When the Data Isn't Enough to Conclude: A Lesson from Vietnamese Football

In the press room after the SHB Da Nang versus Hanoi FC match in the 2026 V.League season, I was the only female reporter present. I asked coach Le Huynh Duc about his team's expected goals figure of 0.4, even though the team had just won 1-0. A male colleague cut in with a line loud enough for the whole room to hear: what does a woman know about football, she just makes up numbers. I did not argue back. I recorded the tracking data of all 22 players in the match, and that night I published a 3,000-word analysis proving that Da Nang's win came more from luck than from a dominant playing style. The piece was shared more than 2,000 times on Vietnamese football fan pages that week.

That memory returned when I held in my hands an analysis built from nine sections, spanning tactics, club finance, media, and league governance. Every section had neat tables, and every cell had a heading. But reading closely, the entire content inside repeated one status: insufficient information. No article title, no information points, no core viewpoint, no entity list. A beautiful frame, hollow at the core.

I have spent seven years in Vietnam replacing the shouts in the stands with numbers that cannot be argued away. But those seven years also taught me the opposite lesson: an empty frame is more dangerous than a wrong number. A wrong number can be caught, but a gap is always waiting to be filled with guesswork.

Method: read the data before reading the conclusion

My way of working always begins with a step many colleagues skip: extracting raw events before interpreting them. I call it step zero. Before asking whether a team played well or badly, I must know exactly when the match took place, on which pitch, with which lineups, and which data actually exists. If step zero returns an empty set, then every conclusion that follows is a building constructed on sand.

For a post-match analysis piece on the V.League, I usually work with three layers of data. The first layer is events: goals, cards, substitutions, timings. The second layer is process metrics: expected goals (xG), key passes, successful passing rate into the final third. The third layer is off-ball metrics: sprint distance, pressing counts, and PPDA — the number of passes an opponent is allowed to make before your team intervenes defensively, a figure where lower means more intense pressing.

These three layers do not replace each other. Events tell you what happened, process metrics tell you whether it is sustainable, and off-ball metrics tell you the price paid. When I explain these foundational concepts in an article, I am not talking down to the reader. I know many people who appear sophisticated still need a clean definition, and a clean definition has never ruined a good analysis.

What I learned after the question that was cut off in 2026 is a simple discipline: always cite raw numbers before offering a judgment, and always cross-check from at least two independent sources. Since then, every analysis of mine has come with a self-built data table with clear sourcing, so readers can verify for themselves instead of trusting the writer's reputation.

And from that same discipline, I noticed a paradox: when data is entirely absent, the writer's natural reflex is to invent a frame to fill the void. The nine-section frame I was looking at is the embodiment of that reflex — it has the shape of an analysis, but not the weight of one.

The nine-section frame and the gap called insufficient information

Try reading that frame the way an engineer reads a blueprint. Section one asks about tactics: sophistication, execution, personnel fit, key data. All four cells answer with the same phrase: insufficient information. Section two asks about club finance: broadcasting revenue, commercial revenue, wage bill, net debt. Four cells, four blanks. Section three asks about results and the opinion cycle. Section four asks about the league landscape and team positioning. Section five asks about rules and governance. Section six asks about the coaching staff and the dressing room. Section seven asks about risk. Section eight asks about media and expectations. Section nine asks about the football industry's transmission chain.

Nine sections, dozens of cells, and one conclusion running through all of it: cannot yet be assessed. What stands out is not that the cells are empty. What stands out is that the frame still confidently presents itself as a finished product, complete with star ratings and a risk-warning section sorted by priority.

I have followed enough reports to recognize this pattern. When an extraction pipeline fails, the system behind it often does not stop. It keeps running, and instead of raising an error, it produces an empty structure. That structure carries every formal feature of a conclusion: a title, a classification, priority levels, and even a note stating that everything is only speculation.

In football, we meet an identical version of this frame every week. After a match, a report appears with a full set of numbers: 62% possession, twelve shots, seven corners. But without xG, the reader cannot tell whether those twelve shots were twelve attempts from outside the box or three clear chances wasted. Possession without a metric for progressing the ball toward goal is just a pretty circle in midfield.

Data without context is not evidence; it is raw material waiting to be bent to the writer's intent. This is why I always ask the first question when I receive a data table: what context produced it, and does that context still hold for the match I am analyzing.

Empty stadiums and the 38% figure

The 2026 season was the greatest test of that principle. When leagues around the world had to pause and then return in empty stadiums, I tracked 156 V.League matches during that unusual period and found a shift never previously recorded: the home win rate fell from 46% to 38%.

Eight percentage points sounds small. But multiply it across hundreds of matches per season and it upends every prediction model built over decades. Home advantage in football has long been treated as an almost natural constant, and when that constant changes, every equation resting on it is distorted.

I wrote a warning piece arguing that traditional models were being warped, and that we needed a new adjustment coefficient for the crowdless period. A data analyst at Hanoi FC shared the article and applied the idea to the team's away-match tactics.

The point I want to stress is not the result, but how I had to check myself. Before publishing the 38% figure, I asked myself three counter-questions. First, was the sample of 156 matches large enough to rule out random noise? Second, did the decline come from losing the crowd, or merely from a compressed schedule in which strong teams happened to play more away games? Third, did away teams press harder because there was no crowd pressure, or because they had to play at a denser rhythm and chose a more energy-efficient style?

Only after all three questions were answered with data did I allow myself to write that home advantage no longer exists. An empty stadium does not erase the truth. It merely strips away the fog that 40,000 shouts once created.

There is a counter-angle worth keeping in mind. Losing home advantage does not mean losing advantage in a linear way. In some crowdless matches, home teams actually played more freely because they no longer feared being booed after every misplaced pass. My data showed home teams slightly increased their rate of risky passes, even though the final effectiveness did not rise accordingly. This is the kind of signal that only surfaces when you are willing to read three layers of data at once, instead of looking only at the win rate.

0.4 xG and a 1-0 win

Back to the 2026 SHB Da Nang versus Hanoi FC match. Da Nang won 1-0, and reading only the result, people would praise a solid defense and a well-timed goal. But when I reconstructed the entire match from tracking data, a different picture emerged.

Da Nang generated 0.4 xG for the whole match. Hanoi FC generated 1.8 xG. In other words, the away team created chances of more than four times the quality and still went home empty-handed. Da Nang won thanks to one rare finish converted into a goal, plus a goalkeeper having an above-average day. It was a real win, but not a pattern that could repeat every week.

When the press room mocks xG, I know I am reading the right book they have not yet opened. Expected goals does not say which team deserved to win. It says which team created more chances, and in football, chances are the only thing a team controls over the long run. A goal is a discrete event, one that can come from a single individual moment. A chance is a process, reflecting the quality of the whole system.

The lesson I drew for myself matters more than the lesson for others. If I had only the 1-0 result and no process data, I would have written a piece praising defensive tactics. If I had only xG and no context, I would have written a piece indicting luck. Both would be conclusions drawn too quickly from too small a sample.

The crowd may remember a goal forever. I remember the third pass before it, where the real decision was made. In that match, the third pass before Da Nang's goal came from a situation in which Hanoi's defense had lost its structure for three seconds. Those three seconds never appear on the scoreboard. They appear only in positional data, which is why I recorded all 22 players rather than only the scorer.

The multi-unknown equation of the transfer market

The same principle applies to the transfer market, where numbers are inflated even more easily than on the pitch. Every transfer deal is a multi-unknown equation. Most reporters look only at the coefficient before the equals sign.

The coefficient before the equals sign is the transfer fee. It is the easiest number to read, the most shocking, and also the least informative. The unknowns lie behind it: wages, contract length, agent commissions, performance-dependent clauses, and the sell-on percentage owed to the former club. A deal announced at 5 billion dong can be twice as expensive as one announced at 8 billion dong, depending on the installment structure and the accompanying wages.

In the V.League, where financial transparency has never been the norm, these unknowns are even harder to see. I once spent weeks cross-checking information from two independent sources on a foreign-player deal, and found that the figure published in the media was only the tip of a far more complex structure. That is why I never judge a deal by the transfer fee alone, and never conclude that a club has grown stronger simply because it spent a lot of money.

The transfer race among big clubs is more often a branding arms race than a capability race. The club that buys the most expensive player is not necessarily the club that benefits most. Real value often lies with small clubs, where one carefully chosen deal can change an entire season at a fraction of the cost.

A single number can lie, but a model validated across 10,000 matches has no reason to pretend. That is why I build my metrics from multi-season datasets, not from one beautiful match. A single anecdote is journalistic waste. Only long data series have the right to persuade.

Correlation is not causation

This is where I must argue against myself the most. After publishing the 38% figure on home advantage, I received many requests to explain the mechanism behind it. The most convenient answer is: losing the crowd means losing mental advantage, and away teams play with more confidence. That answer sounds reasonable, but reasonable is not the same as evidence.

The correlation between losing crowds and a falling home win rate does not automatically prove causation. There are at least three other mechanisms that could explain the same phenomenon: a compressed schedule forcing more rotation, continuous play without rest eroding the advantage of teams with greater squad depth, and changes to substitution rules allowing away teams to adjust tactics more flexibly. Each mechanism needs its own verification method.

In the nine-section frame I read at the start, one warning was ranked at a high level: any conclusion generated without source data risks becoming pure speculation. That warning is correct, and it applies even to those who consider themselves the most careful. I have seen the best analysts get so absorbed in perfecting a model that they forget the basic question: what question is this model answering, and does the input data actually exist?

The biggest blind spot of a data person is not a lack of numbers. The biggest blind spot is over-trusting one's own structure. When you have a beautiful frame, you tend to want to fill it, even with guesswork. The only cure is to apply the rule of one article, one question: before writing, define the single question the piece answers, and if the data for that question does not exist, then the piece should not exist.

A signal for the next cycle

An empty analysis is not a failure of the analyst. It is a correct signal placed in the wrong position. When the extraction pipeline returns an empty set, the real message is not that this team cannot be assessed, but that the pipeline needs to be re-run before any conclusion is allowed to be born.

In 2026, they called me crazy when I spoke about Croatia. I relied on a PPDA pressing figure of 8.2 — the highest in Europe during qualifying — and a successful passing rate into the final third in the top three, to predict they would reach the final. Croatia did reach the final. But if I had not had those two metrics that year, I would have had no right to say anything at all. The difference between a bold prediction and empty talk lies precisely there.

For Vietnamese football next season, the signal I am watching is not which team will win the title. The signal I am watching is whether clubs will begin publishing their own process data. The day a V.League team dares to post its own xG and PPDA after each round, that day will mark a bigger turning point than any transfer deal. Because once data is public, premature conclusions have nowhere to hide, and fans will have the right to verify rather than to believe.

As for that empty analysis, I will keep it. Not as a product, but as a reminder. It reminds me that in this profession, the greatest courage is not making a bold prediction. The greatest courage is saying I do not yet know, and waiting until there is enough data to know.