Vietnamese Volleyball: 41% Second-Ball Points and the Data Gap in the National Championship

On November 18, 2026, in the second phase of Vietnam's national volleyball championship, one team entered the match with a positive first-pass rate six percentage points lower than its opponent and lost all three attacking metrics shown on the electronic scoreboard. It won 3-1. The road to that win lay in an area for which the scoreboard has no column: 41% of its points came from second-ball situations after the opposing block touched the ball, and its break point rate reached 38%, nearly eleven percentage points above the league average.

I rewatched those four sets until 2 a.m., coding every rally by hand because there was no data file to download. Numbers are like dust: they only mean something when we are calm enough to look through them. And looking through, I did not see a better team. I saw a league operating on the weak light of a scoreboard.

The scoreboard of a Vietnamese volleyball match records four things: points, aces, direct block points and errors. Those four columns are enough to broadcast a match. They are not enough to make a decision. They do not say how far from the net a libero's pass travelled, how many seconds a setter needed to release the ball, or that a spike absorbed by a one-touch block opened a counter-attack finished by a teammate. The best rally of a match can end with a point credited to someone else.

The consequence outlasts a single match. Players are judged by points, contracts are signed on highlight reels, and the most valuable positions in modern volleyball produce the fewest points. Liberos score nothing. Setters score very little. Nobody pays a premium for a great passer who never appears in a thirty-second clip on social media.

World volleyball passed this milestone long ago. Rally-level coding systems such as DataVolley have been used in European national leagues and continental cups since the early 2000s, breaking a match into thousands of geolocated events. In Vietnam, the most common tool in a club meeting room is still a person writing on an A4 sheet and a coach who trusts his memory.

This is not an indictment. Across the 2026, 2026 and 2026 seasons I coded 96 national championship matches and a number of national team matches in regional competition. I did it to answer a narrow question: if Vietnamese volleyball is strong at one thing and weak at another, what are those two things, and how wide is the gap between them.

Every rally begins with a coin flip: the serving team is at a disadvantage. Winning a rally while receiving is called a side-out, and a strong team's side-out rate usually sits between 55% and 62%. Winning a rally while serving is called a break point, and that is what separates levels. Side-out is survival. Break point is the ability to kill a match.

My coded data shows a troubling, repeating pattern. The national champions of the last three seasons all ranked in the top three for break point, at 30% to 32%, against a league average of about 27%. But when those same players step into Asian competition, their break point collapses to roughly 19% to 21%. The reason is concrete: serves that disrupt a mid-table domestic passer are received almost perfectly by a national-team libero.

Vietnamese volleyball's domestic strength is built on the serve, while international volleyball is decided in transition. A team can win at home by serving four percentage points better than everyone else. It cannot win regionally that way, because everyone receives well there, and the match falls into what I call the second ball.

The second ball is the attack after the ball has been blocked or dug. It has no fixed pattern, no diagram, no coach who can draw it on a whiteboard before the match. It is decided in 0.8 seconds. And it is where Vietnamese teams lose the most points against equal or slightly stronger opponents.

I measure setter distribution with entropy, a concept borrowed from information theory that captures how spread out the choices are. High entropy means the setter shares the ball among many attackers, so the block cannot guess. Across the 96 coded matches, the champions averaged 2.31 bits and the bottom team 1.84 bits. Half a bit sounds small, but it equates to the opposing block guessing wrong roughly one extra rally in seven.

What struck me more is that entropy is not a constant. It moves with the score. When a team leads by three points or more, the best setters in the league rise to 2.5 bits. When a team trails by three, entropy collapses to about 1.6 bits. The setter funnels the ball to the number one hitter. The opposing block reads it before the ball leaves the hand.

I do not blame the setters. They are behind, they need a sure point, and the surest point is the player who has scored 20 in three sets. The problem is that the team has no second option drilled well enough to be a real option. When a team trails, every attacking system returns to instinct, and the instinct of a team that has never been measured is the instinct of one player.

Attack volume is the second variable I track. In the national league, some outside hitters routinely take more than 50 swings per match. I split hitters at a threshold of 45 swings and compared efficiency. The result is blunt: below the threshold the group averaged 44% attack efficiency; above it, efficiency fell to 34% and the rate of being blocked outright nearly doubled.

That number does not say the best hitters get worse. It says they are forced to attack in harder situations: passes far from the net, slow second balls, a block already in position. A team without a second weapon turns its best player into its biggest burden. At national-team level the mechanism is clearer, and it explains why Vietnamese defeats against stronger sides follow the same script: a level first set, a fade in the third and fourth.

Fitness is the piece the scoreboard hides entirely. Splitting attack efficiency by set produced a stable curve across three seasons: sets one and two are nearly identical, set three drops about 4%, set four drops an average of 11% against the opening set. The drop varies by team. Squads with real depth lose only 5% to 6% in set four, thanks to rotation and to a smarter spread of attack volume in the first two sets.

Blocking is where Vietnamese data is most distorted. Official figures count only direct block points, while the real value of a block lies in the one-touch that slows the attack and sets up the defence behind it. In my sample, the champions averaged 6.8 block touches per set and the bottom team 4.1. The direct point difference from blocks between those two teams is only about 0.9 points per set, but the difference in points generated from transition after a block touch is close to 3 points per set.

Height explains little of that gap. I matched average on-court height against block effectiveness and found a very weak correlation. What correlates much more strongly is block timing and the distance between the two blockers. A tall block out of sync is worse than a short block in sync, because the seam between two blockers does not depend on height.

Here I have to turn back to myself. In 2026 I wrote a prediction that got the nature of a match completely wrong because I trusted form and feeling. That lesson became discipline. When a model fails, I do not blame the data; I blame myself for believing it blindly. On this volleyball question I must state the limits of the sample: 96 matches is small, the coding was done by one person, and every figure above carries the error of the person writing it down.

The second limit is more serious. Correlation is not causation, and in volleyball it appears in a deeply uncomfortable form. A team with a better first-pass rate does not necessarily win, because a good rate may simply reflect weak opposing serving. A libero with a beautiful number against the last-placed team is not better than a libero with an average number against the champions. Put those two numbers side by side in a ranking table and you have manufactured a lie with fully sourced statistics.

The biggest blind spot of any volleyball model is what does not produce points. A perfect pass that lets the setter run a fast attack counts for nothing if the teammate hits out. A libero's movement that forces the opponent to change serving direction appears in no data file. I once judged a libero by points scored and reached the wrong conclusion about her. That method is convenient for the writer and unfair to the player.

There is a layer I cannot measure and do not pretend to. The state of a team trailing 2-1 in the fourth set, the voices in the locker room at the interval, a setter choosing the third option in the decisive rally. Data reveals only the visible part. The rest belongs to people, and I leave it there rather than assign it a number that looks scientific.

When the ball stops bouncing, I write a plan for the one thing beyond argument: preparation. If the national championship publishes rally-level data from next season, three changes arrive quickly. First, the market value of liberos and setters will rise, because for the first time there is evidence. Second, teams will rotate hitters against a volume threshold instead of a feeling, and fourth-set efficiency will become the most requested metric. Third, break point will become a selection criterion instead of total points.

I do not bet on passion; I bet on probability verified three times. And the highest probability here is this: a volleyball ecosystem whose 96 matches were coded by one person will not travel further than that person. What is needed is not a better model. What is needed is a recording system so that every decision about an athlete no longer depends on what her coach happens to remember the next morning.

On November 18, 2026, the winning team did not know it had won with 41%. It won with an instinct forged over years of training that nobody wrote down. The question for next season is not who will be champion. The question is who will be the first to pay for knowing exactly how they won.