The Empty Data Column of V.League 1: When Vietnamese Football Is Told in Adjectives

Minute 88, a penalty. The player sets the ball on the spot, takes four steps back, runs up, strikes. The ball clears the crossbar by roughly a hand's width. The stand holds its breath, then breaks into a long exhale. On the pitch, the taker drops to his knees. In the stand, a man presses both hands to his face.

That night I reopened the match record in my working folder. The first line is complete: the scoreline. The second line: date, time, stadium. The third line — the shot-distance column — empty. The foot-placement column: empty. The number of defenders in the box: empty. The time the goalkeeper left his line: empty.

For twenty years, every shot in Ligue 1 has entered my computer with fourteen attributes attached. In this record, the most important shot of the match entered with a single dash.

By the next morning, Vietnamese social media carried thousands of comments about that penalty. I read the first three hundred. Every word was true, and none could be verified: decisive, weak, character, brittle mentality, scapegoat. Nobody asked the distance in metres. Nobody asked whether the goalkeeper moved early. Nobody asked which way that same player had struck the ball in his four previous attempts that season.

I am 66. Old enough to know a number never tells a story unless we ask it a question.

The summer of 2026 and the habit of verification

In the summer of 2026 I learned to trust something nobody had yet named: xG. When Opta first published Expected Goals for Ligue 1, I was 57, working as a transfer-market administrator in Marseille. Colleagues printed the table, stuck it on the wall, and went back to judging strikers by goals. I did not.

I hand-recorded 1,204 shots from all 20 clubs across the first half of 2026-18: location, body part, the phase that produced the shot, the number of bodies between ball and goal. Then I matched them against actual goals. The correlation coefficient came out at 0.84. That figure does not prove xG is right; it proves that across 1,204 observations, the model ranked chances closer to reality than the method I had used before. Two different claims, and I needed to separate them before using either.

Colleagues said my reaction was slow. I answered with a line I still keep: verify first, use later. A metric without sample size, confidence interval and match context is a slogan in bold type.

A year later, at the 2026 World Cup, I was invited to contribute after my striker-valuation dataset reached a sports newspaper. I was 58, tracked all 64 matches, and counted PPDA for every team. In the semi-final between Croatia and England, Croatia allowed England just 8.2 passes per defensive action, while England allowed Croatia 12.5. I wrote that Croatia would win through pressing in extra time. They won 2-1.

I did not shout. I reopened the spreadsheet and went looking for outliers, because a correct prediction can be correct for the wrong reason. If Croatia had won through two set pieces rather than through pressing, my article would have been right by luck, and luck does not reuse.

In 2026, when football restarted after the pandemic, I sat in Marseille and analysed 81 matches played in empty stadiums in the 2026-20 Bundesliga. Home teams won only 26 per cent, against 43 per cent before the shutdown. My report reached a Ligue 2 club, Le Havre, and they used it to negotiate down the price of a young striker who had just enjoyed a strong home run.

An empty stadium is the finest laboratory a data obsessive could ask for.

Those three episodes taught me a sequence: verify, contextualise, then hunt the exception. When a colleague in France asked me about Vietnamese football, I applied exactly that sequence. And I hit a wall.

Why I opened the V.League file

In 2026, Nguyen Quang Hai moved to Pau FC in Ligue 2. It was the first time I had to open a V.League file for a French colleague, and the first time I realised I could not answer very basic questions.

He asked: where does he shoot from, under what pressure, when does he dribble and when does he pass. I opened three Vietnamese data sources. The first had a complete league table. The second had shot counts, possession, corners — but each defined them differently, and none stated its definitions. The third, the one I needed, barely existed in public.

I closed the laptop and gave him an honest answer: I know he is good, but I have no way of showing you where he is good.

For a man who has spent a career demanding evidence, that was the most uncomfortable sentence I have ever had to say.

Three layers of data, and the third left blank

Football data comes in three layers, each with a different price.

The first layer is results: scorelines, tables, fixtures. Free, public, uncontested. It answers who won and nothing about why.

The second layer is aggregate statistics: shots, possession share, passes, fouls. Cheap, ubiquitous, and the most dangerous, because it creates the impression that analysis has already happened. A team with 65 per cent possession may be playing well, or may be circulating the ball in front of a low block. One number, two opposite stories, and this layer cannot adjudicate.

The third layer is event-level data: every touch with coordinates, pressure, direction, the player executing and the player beaten. It is the expensive layer, and the only one that permits diagnosis rather than description. PPDA lives here. xG chains live here. Everything I have ever done in Ligue 1, the Bundesliga or a World Cup lives here.

In the V.League, the third layer is close to empty in public. Vietnamese football does not lack emotion; it lacks a detailed record of what produced that emotion.

One structural detail makes this gap more expensive than it looks. A 14-team V.League 1 season gives each club roughly 26 matches. The Premier League gives 38. Across 26 observations, a single red card, a single injury, a single poor pitch is enough to distort the whole sample. In a short-season league, every missing data point carries more weight than it would in a 38-round competition.

And there is a variable rarely discussed: the pitch. In Europe, turf quality is close to a constant across venues. In the V.League it is a genuine variable, affecting ball speed, bounce, and therefore the accuracy of any model trained on European football.

xG does not emigrate easily

This is the part I most want Vietnamese readers to hear, because it runs against the habit of importing metrics wholesale.

An xG model is not a law of physics. It is a probability function trained on a specific population of shots. Feed it hundreds of thousands of European shots and it learns that a shot from 12 metres centrally under light pressure converts at one rate, a header from eight metres off a cross at another. Those rates encode European defenders, European goalkeepers, European turf, and European shooting habits.

Move that model to a league with a different set-piece share, more second balls, deeper defensive blocks and a wider spread of goalkeeper quality, and the output will be wrong systematically rather than randomly. Systematic error means it always leans the same way, and that lean will be read by audiences as fact.

An imported xG model is not wrong because it is poor; it is wrong because it was taught a different football.

The answer is not to throw xG away. The answer is to change the unit of reading: instead of absolute values, read rank within the league. A team with a PPDA of 11.4 says little. A team ranked third of fourteen on PPDA says a great deal, because it is being compared with the opponents it will actually face. Within-league ranking is far more immune to calibration error than absolute value.

Simple to state. But it requires something the Vietnamese data market does not yet have: an event-level source consistent enough to rank fourteen clubs on one scale.

PPDA in a league that does not press as its axis

PPDA measures passes allowed per defensive action. It assumes one thing: the defending team wants to win the ball high. If a league defaults to a low block and presses only situationally, PPDA starts measuring the opponent's patience more than your aggression.

At the 2026 World Cup I counted PPDA across all 64 matches and drew a lesson I still repeat whenever anyone asks me about the metric. Croatia won a tournament of low PPDA? Then PPDA is merely a letter.

One letter in an alphabet. It needs the other letters to become a word. For the V.League, the other letters are: the share of goals from set pieces, the second-ball recovery rate, the number of defensive actions inside the box, and the long-pass rate under pressure. Those four together describe what a V.League match actually looks like far better than an unmodified imported metric.

I once tried to build that four-metric set for a group of V.League matches by rewatching footage and counting by hand. Four hours per match. That is the real price of data: not software licences, but the time of one person willing to sit down afterwards.

The minimum evidence set

After years of reading sports journalism in several languages, I have distilled a four-question test. If an analysis cannot answer all four, it is an essay, not a report.

First: which coach, and what system does he prefer. Without a name and a system, every tactical remark floats free.

Second: at least one of three data points — starting formation, PPDA band, or possession band. Not all three; one, with a source.

Third: the specific opponent against whom the tactic is being tested. A system is neither good nor bad in a vacuum; it is good or bad against a type of opponent.

Fourth: whether the claim is about one match or a trend. This is the most frequently skipped question and the most damaging. One match is not a trend. Three matches begin to suggest one.

Those four questions cost less than any analytics platform. They require one person to sit down after the match, and one editor willing to ask again.

What frustrates me in Vietnamese coverage is not the absence of numbers. It is sentences like “team A pressed better” appearing with no metric attached, when three minutes of video review would yield a count of recoveries in the opposition half.

The undisclosed fee

In the transfer market, the problem changes shape but does not shrink.

In the V.League, most deals do not disclose fees. The common structures are loans with options and one- to two-year contracts for foreign players. Which means: when a newspaper writes that club X paid heavily for striker Y, there is no denominator against which to test the word heavily.

I once tried to compute a positional premium rate for a V.League transfer window. The number of comparable deals for one specific position in one specific window typically came out at zero to three. With three observations, a percentage is decoration. With three observations, the honest sentence is qualitative: this fee is above the domestic norm for a 24-year-old striker.

With three comparable deals, a percentage is decoration; a qualitative sentence can be verified.

There is a subtler consequence few notice. If foreign contracts are typically one to two years, the concept of the contract year compresses: every season is a contract year. The volatility of foreign players in the V.League may therefore be structural rather than psychological. Proving that would require contract data and season-by-season performance data — neither of which is public.

The foreign-player quota is the real constraint

When people analyse Vietnamese club finance, they borrow the European frame: broadcast revenue, commercial revenue, wage bill, net debt. The frame is not wrong, but it is not the binding constraint.

The binding constraint in the V.League is the foreign-player registration quota and the transfer windows. A club can have money, need and a partner — and still be unable to buy, because there is no slot left. Conversely, a cash-poor club can hold on to a key domestic player, because that player has few equivalent destinations at home.

This structure produces a very specific distortion: a domestic player's value is set not by his absolute quality but by whether the club has already used its foreign slots. That is the variable to read before any wage table.

I say this as a man who worked a transfer market governed by entirely different rules. Players are variables, the market is a function, but most of my life has been a constant.

The coaching seat is not decided by a results curve

In Europe, sack-pressure indices are built on points per game. In the V.League that model underperforms, because managerial change tends to track ownership relationships and sponsor politics more than a results curve.

That does not make analysis impossible. It raises the evidence threshold: two independent Vietnamese-language outlets plus a club-level signal before the label under pressure may be applied.

One detail strikes me as the most undervalued analytical signal in Vietnamese football: the announcement channel. At many clubs, news of a coaching change or a player sale surfaces first on unofficial channels. Who announced it, where, and when — that sequence is itself data. An official statement arriving three days after social media has already circulated tells a different story from a press release issued first.

The academy pipeline and one export case

If I had to pick one lens for Vietnamese football over the next decade, I would pick the academy pipeline.

The reason is arithmetic. A league can grow broadcast revenue slowly and cannot restructure ownership overnight. But one successful export raises the price of the entire cohort behind him. Hoang Anh Gia Lai was the reference case for an academy built with a foreign partner; what deserves measuring is not the number of trainees but the conversion rate from trainee to first team, and from first team to abroad. Both rates require event-level data to measure honestly.

The Quang Hai to Pau FC case is worth dissecting because it shows what actually changes when a Vietnamese player goes abroad.

What changed was not technique. He still dribbled with his right foot, still struck with his left instep, still operated in the same distances. What changed was the measurement environment: every touch in Ligue 2 is logged, every loss of possession counted, every duel assigned an outcome. The question where is he good becomes answerable, whereas at home it could only be answered by impression.

A player going abroad does not just change shirts; he changes measurement systems.

That is why I say the V.League does not lack good players. It lacks a measurement system that lets outsiders see the quality before the player has to prove it a second time on foreign soil.

Adjectives and their price

There is a loop I have observed in many smaller football nations, and Vietnam is no exception.

A young player has two good matches. The press calls him the future. International outlets pick it up 24 to 72 hours later, usually stripped of detail and seasoned with hype. The player is called into the national team. He has one poor match against a stronger opponent. The same outlets call him a victim of expectation. The loop closes, and nobody inside it ever counted a metric.

For a data person, one rule follows from this loop: Vietnamese-language sources must be weighted more heavily than foreign ones when the story concerns the V.League. International agencies routinely recycle V.League stories with delay and reduced accuracy. A story that appears only in a foreign language, with no Vietnamese source confirming it, should be down-weighted rather than upgraded because it looks international.

A V.League story that appears only in a foreign language should be down-weighted, not upgraded.

At the same time I must be blunt about my own side: European data journalism generates its own loop, decorated with charts. A chart without a sample size is still an adjective, drawn.

Correlation is not causation, and a blank is not an excuse

Here I have to argue against myself, because two symmetrical mistakes are waiting.

Mistake one: turning the blank into an excuse. There is no lazier sentence in this profession than we cannot analyse it because we have no data. At 57 I hand-counted 1,204 shots. Four hours per V.League match to build a four-metric set is doable. No software does not mean no method; it means higher cost, and a choice about which matches are worth four hours.

Mistake two: turning a correlation into a cause. The home win rate falling from 43 to 26 per cent behind closed doors is a strong correlation, but it does not prove the crowd creates home advantage. At least three rival hypotheses must be eliminated first: compressed post-pandemic schedules, altered travel routines, and changed referee behaviour in front of crowds. I wrote that report with those three hypotheses at the top, and I still think that was its most important section.

With the V.League the principle bites harder. In a 26-round season, every correlation is more fragile than it looks. A team on a seven-match winning run may simply have met seven opponents from the bottom half.

There are matches won on the pitch and lost on the spreadsheet — I choose the spreadsheet.

And I should note one place where Vietnamese football is ahead of Europe in a narrow sense: esports. Vietnamese clubs invested in esports rosters earlier than many European clubs of comparable size. There, every action is logged automatically, and a mouse click on an esports screen carries the shape of a pass — with coordinates, a timestamp, a receiver. The paradox is that those clubs hold richer data about their esports roster than about their own football team.

Signals for the next cycle

I will not close with a summary, because summaries belong to people who are finished. I will close with four signals I will track over the next twelve months, and how to measure them.

Signal one: whether any V.League club publishes event-level data for even one match. Not a season. One match is enough to prove capability.

Signal two: the share of Vietnamese analytical pieces that name a metric's source and sample size. This is the easiest indicator to measure and the fastest to reflect change in the trade.

Signal three: whether a second export case follows Quang Hai, and whether that player was measured with data before leaving or merely praised with adjectives.

Signal four: the trainee-to-first-team conversion rate at two or three of the largest academies, published as an annual indicator rather than a press number.

None of these four requires a federation budget. They require a decision: to treat record-keeping as part of football rather than a by-product of it.

Vietnam's post-2026 generation plays in a league that records less about them than a Ligue 2 reserve match records about a substitute entering in the 85th minute. That gap will not be closed by the fee for a foreign striker. It will be closed by one person sitting down after the match, opening a spreadsheet, and starting to count.

A cancelled match is not a lost point; it is a lost page of the diary. A missed penalty in the 88th minute is the same — if we do not record the distance, the angle and the moment the goalkeeper moved, tomorrow leaves us one adjective, and adjectives teach nobody how to take the next one.