Vietnamese Basketball's Missing Data Infrastructure: What Remains After the Final Buzzer

For many nights during the past season I did something that had become reflex: after the final buzzer, I opened my laptop on the arena seats and went looking for data on the game I had just watched. At Vietnam's professional basketball league, what I got was almost always the same. A tidy box score: points per player, minutes, fouls, sometimes a shooting percentage. Then nothing.

No play-by-play. No possession log. No positional data, no shot chart, no time of possession. If I want to know whether a team plays fast or slow, where it scores from, why it turns the ball over, I have to reconstruct it from memory and from video — two tools far less reliable than a game log file.

This piece does not retell any single game. It is about what is missing, and what that absence costs.

Vietnamese basketball has outgrown its measurement infrastructure

Within roughly a decade, Vietnamese basketball has changed faster than any other team sport in the country. The Vietnam Basketball Association launched in 2026, gathering teams that had existed in the amateur scene and turning them into brands with schedules, sponsors and television coverage. Before that, a Vietnamese club had competed in a regional competition from 2026, opening the idea that Vietnamese players could face Southeast and East Asian clubs at club level.

At the national team level, recent SEA Games brought basketball into the group of sports that draw attention. The 3x3 format exploded in schools and parks. A generation of young players grew up watching the NBA daily, knowing what shooting efficiency means, knowing what switching defence means.

But one thing did not grow with it: the capacity to produce and publish data.

Placed next to football, the gap becomes obvious. Vietnamese football is used to xG, to passing maps, to distance covered, simply because international providers sell those data packages for competitions in the region. Vietnamese basketball has no such luck at a comparable scale. No company goes out to log every possession for a league whose online audience is still modest. Nobody pays for data, so nobody produces data.

That is the starting point of everything that follows.

How I built a dataset myself

Because no source existed, I worked as a manual recorder. My method had four steps, and I lay them out so readers know what kind of evidence they are reading.

First, I reconstructed each possession from video, recording the team in control, the start and end time, the type of ending (shot, turnover, free throw, offensive rebound), and the shot location at a coarse level — in the paint, mid-range, beyond the arc. Each game took three to four hours, plus cross-checking time.

Second, I calculated team-level basics: effective field goal percentage, turnover rate per possession, offensive rebound rate, free throw rate, and pace — possessions per 48 minutes.

Third, I tagged defensive actions at a coarse level: man-to-man, switching, zone, full-court pressure, trap. In basketball I could not build an equivalent of football's PPDA — passes allowed per defensive action — because basketball has no dozens-of-pass sequences. But there is a substitute: seconds the opponent needs to cross half court, and possessions in which the opponent must pass more than seven times before finishing.

Fourth, I recorded everything that could not be measured. That is the most important column and the one I hate most, because it reminds me that the data I hold is a very small part of the truth.

My sample is small. I lack the resources to log a full season for every team. I picked games with the best video quality, which creates a bias known in the trade as selection bias. I state this plainly before stating any conclusion.

What the data shows, and what it does not

Numbers show trends, not prophecies.

In my sample, club-level Vietnamese basketball plays at a lower pace than I expected. Possessions per 48 minutes frequently sit in a low range compared with professional leagues of similar audience size. There are three explanations, and I cannot yet pick one decisively.

The first: roster height. When domestic players are not tall enough to operate in the paint, teams must organise offence through outside shooting — and outside shooting at this level means many possessions stall while waiting for a teammate to create space, which slows the game down.

The second: shot quality. If conversion is low, teams must recover the ball more often, which means slowing down to reorganise defence and limit opponents' transition.

The third: coaching. Many Vietnamese teams run set offences, work the system, wait for a good situation. There is nothing wrong with that. It is simply different from the fast basketball many NBA viewers are used to.

I do not have enough data to separate these three causes. And I refuse to pick one and call it the truth.

Another observation held steady across the games I logged: offensive rebound rate is low. When a team misses, the probability of regaining the ball is not high. This fits the height hypothesis, and it also fits a tactical reality: many teams prioritise retreating on defence the moment the shot leaves the hand rather than crashing the boards. That trade-off is rational when you fear transition more than you fear losing a possession.

And here is the most important part: the three-point rate in my sample was higher than I predicted before logging. I had assumed Vietnamese basketball would shoot fewer threes, because without height, threes are the logical compensation — but good shooters are a scarce resource. The result showed teams shooting plenty of threes, most of them contested, not open. The difference between those two kinds of attempts lies outside the data I collected.

This is the limitation I want readers to see. A contested three and an open three both count as three points on the scoreboard. On the raw data sheets of many Vietnamese competitions, they also look identical. To an analyst, they are entirely different shots in tactical value.

The import question through a numerical lens

The Vietnamese professional league allows each team a group of players with a foreign element, rotating on court under a set mechanism. The rule is necessary for the league to reach an attractive competitive level. But it creates a question nobody can answer without lineup-level data.

That question is: in whose hands do most of a team's offensive decisions sit, and how does that change when a domestic player has the ball?

To answer, you need usage rate — the percentage of possessions ending in a shot, turnover or free throw by a given player while he is on court. That metric appears in no game report of the league. So the debate over whether imports are taking development opportunities from domestic players stays permanently at the level of feeling.

In my small sample, I saw signs that the usage rate of the import group was very high in early quarters and declined as the score became safe. If that signal holds across a full season, it says something simple: domestic players are not short of opportunities, they are short of opportunities in the most consequential moments — and those moments are where skill is forged.

I stress the word "if". Small sample, selection bias, and no data from teams I could not log. The correct conclusion here is: this is a hypothesis to test, not a finding.

The cost of the shortfall

I once wrote that a contract is only truly right when the number is signed alongside the signature. In Vietnamese basketball, the signature exists, the number is blurred.

When the domestic transfer market operates without publicly available lineup-level performance data, player valuation depends on reputation, on media image, on scoring average. Scoring average is the worst metric for evaluating a basketball player, because it says nothing about how many possessions he used to get it, the quality of his shots, or his defensive contribution. But it is the only thing everyone sees.

The consequence is a market that pays for output instead of paying for efficiency. A guard scoring 18 points on 22 shots may be valued above a guard scoring 12 on 9 shots, plus three steals and four assists. The second helps his team win more, but the one visible on the box score is the first.

This is the kind of bubble that needs no enormous numbers to form. It only needs opacity. And in a league that publishes no advanced data, opacity is the default.

The counter-intuitive point

There is an assumption I want to flip, and it runs against my own work.

That assumption is: Vietnamese basketball needs more data to get better.

The data shortage in Vietnam does not come from a lack of money, it comes from a lack of incentive. Nobody in the system is harmed by the absence of play-by-play. Coaches still win games. Players still sign contracts. League organisers still sell sponsorship. Media still write stories using scoring averages and highlight plays. Once nobody pays a price for imprecision, measurement capacity never appears on its own.

And there is another paradox: more data does not automatically make basketball better. If poor data is used to justify poor decisions, it is worse than no data at all. A coach who looks at an opponent's shooting percentage while ignoring context — the opponent shoots well because it plays weak defences — will prepare wrongly. I made exactly this mistake.

In 2026, I built a model for a major tournament using pre-tournament cumulative metrics and felt confident about a strong team with the best attacking base in its group. That team went out in the group stage. Looking back, my model lacked data on opponents' defensive intensity — something I had not collected because I did not think it mattered. Three months later I was still thinking about it. Since then, every analysis of mine must include a dedicated section on what the model cannot see.

I tell this story not to ask for sympathy. I tell it because Vietnamese basketball is exactly at the point where an analyst can run one step ahead of the media, or drift a long way off.

On what cannot be measured

In 2026, when stadiums closed, my home-advantage model collapsed. I predicted home win rates would fall below 50 percent, and they did. But when I used that model to forecast the recovery pattern, I failed, because I had not put into the equation the differences in training conditions and mental states between teams.

When the stands were empty, my model collapsed. I knew I had forgotten the human factor.

Vietnamese basketball is full of such factors. A young player starting for the first time in front of three thousand home fans. A coach choosing between winning tonight and developing a player for two seasons from now. A team training twice a day in an unairconditioned arena in midsummer. None of this appears in any dataset, and none of it should be treated as noise. It is part of the model.

That is why I write this section in the middle of the piece rather than at the end. Readers need to know that what I present above is only the visible part.

And for the same reason, I do not believe in hunches. But I believe in what hunches confirm when data backs them.

What would change the picture

Three things, in my view, will decide whether Vietnamese basketball escapes this situation in the coming years.

The first is a decision to publish game logs. A single league releasing possession-level data in an open format would give the entire domestic analytics field its raw material. No positional tracking technology needed, no multi-angle cameras. One careful recorder per game is enough to generate a decade of data.

The second is the emergence of a small group of independent analysts willing to work with self-built data and to publish their methods. Whether this profession can survive in Vietnam, I dare not assert. But it starts with a few people willing to sit for three hours on a game nobody pays them for.

The third is clubs beginning to use numbers for recruitment instead of instinct. This is the slowest change, because it requires someone on the coaching staff who can read a data table. In major leagues that role is called an analyst. In Vietnam, it is usually an assistant coach doing it on the side with a personal spreadsheet.

None of these three changes requires a large sum of money. All of them require a choice.

What I am tracking next

Numbers never need us to defend them. On the contrary, we need them so we do not fool ourselves.

Next season I will keep logging. I will count possessions, time each possession, classify shots into three zones, and record everything I cannot explain. I will publish the method before publishing the conclusion, and I will leave the gaps where the data says nothing, instead of filling them with a plausible-sounding sentence.

If one day I open my laptop after the buzzer and find a real game log file, I will be the first to download it. Until then, anyone writing basketball analysis in Vietnam must accept one thing: they are drawing a map of a territory they have never stood above to look down on.

That is why I do not rush to conclusions. A basketball scene that is growing will generate its own demand for truth, and that demand will eventually force someone to sit down and take notes.