the next five years is not how much it spends on software. It is how many spreadsheet rows are correctly labelled, because every correctly labelled row is a moment when somebody refused to guess.

I once staked my reputation on a call, and football answered with data. But I also learned that the residual inside every metric, the gap no model can fill, is where human fate actually sits. A coach sacked after a losing run is not sacked because his metrics were poor. He is sacked because his report was empty in exactly the column the board cared about most, and nobody told him that column had never held data at all.

Next season, when a V-League club announces a new analyst, the first question should not be which software he uses. It should be: who is responsible for labelling the empty cells?

Without an answer, we will keep losing 0-2 because of a tidy sheet of paper.