The Data Void: Vietnamese Volleyball Plays on Memory, Not Statistics

The Data Void: Vietnamese Volleyball Plays on Memory, Not Statistics

Fourteen-Eleven

That night in a provincial arena, fifth set, 14-14. The rally ran through eleven touches: the libero saved the ball with the back of her hand as her body tipped toward the stands, the setter pushed the ball wide to the left wing, the middle blocker ran a block one beat late, and the opponent's swing passed through the gap between two blocks nearly 1.90m tall. The whistle. The crowd broke into a sound you could no longer tell was cheering or screaming.

I was in row seven, blue pen in my left hand, an A5 notebook yellowed at the spine in my right. I wrote the last line of the set: "First contact to zone 2 — error. Out-of-system attack — point." It was line one hundred and eighty-one of the match. Then the match ended, medals were handed out, coaches shook hands, spectators filed toward the gates.

The person at the stats table handed me an A4 sheet pre-printed with thirty boxes. Nine boxes had numbers. Twenty-one were blank. No perfect-pass rate, no attack efficiency, no block touches, no point distribution by rotation. A five-set final, two hours and nineteen minutes, seven thousand people in the stands — and all that survived was collective memory and a few vertical phone videos.

I kept that sheet. It sits in my desk drawer next to identical sheets from previous seasons. That drawer is the most complete volleyball data archive I have access to.

A Void With a Name

If you work with sports data in Europe or Japan, you know the term event tagging. A professional volleyball match in Italy, Poland, Turkey or Japan is logged by two or three people simultaneously, each covering a layer of events: first contact and distribution, attack and block, serve and defence. At full time they export a file containing one thousand five hundred to two thousand five hundred coded touches. The International Volleyball Federation runs its VIS system for world-level events; major national leagues use DataVolley or VolleyStation as a mandatory standard for index calculation and opponent reports.

In Vietnam, the operational structure of a national-level match usually consists of: one person pressing the electronic scoreboard, one secretary keeping the official record, and for televised matches, a graphics unit displaying the score plus a few simple figures. There is no tagging layer. There is no publicly archived raw data file. The result is published; the process that produced it is not.

This leaves a specific consequence: Vietnamese volleyball is advancing faster than the data layer it is building. The women's national team appears regularly at Asian competitions, has players competing in Japan, and plays five-set semifinals against Southeast Asian rivals. Each of those steps demands a corresponding analytical capacity: which opponent attacks fast in which rotation, which wing their setter distributes to when trailing, which middle blocker blocks cross better than line. Those questions are currently answered by the eyes and memory of the coaching staff.

I came to volleyball from football. In football I spent years arguing about the limits of xG — about how the model measures the quality of a chance but not what happens inside a defender's head in the 90th minute. I thought I had reached the bottom of helplessness in working with data. Then I moved to Vietnamese volleyball and realised that here, the problem is not the model's limits. The problem is that there is often no model to have limits.

A Scoresheet Cannot Tell the Story

Start with the smallest unit everyone has: the score. A team wins in straight sets, loses in straight sets, or wins in five. That is complete information about the result. It tells you the result, not the cause.

In volleyball, four statistical blocks make up a match, and each answers a different question.

The first is first contact — the first touch after the opponent's serve. The key index here is perfect-pass rate: the ball delivered to the setter's zone, at a height and trajectory that lets the setter run the full attack menu, meaning all three front-row attackers are available. When that rate falls, the team is forced into out-of-system attacks — the hitter improvises, no tactical option exists, and efficiency drops steeply.

The second is attack. Here it is not measured by points but by efficiency: points minus errors minus blocked balls, divided by total swings. A hitter scoring eighteen points sounds impressive until you learn she took forty-five swings. That efficiency is 0.31 — good, but the workload placed on her is double that of the player beside her.

The third is blocking. There are two distinct numbers here and many people confuse them: direct block points, and block touches — balls that contact the block but are still dug, slowing the opponent's attack and creating counter-attack chances. A middle blocker may score no block points and still be the most important defensive factor on the court.

The fourth is rotation structure. A team has six positions cycling in sequence, and the setter's position determines whether the team has three or two front-row attackers. Two-attacker rotations are the structural weakness of every team in the world, and how each team covers that weakness — by pushing the setter up, by raising the share of back-row attacks, by substituting — is where tactical identity shows most clearly.

Now try to read a straight-sets defeat from the score alone. Team A loses 0-3. That is all the scoresheet tells you. You do not know whether Team A lost because first contact collapsed — a problem of individual technique and the reception system — or because attack efficiency was low while first contact held — a problem of hitter quality and the setter's choices — or because a weak rotation was repeatedly exploited — a problem of squad structure. Three causes, three completely different coaching solutions, and an identical scoresheet for all three.

I once received an analytics report from a service provider. The file had every section heading, every index name, every table cell — and every cell was empty. Not a single number. I reopened it three times in two days, each time hoping I had misread the format. That file looked exactly like the A4 sheets in my drawer, only better presented.

The Notebook of a Lone Recorder

Because there is no data file to download, I make my own. My method is crude and ugly: an A5 notebook, two coloured pens, one rule — every rally must be recorded before the ball is replayed.

I sort first-contact outcomes into four groups: perfect (ball to the setter's zone, all three attack options available), acceptable (ball near the zone, two options left), poor (ball outside the zone, one option left or out-of-system), and error (direct point lost). Then I record the outcome of the attack that follows.

Based on my experience tracking matches this past season, I logged forty-six sets across eleven matches in the women's national championship. It is a small sample, single-observer, with no inter-rater cross-check, and I had no video to review for two of the eleven matches. I state those limits before stating results, because one of the worst habits in this profession is presenting a small sample as if it were truth.

The findings, exploratory in nature: when a team's perfect-pass rate sat above roughly 55%, its attack efficiency typically landed in the 0.34 to 0.38 band. When that rate fell below 45%, attack efficiency dropped into the 0.18 to 0.24 band, and the share of out-of-system attacks exceeded 40% of all swings. The break threshold appeared to sit around 50%.

What caught my attention was not the strength of the relationship — anyone in volleyball knows that intuitively — but its shape. The relationship is not linear. From 60% down to 55%, efficiency barely moved. From 52% down to 47%, efficiency lost nearly a third. There is a cliff, and the cliff sits exactly where the naked eye cannot see it, because within a single set the difference between 52% and 47% is two balls.

For comparison of resolution: a professional tagging file for one match holds roughly one thousand five hundred to two thousand five hundred coded touches, with court coordinates, technique type, executor and outcome. My notebook for one match holds about one hundred and eighty rows. The gap between those two figures is not a gap in technology. It is a gap in the number of people sitting down to record.

The Blind Spot in Talent Identification

If you want to know why the data void matters more than any tactical story, start with talent identification.

A young player from a rural province walks into a national youth tournament. She is 1.78m tall, touches 2.95m, and plays middle blocker. After the tournament, the youth national coaching staff must decide whether to add her to the list. What is the basis for the decision? Height measured with a tape. Vertical jump measured by hand. The impression of whoever was watching. And in the best case, a few short video clips.

Nobody knows what her first-contact level is, because nobody records first contact at youth level. Nobody knows whether she blocks cross better than line, because block touches are not counted. Nobody knows whether she holds her attack efficiency when trailing in the third set, because situational data does not exist.

If you have ever worked with a data pipeline, you will recognise a familiar class of failure. Automated text processing includes a step called entity extraction: the system reads a document and identifies names of people, teams and competitions. When that step fails, the output is not a wrong list — it is an empty list. The entity does not vanish from the world; it vanishes from the record.

Vietnamese volleyball is running that process by hand, and the output is a series of empty lists. Every season, a number of players leave this sport's memory without leaving a single data line behind. They played three seasons, they blocked balls at important moments, and there is no way to prove it beyond the testimony of those who were present.

I have my own method for this. Every time I log a match, I mark the names of players for whom I cannot find any public data online. Last season, more than half the players I logged fell into that group. Not because they played badly. Only because nobody ever opened a file for them.

When Data Becomes Money

There is a material reason behind all of this, and it is simpler than people assume.

To have a complete tagging file for one match, a club needs four things: a software licence, a person trained to use it, an archiving process, and time. All four cost money. In Vietnamese football, clubs beginning to hire video analysts is a phenomenon of roughly the last ten years. In volleyball, it has not happened.

The financial picture of Vietnamese volleyball is starkly asymmetric. One group of clubs has large, stable sponsors that treat volleyball as a long-term media channel. Another group lives on provincial budgets, on parent units, on short-term contracts renewed annually. The first group could pay for an analyst if it wanted to. The second has no room to put that desk.

The result is a league in which a few teams can build knowledge about themselves and their opponents, while most cannot. That gap does not show on the standings immediately. It shows three to five seasons later, when the club with accumulated data knows exactly what profile of player it needs and the club without data is still recruiting by feel.

The transfer market is where the void is most visible. In Vietnamese volleyball, domestic transfers almost never carry public figures. No transfer fees, no salaries, no contract lengths are published. This leaves the market without a price signal. Without a price signal, every valuation is sentiment, and sentiment always leans toward the familiar.

Every contract is a thread; I weave the transfer market out of data. But in Vietnamese volleyball, I often have to weave with threads broken in the middle — a name appears at club A one season, disappears, then reappears at club B the next, and no line explains the interval.

The First Vietnamese Player in Japan's V.League

In August 2026, Tran Thi Thanh Thuy became the first Vietnamese volleyball player to compete in Japan's V.League, wearing the PFU BlueCats shirt. It was a landmark for Vietnamese women's volleyball, and it was also a natural experiment for this entire story.

In Japan, a player enters an ecosystem where everything is recorded. Training load is tracked by wearables. Every match has a full tagging file. Before each match, the coaching staff holds an opponent report showing distribution tendencies by rotation, the weaknesses of each hitter against a double block, and the first-contact rate of each opposing libero by serve zone. The player does not need to understand that entire system. But she breathes it every day.

The question I care about is not whether she played well. The question is: when a Vietnamese player lives inside a complete data culture for one or two seasons, and then returns, what happens to that knowledge?

Two possibilities. The first: the knowledge transfers. She returns to the national team, tells her teammates how an opponent meeting is run, how people read an opponent's first contact to choose serve zones, how a middle blocker prepares for each rotation. When one person brings back a system, that system can take root, however slowly.

The second: the knowledge evaporates. She returns, enthusiastic, proposes things, and finds there is no software, no recorder, no archive, no meeting that lasts longer than fifteen minutes because the coaching staff is busy with training, lodging, travel and budget. After a season, the old way returns, because the old way is the only one that operates under current conditions.

I do not know which will happen, and I do not want to pretend my data is sufficient to answer. But I know this: when an individual carries a system, that system depends on the individual. When an organisation owns a system, that system survives the individual's departure.

The Counter-Intuitive Point

The first reaction most people have on hearing this story is: then buy the software and hire the people. I have heard that answer many times, and I think it is half right.

The wrong half is the assumption that more data is automatically better. In practice, incomplete data is more dangerous than no data. A file with all its cells defined but only half of them filled will be read as a complete picture, because its form looks complete. People will make decisions from it with the same confidence they would bring to a fully populated file.

A concrete example. If you record only scores and each hitter's point tally, you will conclude that your opposite hitter is your team's most important player, because she has the most points. You will build the tactics around her. You will look for every way to feed her the ball more. What you cannot see, because you do not measure it, is that her efficiency is only high when the team's first contact is good, and when first contact collapses, that efficiency falls below that of an ordinary hitter. You have built the house around a pillar whose load you never measured.

The second wrong half is the assumption that a system can be imported without the people. A professional analytics software licence sitting on a club laptop with nobody trained to operate it is a very expensive ornament. I have seen this happen in football, and I see no reason it would not happen in volleyball.

There is another layer of complexity here, and I want to say it plainly. The parties most willing to pay for Vietnamese volleyball data will not be the clubs. They will be betting companies. Live data fed to betting operators is the darkest side effect of sports digitisation, and anyone building data infrastructure for a young sport should first answer one question: who does this data serve.

I do not say this to oppose building data. I say it to put the priorities in the right order. Infrastructure built for the needs of coaches, scouts and the players themselves will take a different shape from infrastructure built for the needs of a betting market. The same software, two different data architectures, two different fates.

Signals for the Next Cycle

I will not close with an appeal. There have been more appeals in Vietnamese volleyball than there are people sitting down to record.

I will close with what I am watching.

I am watching the number of people at the stats table each match. If over the next two seasons that number shifts from one person recording the score to two people recording events, that is a bigger signal than any signing.

I am watching the academies and youth development centres. If one of them starts archiving player data from age fifteen — height, jump, medical outcomes, minutes played — then seven years later it will own something no club in Southeast Asia has.

I am watching the players who go abroad. Every return is a chance for a system to take root, or a chance for it to be wasted. I will track which of them is given an official role in the coaching structure, and whether that role comes with data access.

And I am watching my own notebook. It has thickened season by season. The figure still smoulders in the spreadsheet; every season I blow on it once. Some nights I wonder whether I am recording in order to analyse, or recording so that those players do not disappear entirely.

xG never explains why a match makes us cry. Data does not soothe pain, and that is why it is not the most important thing in the world. But when there is no data, we also lose the ability to argue about that pain seriously. Every defeat can be explained by spirit, by nerve, by bad luck, and nobody has to be accountable for anything.

My data leans toward a simple and unglamorous conclusion: the problem with Vietnamese volleyball is not that we lack an advanced analytics system. The problem is that nobody has ever sat still in one place for two hours to record what actually happened on court.

The question I leave, like a cup of tea set on the table: if next season every national championship match had exactly one more person — a young person, trained for three weeks, recording first contact and attack outcomes — then after how many seasons would Vietnamese volleyball know itself?

I keep a small hermitage where volleyball and data bow to each other. Right now only one person sits in it. The door stays open.