Vietnamese Football and the xG Void: A Decade of Winning by Organisation, and the Price Nobody Has Paid
In my tracking notebook, that match is circled in red: a 2-0 win, an xG of...
In my tracking notebook, that match is circled in red: a 2-0 win, an xG of 0.71. The losing side walked off with 1.94 xG, fifteen shots and six touches inside the box. I stayed forty minutes after the final whistle to chart every phase again, rewinding the last ten minutes three times, and the more I watched the more one thing unsettled me: the winning team barely controlled the game by any measure I could quantify.
It was an ordinary evening in the V.League. No packed stand, no drawn-out VAR controversy, no goal for the ages. Just a result that drifted away from the run of play, and a man taking notes who realised he had no systematic way to verify that.
My job is to turn what happens on the pitch into numbers that can be compared. In Europe I have hundreds of matches a season with ball-tracking data, player-movement data, and an xG model trained on hundreds of thousands of shots. In Vietnam I have eyes, video, a notebook, and a belief that if I am patient enough I will hear the story the scoreline is not telling.
That night I learned something else: that belief is only worth anything if I am willing to admit my own limits.
The data gap is not a technical accident. It is a structural feature of Vietnamese football, and every conclusion about this game has to start by admitting that.
In 2026, when I was a first-year student writing a blog about European football, I recalculated every shot in the UEFA Youth League semi-final between Barcelona U19 and Chelsea U19. Barcelona won 3-0, but Chelsea's total xG was higher. I wrote a piece arguing that Chelsea had created more and been buried by the scoreline. It drew more than twelve thousand reads. The lesson I carried away was simple: a scoreline is a fact, not a verdict.
But the second lesson took years to understand: you can only argue against a scoreline when you have a measurement system good enough to replace it. Without xG, without positional data, without a model, your scepticism is just a feeling. And in this trade, a feeling is the cheapest thing there is.
Context: a league running without a compass
V.League 1 has fourteen clubs, plays a double round-robin, roughly twenty-six rounds a season. That is a league large enough to produce a statistically meaningful sample, if the data existed.
It does not exist in the sense I need.
No public tracking-data provider supplies the V.League with the ball-position and player-movement metrics available in Europe's top divisions. There is no standardised xG model published match by match. There is no automatically calculated, widely shared PPDA dataset. What exists is the basic statistical table: shots, shots on target, possession, passes, cards, and the score.
For a data analyst, this is a dark room with a few small lamps. You can see the shape of objects, not their texture.
The first and most serious consequence sits in recruitment. V.League clubs mostly recruit by eye, by highlight reels, by agent recommendation, by personal networks. A good scout can spot a midfielder's qualities across three live viewings. But the human eye carries systematic bias: it remembers the spectacular and forgets the repetitive. It overvalues speed in a moment and undervalues speed across a match.
I have seen this from the inside. In January 2026, when I was twenty-three and working as an analyst at a consultancy in Shenzhen, a club asked me to assess a young Argentine midfielder playing in Argentina's top flight. I showed that he produced an xG chain of about 0.45 per match, top five per cent in the league, but ran only about 9.8 km per match, below the regional benchmark of 11.2 km. I concluded he was worth signing, because involvement in attacking chains matters more than distance covered.
The sporting director looked only at the physical data, rejected the report, and signed a different domestic midfielder. The player I recommended went on to shine at a World Cup and move to a major English club for a record fee. I wrote a piece about it and it resonated across the scouting community.
The lesson was not that I was right. It was that when you have only one metric, that metric will always lead you. That sporting director had physical data and nothing else. I had attacking-chain data and nothing else. Both of us were reading half the story and believing it was the whole thing.
Since then every report of mine uses a multi-dimensional scale: distance, xG, PPDA, xG chain, receptions between the lines, duel win rate in the opposition half. The full picture has to be built before I write a single line.
But in the V.League, even that multi-dimensional scale lacks raw material. You cannot compute xG chain without sufficiently detailed event data. You cannot compute PPDA without counting the passes an opponent makes before each defensive action. You can count by hand. I have done it. It takes three to four hours per match, and my error margin is roughly ten per cent.
A league without public data is not a league without data. It is a league where the data belongs to those with the time and money to collect it themselves.
Where the national team's winning model is built
If you ask ten Vietnamese fans where the national team is strong, you get ten different answers: spirit, unity, the crowd, the coach, the golden generation. All are true as feelings. None is a testable claim.
I want to rebuild that winning model from what can be counted.
Modern Vietnamese football, since Park Hang-seo took over in 2026, runs on four identifiable pillars. First, a low-to-mid defensive block that protects the central corridor and forces opponents wide. Second, very fast transition, with counter-attacks launched down both flanks. Third, set pieces, where the national team regularly generates more chances than its own open-play average. Fourth, a physical base built through long training camps and a high density of international fixtures.
These four pillars are not a philosophy of beautiful attacking. They are a system optimising probability under resource constraints. When you cannot create more chances than your opponent in open play, you optimise other things: the quality of set-piece chances, expected goals conceded, and the ability to withstand pressure in the final fifteen minutes.
Here I must be explicit about method. I do not have a complete xG dataset for Vietnam's matches in regional competition. What I have is my own notes from live viewing and rewatching, plus basic published metrics. So every conclusion below carries a medium confidence tag, and I will say clearly when I am inferring rather than measuring.
xG is not the truth - it is a compass, and a compass never points to a shortcut.
The 2026 ASEAN Championship final through a data lens
In December 2026 and January 2026, Vietnam under coach Kim Sang-sik won the ASEAN Championship, a tournament many still call by its old name, the AFF Cup. It was Vietnam's third title in this competition, after 2026 and 2026.
The final was played over two legs against Thailand. The first leg on 2 January 2026 at Viet Tri ended 2-1 to Vietnam. The second leg on 5 January 2026 at Rajamangala in Bangkok ended 3-2 to Vietnam. The aggregate was 5-3.
Nguyen Xuan Son, the naturalised Brazilian-born striker, finished the tournament with seven goals and the golden boot. He broke his leg in the second leg and had to leave the pitch.
Everyone knows those facts. What I want to discuss is what lies behind them.
If you look at Vietnam's sequence of matches in that tournament, a clear pattern appears: the team did not win by dominating from the first minute to the last. It won by controlling the decisive windows. In the first half of many matches Vietnam often let the opponent have more of the ball, accepted a high number of opponent passes in midfield, and waited. In the second half, as opponents began to lose structure through fatigue and ambition, Vietnam accelerated.
This pattern is not new. It is a refined version of the model Park Hang-seo built between 2026 and 2026. What Kim Sang-sik added was a central striker capable of holding the ball and finishing inside the box at a level above the tournament's baseline.
Here is the point I want to put in context. In a regional tournament, the quality gap between teams is usually small, and the biggest difference often comes from one player who can create a goal from half a chance. Nguyen Xuan Son was that player. He did not merely score. He changed how the team could play, because when you have a striker who can win duels and hold the ball up top, you can defend deeper while keeping your counter-attacking threat.
In a short tournament, one right player can raise a team's title probability more than any tactical change. Over a long season, that stops being true.
That is the boundary I want people to see. A regional tournament is a test of optimisation in a small sample. A domestic season, or a continental qualifying campaign, is a test of a system's sustainability. These two demand different kinds of evidence, and Vietnamese football has plenty of the first and very little of the second.
The contrarian angle: a regional title proves nothing at continental level
This is the part where I know I will annoy people.
A Southeast Asian title is a real achievement. It is not an easy tournament. It demands fitness, organisation, and the ability to absorb pressure across seven matches. But it is not evidence for a broader claim, that Vietnamese football has crossed some quality threshold.
The reason lies in the tournament's structure. In Southeast Asia, no team has the resources to build a large, stable quality gap. The distance between the best team and the fifth-best is usually smaller than the distance between the region's fifth-best and a mid-tier Asian side. The result is that a regional tournament has high variance, and high variance means results carry little information.
This is where my Croatia 2026 memory returns. Croatia 2026 taught me that a 12 per cent probability is still a number worth betting on. But it also taught me the opposite, which fewer people remember: a team can reach a final on a low probability and still not be a structurally stronger team. Croatia 2026 reached the final through three matches that went to extra time and penalties. They deserved it. They were also lucky. Both are true, and any analysis that keeps only one of them is a bad analysis.
Applied to Vietnamese football: the 2026 ASEAN Championship title confirms that Vietnam was the best team in Southeast Asia in that window. It does not confirm that Vietnam has closed the gap with Japan, South Korea, Iran or Australia.
And here is the problem with how results are read in Vietnam: a regional title is often used as evidence for a much larger claim than it can support. It becomes the basis for saying the game is on the right track, that youth development is working, that nothing needs to change.
No data supports that chain of reasoning. And the truth is we do not even have the data to refute it.
I want to build a counter-argument against myself, following the discipline I set for myself. The optimistic case would be: Vietnam won with a new coach, in a transition cycle, with many young players, and with a naturalised striker. That signals a system with depth, not a lucky generation.
That argument has weight. It cannot be dismissed by feeling.
But it cannot be confirmed by feeling either. To test it I would need data on those young players' performance at club level, in continental competition, in matches against opponents outside the region. I do not have that data in a comparable form.
So my most honest conclusion is: I do not know. And in my trade, saying you do not know is a professional act, not a confession.
Every number is a testimony; only the patient enough will hear the full trial.
The transfer market: where data and emotion share a table
The transfer window is when Vietnamese football's data gap becomes most expensive.
In major leagues, a transfer is judged against a fairly standard metric set: age, minutes played, xG, xA, seasonal development trajectory, wages, and expected transfer value. In the V.League, most of those metrics either do not exist or are not shared.
As a result, transfer decisions rest on three kinds of signal: relationships, memories of a handful of matches, and price.
Look at how Vietnamese players go abroad. Nguyen Quang Hai moved to Pau FC in France's Ligue 2 in 2026. Nguyen Cong Phuong played in Japan for Mito HollyHock and Yokohama FC, and in South Korea for Incheon United. Doan Van Hau was loaned to SC Heerenveen in the Netherlands in 2026-20.
Three different paths, one pattern: Vietnamese players abroad tend to arrive on loan, on short contracts, or on free transfers. None of those deals carried a large transfer fee. That is a more important fact than it looks.
A market without large transfer fees is a market without valuation. And a market without valuation is one where a player's worth is set by something else: by reputation, by a relationship with a coach, by whether he has been called up to the national team.
This creates a feedback loop I find dangerous. A player called up to the national team becomes more valuable in the domestic market, regardless of club form. A player more valuable in the domestic market is more likely to be called up. Selection criteria and valuation criteria become the same thing, and neither rests on performance measurement.
In the transfer market, a figure of 80 million euros can be... a joke. But in a market with no figures at all, anything can be a joke, and no one can prove otherwise.
There is another trend I track in recent windows: the concentration of resources into a small group of clubs. Teams with strong financial backing such as Cong An Ha Noi, Thep Xanh Nam Dinh, and Hanoi FC can sign and retain the best players. The rest operate in a market where they are usually sellers.
This concentration is not inherently bad. Every league has rich and poor clubs. But it has a specific consequence for data quality: when a small group of clubs can buy almost any player they want, they have little incentive to build data-driven recruitment capability. Why analyse when you can simply outbid?
The answer is: because outbidding does not guarantee buying right. But to prove that, you need data. And we are back where we started.
Club finance: owners, invoices and the season
I do not have access to detailed financial statements from V.League clubs. That is itself information.
In developed leagues, club financial statements are public to a degree, because financial fair play and club licensing rules require it. In Vietnam, transparency is much lower, and many clubs' ownership structures are tied tightly to a company or an individual.
This creates what I call the single-source model. The club survives mainly on cash flow from its owner, not on self-generated revenue.
A V.League club's own revenue comes from three main sources: ticketing and matchday, sponsorship, and its share of broadcast rights. The broadcast share in the V.League is low relative to the size of the league, and matchday revenue depends heavily on whether the club is a big draw.
This means a club's spending often does not correspond to the cash it generates, but to its owner's capacity and patience.
The single-source model carries two risks. The first is disruption risk: if the owner stops spending, the club loses competitiveness almost immediately. The second, subtler risk is time-horizon risk. An owner may have a three-year horizon, while a youth-development cycle needs ten. When those horizons do not match, the club prioritises short-term results.
And short-term results mean buying ready-made players rather than developing young ones. Buying ready-made players means high wages. High wages mean fewer resources left for analysis infrastructure, scouting, and academies.
This is a loop, and it explains why so many V.League clubs compete for two or three seasons and then fall away. They are buying results, not buying a system.
I want to be clear that I do not have the data to quantify this loop. What I have is a theoretical model and a set of observations repeated across seasons. This is the kind of conclusion I tag as medium confidence and stand ready to revise when better data arrives.
Governance, rules and the boundary lines
Vietnamese football operates in a multi-layered governance system: the Vietnam Football Federation at national level, the Vietnam Professional Football company at league level, and Asian Football Confederation rules at club level.
Each layer has its own rulebook, and these rulebooks intersect in ways fans rarely see.
Take continental competition slots. V.League clubs have slots in AFC continental competitions, allocated according to the league's ranking in the AFC points system. When a Vietnamese club enters a continental competition, it must meet licensing criteria on facilities, finance and staffing.
This is where regulation, which looks like mere procedure, actually shapes club behaviour.
An academy licensing criterion can force a club to invest in youth development even when the board does not really want to. A stadium criterion can force a club to upgrade infrastructure. These external pressures are often more effective than any internal incentive.
On the foreign-player quota, this is a tool I have a clear view on. Limiting the number of foreign players protects opportunities for domestic players. At the same time, it limits internal competition within the squad, and competition is one of the determinants of young-player development.
I do not think there is a correct number. I think there is a correct approach: any rule on foreign players should come with a mechanism to measure its impact on domestic young players' minutes. Without that mechanism, we are regulating by belief, not by evidence.
On naturalisation, this is a subject I want to handle with the greatest care. A foreign-born player wearing the national shirt is a decision with both a sporting dimension and an identity dimension. Sportingly, it can solve a specific positional problem. On identity, it raises the question of what the national team represents.
I have no authority to answer the second question. I can only say that on the first, a naturalised player is only worth it if he fills a gap the development system cannot fill in the short term. If he is merely slightly better than an existing option, the identity cost is too high for the sporting gain.
The dressing room and the art of managing people
This is the part data cannot reach, and I must be honest about that.
In my analytical file there is a column I call column X: the factors I know matter but cannot measure. A player's motivation after being dropped. The relationship between captain and coach. A young player's feeling when thrown on in the eightieth minute of a big match. These decide results more than people think.
In Vietnamese football the human-management factor carries unusual weight, for two reasons.
First, the quality gap between players in the same position is often small. When technical skill is near-equal, the deciding factor becomes psychology and organisation.
Second, the power structure at many V.League clubs is concentrated in one or two individuals. When power is concentrated, the people-management quality of whoever holds it becomes the key variable.
In the national team's recent history we have seen two different management models. Park Hang-seo's model rested on strict collective discipline, a clear tactical system, and defined roles for key players. Philippe Troussier's model rested on restructuring the playing style towards possession, with a pronounced generational transition.
The second model failed on results in the short term. But I want to ask a question few ask: did it fail because the idea was wrong, or because the time frame did not allow a right idea to be tested?
I do not have a definitive answer. What I have is an observation: when a coach tries to change a national team's playing model, he needs a large enough match sample for his idea to have a chance. A national team plays only about ten to fifteen matches a year. In that small sample, variance dominates.
I do not believe in luck - I believe in a large enough data sample. And the national team, at fifteen matches a year, has never had a large enough sample.
This is one reason judging a national-team coach is far harder than judging a club coach. At club level you have thirty to fifty matches a season to observe. At national level you have a third of that, and the matches differ in objective, opponent and conditions.
Risk: what I cannot measure
When I build a risk table for a club, I split it into six categories: sporting, financial, personnel, regulatory, public opinion and systemic. For Vietnamese football right now, I want to flag the three risks I consider most important.
The first is misidentification risk. This is the risk the whole game carries, and it is larger than any specific sporting risk. It is the risk that we read regional results as evidence of continental capability, and therefore allocate resources to the wrong places. If you misjudge your position, you buy the wrong players, choose the wrong coaches, and set the wrong targets.
The second is resource-concentration risk. When a small group of clubs holds most of the talent and budget, league competitiveness falls, and national-team quality becomes dependent on a small group of players from a small group of clubs. This raises the variance of results.
The third is generational-transition risk. The generation that took Vietnamese football to its current position is entering the final phase of its peak. Replacing them requires a stream of young players of comparable quality, and that stream must be built before the old generation leaves, not after.
This is the kind of risk you cannot measure with one metric. You can only measure it as a trend over time, and that trend needs data.
Media and the expectation spiral
Vietnamese football media has a feature I have observed for years: it runs in short cycles with large amplitude.
After a big win, a young player becomes an icon. After two poor matches, that player is questioned. The amplitude of this cycle is far larger than the amplitude performance data could justify.
This is not unique to Vietnam. It is a feature of any football culture with a large fan base and a small public dataset. When the public has no objective yardstick, it uses emotion as the yardstick. And emotion swings faster than form.
The consequences for players are clear. A young player over-hyped can be crushed by expectations out of proportion to his development stage. An over-criticised player can lose confidence in a phase where confidence is his greatest asset.
For me, this is why my trade exists. A good metric does not make football less exciting. It makes arguments less meaningless. It lets fans say a player is performing better than his results, or worse, and have grounds for saying it.
In the transfer window this cycle becomes most intense. Every rumour creates a wave, and every wave has a life cycle of about forty-eight hours. The only way not to be swept along is to rank rumours by evidence quality, track the money and contract structure, and watch agent behaviour.
A rumour from a tier-one source, confirmed by two independent outlets, is worth something entirely different from a rumour from a social-media account. The surprising thing is how few people distinguish the two when they argue.
The transmission chain of a football economy
When I analyse an event in football, I always ask where it travels.
A change at academy level reaches club level in three to five years, and national-team level in five to eight. A change in club finance reaches league quality in one to two seasons. A change in regulation reaches club behaviour almost immediately.
In Vietnamese football I see three transmission chains running at once.
The first is the talent chain. The supply of high-quality players in Vietnam remains concentrated in a few academies and a few major football centres. When a good player emerges, the question is not whether he is spotted, but who buys him and for how much.
The second is the financial chain. Money from corporate owners flows into clubs, out again as wages and transfer fees, and a small share into infrastructure. The split between wages and infrastructure is a metric I would love to measure but cannot.
The third is the media chain. National-team success creates a wave of interest, which creates sponsorship revenue, which flows down to clubs and academies to some degree.
These three chains are not synchronised in time. The media chain reacts in weeks. The financial chain reacts in months. The talent chain reacts in years.
The desynchronisation of these three chains is why a regional title can create a feeling of progress far faster than actual progress.
This is what I want football policymakers to see. If you judge success by the media chain, you will always find yourself successful. If you judge by the talent chain, you will see the truth later, but more accurately.
What I carry into the next round of fixtures
When I return to my notebook, the match with the 2-0 scoreline and the 0.71 xG is still there, still an open question.
I cannot conclude the winners were lucky, because I do not have a large enough sample to say that result deviated from their norm. Nor can I conclude the losers deserved a point, because xG measures chance quality, not the quality of the final decision. A good striker positions himself slightly better, a goalkeeper reacts a hundredth of a second faster, and my model does not see those things.
What I can do is keep taking notes. Keep building the sample. Keep waiting until I have enough matches to say this pattern repeats, or to say it is random.
The 2026 season taught me a lesson I still carry. When the pandemic disrupted every league and I lost my source of fresh data, I used the time to reassess five European seasons. I found a pattern in home teams' average PPDA: it fell when stadiums were empty. Home teams pressed less when there was no crowd. I wrote a study about it and it opened a new direction of work for me.
The empty stadium is the largest laboratory modern football has ever had. It lets you isolate a variable nobody could isolate before: the effect of the crowd on player behaviour. The 2026 season was not an exception - it was a test for every old hypothesis.
With Vietnamese football, I ask what the equivalent laboratory is. Perhaps matches at neutral venues. Perhaps long national-team camp periods. Perhaps matches played in extreme weather, when the fitness advantage is neutralised.
These laboratories already exist. They simply have not been used.
A number never lies - only the way we read it is mistaken. And in a game that does not yet have numbers, the greatest mistake is to believe we do not need them.
I will keep staying behind after every match, charting every phase, counting every pass, recording what my eyes see but the scoreboard does not tell. Not because I believe data will answer every question. But because I believe a well-posed question is worth more than an answer given too soon.
Vietnamese football has won a great deal over the past decade through organisation, through discipline, through things that cannot be measured by a single number. That is a real achievement. But to go further, this game will need to learn how to measure what it is doing well, so it can repeat it deliberately instead of relying on the intuition of a few individuals.
My question for the coming season is simple, and I leave it open: if one day we have a full xG dataset for the V.League, will we dare to look at the results, or will we keep believing the scoreboard because it is easier to read?"
