No Red Flags: The Silent Failure Inside Vietnamese Sports Data Analysis

On the analysis room screen, nine data cells. All nine empty. No tournament name, no patch number, no team name, no financial figure of any kind. Thirty minutes later, the summary sent upstairs carried one line: “No major risks identified.”

The person who wrote that line did not lie. He simply had not checked anything.

I first met this kind of failure on Vietnamese grass, in the 2026 V-League season, when an xG spreadsheet built from 26 rounds was sent back with a single sentence: football is not mathematics. Seven years later, I still receive reports shaped exactly like that spreadsheet — full of templates, full of headers, and empty precisely in the cells that decide things.

In sports analysis, this is the most expensive error class, because it makes no noise.

To understand how an empty report walks into a boardroom unchallenged, look at how sports data is produced in Vietnam today.

The chain has three layers. The source layer: international data vendors selling metric packages by match, by season, by league. The extraction layer: where data is pulled by API, by automated scrapers, or by a human typing from video. The interpretation layer: where an editor, a scout or a sporting director reads the table and turns it into a decision.

In Vietnam, the second layer is the thinnest. Clubs buy dashboards. Newsrooms buy subscriptions. Very few hire anyone to verify that data is actually arriving.

The second layer is also the most fragile. A source page switches to JavaScript rendering, an API rotates its key, a character encoding breaks, a page moves behind a paywall — all of them produce the same output: empty cells. But software rarely says “I failed.” It says “not available.”

On the esports side, where I work daily as a transfer market administrator, the problem is sharper because the clock runs faster. A transfer window can compress into three weeks. Roster registration deadlines, patch publication deadlines, contract lock dates — all hard. If the data pipe dies at two in the morning before a deadline, the decision-maker at breakfast still receives a file with perfect formatting and no content.

Seventeen years in this industry, starting as a competitor and then a tournament organiser in 2026, taught me one thing: nobody gets fired for a report with no red flags. People get fired for results on the pitch.

“Unverified” and “clean” are two different states

In medicine, a negative test means something only if the machine is working. If the machine is broken, the return is not negative — it is no result. Nobody is discharged on that basis.

In sport, people are discharged on that basis every week.

A fitness report missing positional data for six players does not mean those six players are healthy. A transfer assessment missing a target’s minutes does not mean the target is reliable. An injury file missing the rehabilitation phase does not mean the player has recovered.

The failure point is that a table always looks like an answer. An empty cell resembles a checked cell. The reader in the meeting cannot see the pipeline; they see a page with no red text. And in an environment where the target is “no bad news,” a page with no red text is good news.

I call that state unverified. It is not the same as clean. The first is an open question. The second is an evidenced answer. Merging the two into one column is the most serious mistake an analysis department can make, and the least likely to be discovered.

The 26-round spreadsheet and the price of one rejection

In the 2026 V-League season I built an xG model from 26 rounds of data. The output flagged Long An: 0.72 xG per match on average, lowest in the league. At that level of chance creation, survival probability was thin.

The report was rejected. The editorial desk at the time held that football is not mathematics, that a club does not get relegated because of a spreadsheet.

At the end of the 2026 season, Long An were relegated.

I was rejected in 2026 over a model. Seven years later, I am paid to write about it.

But the lesson I kept was not that the model was right. It was that the spreadsheet had a column the 2026 report lacked: a column stating which matches had no data, which rounds had been entered by hand, which metrics were estimates only. I knew my limits. The desk never reached that line, because it sat at the bottom of the page.

Had I put the data-limits section at the top, the outcome might have differed. Not because the numbers would have been more persuasive, but because the reader would have known which parts of the spreadsheet were stone and which were sand.

What I learned from V-League 2026: a truth that gets rejected still comes back, only next time it arrives with more data attached.

Croatia’s PPDA of 9.8 and the counterintuitive flip

In 2026 I extended the model to the World Cup. I calculated PPDA — passes allowed per defensive action — for all 32 teams. Croatia averaged 9.8, among the lowest in the tournament. The conventional read: Croatia do not press.

But when successful presses were divided by the opponent’s total passes, Croatia led the tournament at 23% efficiency.

Two metrics, two stories. The first says lazy. The second says economical. Croatia did not run more; they ran at the right moment, at the right player, in the right zone. They let opponents circulate the ball in harmless areas and struck at the point of arrival.

I wrote that Croatia would reach the final. The piece was mocked, mainly on the grounds that the team was strong only because of one midfielder. Croatia reached the final by beating England 2-1 in the semi-final on 11 July 2026 at Luzhniki, then lost the final 4-2 to France on 15 July 2026. The article was shared more than 5,000 times. A European data company approached me to collaborate.

Croatia did not win, but they proved that pressure is also a form of data that knows how to move.

The counterintuitive flip has two faces. The first: a low metric is not automatically a bad sign; it can be the signature of a choice. The second, rarely mentioned: with only one of the two metrics, I would have concluded the opposite of the truth. Enough data to prove, and too little data to destroy, are separated by a single division.

4.2 touches and the collapse of the word “miracle”

Qatar 2026 gave me a live data feed. I tracked Morocco throughout. They allowed opponents an average of 4.2 touches inside their own box per match, courtesy of a 5-4-1 block that held its distances almost unchanged across halves.

In the match against Portugal on 10 December 2026 at Al Thumama Stadium, Morocco won 1-0 through Youssef En-Nesyri’s 42nd-minute goal. I recorded Sofyan Amrabat completing 6 successful tackles and 9 ball recoveries. Those numbers did not appear in the evening bulletin. The evening bulletin spoke of fighting spirit.

I wrote a piece explaining how Morocco neutralised Portugal, using the opponent’s receiving positions and a tackle map instead of adjectives. It spread widely, and a Vietnamese television station invited me on air to comment on data.

Since then I have removed two phrases from my professional vocabulary: fighting spirit and miracle. Both are other names for a data gap. When we cannot yet measure why a weaker team holds firm, we call it a miracle. Once measured, it becomes a structured defensive system, and structure can be taught, bought and repeated.

One match is a story. Fifty matches are the truth.

The esports side: when the pipe dies at two in the morning

Back to the transfer market. My work sits between two tables: contract valuation and performance metrics. Between the transfer board and the pitch, I choose to stand in the middle, measuring both sides.

From the middle, I watch deals decided on a data file whose provenance was never checked. A club reviews a prospect, sees the minutes column empty across three recent competitions, and concludes the player is barely used. In fact the data pipe for those three competitions died six months earlier.

Even a billion-đồng contract begins with a small note about minutes played.

Esports amplifies the error. Patches change constantly, meaning any season-to-season comparison needs an exact timestamp. Rosters change mid-season. A tournament can change format midway. If the extraction layer does not record when the data was pulled, the reader will compare two things belonging to two different versions of the same game without knowing it.

In esports, silence is not exoneration. A file with no red flags is a file that has not been checked.

Anatomy of a silent failure

Looking back at those four episodes, the same pattern appears. Silent failure is born at three points, and one broken point devalues the entire chain.

Point one is the source. If the source page will not load, there is nothing to analyse. Operators usually know this, but do not record it in the report.

Point two is extraction. If the scraper fails, the output is an empty cell rather than an error message. This is the true death point, because it raises no exception for the system to flag.

Point three is interpretation. If the reader treats an empty cell as a safe cell, the first two points may work perfectly and it will not matter.

Vietnamese clubs typically invest in point three — attractive dashboards, specialists to read numbers — and skip point two. That is how an analysis department can look thoroughly modern while making decisions on hollow data.

Fifteen percent and the 8.5 kilometre line

In 2026 global football stopped. My firm took a consulting contract with a V-League club. I pulled distance-covered data for 11 starting players from the 2026 season, modelled the physical decline after three months of non-ball training, and produced an average forecast of 15%. On that basis I proposed cutting the wage bill by 20% on long-term contracts, arguing injury risk would rise in the early weeks of the restart.

The head coach objected, because these players had brand value.

When football returned, that group averaged 8.5 kilometres per match, 1.2 kilometres below their pre-pandemic level. The club had to adjust its policy.

When I delivered the wage-cut advisory, they looked at me as a man without feeling. I was only delivering data, not emotion.

But there is another side to that story I rarely tell. The 15% model had a hole: positional data for three players in the group covered only 60% of matches, and I interpolated the rest. If those three had recovered faster than forecast, my conclusion would have skewed. I stated this explicitly in the appendix, and luckily the appendix was not skipped.

The difference between usable analysis and dangerous analysis is not model accuracy. It is whether the writer dares to state where the model is weak.

When metrics travel without their sources

Over the past three months I collected Vietnamese-language articles citing player metrics from the two top competitions. Most cited figures without a source, without a timestamp, without a sample size. The same metric appeared in five different articles with five different values.

When provenance disappears, readers lose any way to separate real data from data typed out of memory. And once a wrong metric has spread, correcting it costs many times more than checking it at the start.

This is why I set a rule for myself: every metric I publish carries a source, a publication date and a sample size. Without those three, I do not publish. Readers have a right to trace; writers have a duty to leave a trail.

A sports data platform is not measured by how many metrics it supplies. It is measured by whether readers can verify each metric for themselves.

The flip: more data does not reduce silent failure

Intuition says that with more data, this error class disappears. My experience says the opposite.

Every new table added to the chain adds a new set of empty cells. Every new empty cell is a new opportunity for misreading. Every new dashboard makes people more confident in a system whose plumbing they have never audited.

Meanwhile, time pressure in professional sport only rises. Shorter transfer windows. Denser calendars. Nobody has three days to verify a data source before locking a roster.

Adding data without adding verification discipline is the fastest way to turn an analysis department into a machine for manufacturing false confidence.

There is a cultural point here, and I make it carefully. Data has no culture, but the people who produce it do. In many working environments, an empty cell is default-read as no problem. In others, it is default-read as not yet checked. That difference lives in how questions are asked in meetings, not in the software.

And this matters most: correlation is not causation. But the reverse proposition holds to the same standard — the absence of red flags is not the absence of risk. In many meetings I have sat in, those two propositions received very different treatment. The first was repeated as a warning. The second was never repeated, because it protects nobody.

That is the professional blind spot of an entire generation of sports analysts: we are trained to distrust conclusions that look too good, but not trained to distrust blank space. Blank space has no author, no faction, no motive. It looks harmless. It is the only thing in the report nobody is obliged to defend.

What to watch in the next round

I do not trust intuition. I trust the kind of intuition that has been validated across seven seasons.

In the next round, the signal I am watching is not on the pitch. It is in how analysis departments present their reports. If over the coming months a V-League club or a Vietnamese esports team adds a dedicated column stating the percentage of unverified data behind each metric, that is a sign the industry has learned the lesson.

If by the end of the season reports still carry only two colours, green and red, I will know the answer. No red flags means nobody went to check.