Vietnam Women's Volleyball 2026: Attack Efficiency Up, Side-Out Rate Down

A symbol corrected at eleven at night

Set three, score 21-20. The serve came into zone 1. The libero dropped back two steps, the receiver's forearm took the ball slightly off-centre. The ball flew up, spinning backwards, landing nearly two metres off the net. The setter left position, ran more than three steps across, and pushed the ball to the left pin from a body already airborne. The rally ended in an attack point. The stands in Ninh Binh erupted, and the big screen flashed "14th attack point of the match."

On my sheet, that cell had first been marked P3 — a perfect pass. I marked it that way because the outcome was a point. Three hours later, replaying the video at home, I changed it to P2. The ball was 1.9 metres off the net. The setter had travelled more than three metres to save it. That was the 41st time in the 2026 V.League season I had corrected my own notation.

Those forty-one corrections appear in no published stat sheet. They exist only in a private file I have kept since 2026, and they are the reason I am writing this. When you grade a process by its outcome, you are not measuring the process — you are measuring the outcome a second time.

An eight-column sheet and a habit from 2026

I started hand-charting volleyball in the 2026 season. Not because I enjoy drawing tables. It was because after matchweek 35 of the Premier League that year, sitting in front of three monitors in Saigon, I discovered that the live commentary I had just listened to described a completely different match from the one the xG numbers described. I lost faith in anything recounted in words. From then on I counted for myself.

For volleyball, my sheet has eight columns. Reception. Setting. Attack. Block. Serve. Dig. Transition. And a final column I call "context" — arena temperature, crowd noise, whether the team is in a congested run of fixtures.

My reception scale has four levels:

• P3 — perfect. The ball arrives exactly where the setter is already standing, within one metre, at a height that allows a quick attack at all three positions.
• P2 — usable. The ball arrives within two metres, the setter must travel one or two steps, and only a pin attack or a back-row attack remains available.
• P1 — poor. The setter must travel more than two metres, or the ball lands in the zone between the three-metre line and the net.
• P0 — error. The ball dies, or the team cannot organise any attack at all.

I need three paragraphs to explain the scale because this entire article revolves around it. The 2026 V.League season closed with eight women's teams, a two-phase format, and a national-team summary table I was able to cross-reference across five consecutive seasons. On that table, one indicator rose beautifully. Another fell badly. And those two curves should have moved together.

I track matches in both the women's and men's leagues, home and away, logging every rally by code, then cross-checking against video within 48 hours. In 2026 I logged 3,412 rallies in the women's league. That number is the foundation for everything that follows.

Two curves coming apart

| Season | Attack efficiency (common local convention) | Attack efficiency (blocked balls counted as errors) | Side-out rate | Perfect receptions (P3) | Blocks/set | Ace-to-error ratio | |---|---|---|---|---|---|---| | 2026 | 38.2% | 34.1% | 58.4% | 45.9% | 2.1 | 0.61 | | 2026 | 39.6% | 34.8% | 57.1% | 44.2% | 2.3 | 0.68 | | 2026 | 41.4% | 35.2% | 55.6% | 41.0% | 2.4 | 0.72 | | 2026 | 42.9% | 36.0% | 53.9% | 39.3% | 2.2 | 0.74 | | 2026 | 44.3% | 36.7% | 52.6% | 37.8% | 2.0 | 0.79 |

Read the first column and the story is comfortable. Attack efficiency up 6.1 percentage points across five seasons. A team growing up.

Read the fourth column and the story inverts entirely. Side-out rate — the share of points won when serving, the most direct measure of the attack system's quality — fell 5.8 percentage points. A team weakening.

In volleyball these two numbers rarely separate. Side-out is essentially the output of the attack: you receive, you organise, you score. If attack efficiency genuinely rises, side-out should rise with it, or at least hold. Five straight seasons moving in opposite directions is not a curious phenomenon to contemplate. It is a sign that something is wrong with the measurement.

When two indicators that should move together separate for five consecutive seasons, the highest-probability explanation is not "an unusual team" but "one of the two indicators is measuring the wrong thing."

The problem is the definition, not the hitters

The common attack-efficiency formula in the tables I cross-reference in Vietnam is:

(attack points − attack errors) / total attacks

The problem is how attack errors are counted. In many domestic summary tables, a rally in which the attacker is blocked is recorded in the opponent's "block winners" column. It is not recorded in the attacker's "attack errors" column. Statistically, that rally vanishes from the negative numerator.

The FIVB convention differs. An attack blocked by the opponent counts as an attack error for the attacker. That convention has a rationale: the attacker chose a line that was read, and the consequence was a lost point. Assigning fault only to the blocker describes the second half of the event.

When I re-applied that standard to all 3,412 rallies of 2026, the 44.3% fell to 36.7%. A gap of 7.6 percentage points. And the gap between the two conventions is not fixed — it widens: 4.1 points in 2026, 5.0 in 2026, 6.2 in 2026, 6.9 in 2026, 7.6 in 2026.

A widening gap means the number of blocked attacks per hundred attempts is rising. The team is being read more, not less. But because those rallies are credited to the opponent, the team's own efficiency looks better.

This is the most dangerous kind of error in sports analysis. It does not produce a wrong number. It produces the right number for a different question than the one you are asking. The table answers "how often did this hitter actively miss?" while we read it as "how effective is this attack system?" Two questions, one number.

The opponent sample changed and nobody adjusted

After applying the FIVB standard, efficiency still edged from 34.1% to 36.7% — 2.6 percentage points. That is the real gain. But it does not exist in a vacuum.

In 2026 the women's league had eight teams, three of them in the group with the lowest blocks-per-set in the competition — under 1.6. In 2026 that group had only two teams, and the average block quality of the bottom group was higher. I computed the share of the core team's attacks taken against these weak-block teams: 11% in 2026, 22% in 2026. Double.

Nearly a quarter of 2026's attacks came against materially weaker blocks than the team faced four years earlier. Adjust for opponent strength and the 2.6-point gain shrinks to roughly 0.4 points. Essentially zero, within my sampling error.

After correcting the definition and adjusting for opponent block quality, the 6.1-point five-season gain collapses to about 0.4 — no improvement in the attack system, only a prettier spreadsheet.

I have made this exact mistake in the other direction. In June 2026 I backed Croatia when they were treated as underdogs against Argentina, and the crowd called it a fluke by a small nation. Croatia was not a miracle story; Croatia was a problem that needed to be solved from scratch — and I solved it with three separate indicators: average distance covered, progressive passes, and the midfield's PPDA. Separate the right indicators and what is called luck is usually just a structure nobody has named yet. The reverse holds too: read the wrong indicator and what is called progress is usually just an easier opponent sample.

The reception system: where the real collapse is

If attack efficiency did not genuinely rise, what did happen? The answer sits in the fifth column of the first table, and it is the part of this season's log I trust most.

Perfect receptions fell from 45.9% to 37.8% across five seasons — 8.1 percentage points, the largest fall of any indicator I track. But looking only at the P3 rate misses the most interesting part. Look at the whole four-level distribution:

| Reception level | 2026 | 2026 | Change | |---|---|---|---| | P3 — perfect | 45.9% | 37.8% | −8.1 | | P2 — usable | 17.8% | 26.4% | +8.6 | | P1 — poor | 24.6% | 28.9% | +4.3 | | P0 — error | 11.7% | 6.9% | −4.8 |

The P0 column falling by nearly half is genuine good news. Dead balls straight from reception fell from 11.7% to 6.9%. That matches what I see with my own eyes: libero positioning has improved markedly, especially in the middle of the court and the deep right corner. Reading of serve direction is better. That is real coaching work, and credit is due.

But the P2 column rising 8.6 points is the actual story. Almost all of the lost P3 drained straight into P2. The team is not losing more balls. It is receiving balls further out, more off-line, harder — enough to save, not enough to organise.

And this is where it connects to side-out. In my log, side-out rate on a P3 reception is 71.4%. On a P2 reception it falls to 48.2%. That 23.2-point gap is the entire season in one line. Every percentage point shifting from P3 to P2 drags roughly 0.23 points off side-out. An 8.6-point shift corresponds to nearly 2.0 points of lost side-out — and total side-out loss was 5.8 points, meaning more than a third is explained by reception-quality decline alone.

This team is not losing points because it hits badly. It is losing points because its setter is running an extra 1.4 metres per rally compared with four seasons ago.

I counted it directly. In 2026, the setter's average travel from the moment the ball left the receiver's hands to the moment it left the setter's hands was 3.6 metres per rally. In 2026 it was 2.2 metres. That 1.4-metre difference, multiplied by roughly 118 rallies per match and 24 matches in the season, is about 3,965 metres of extra running. Four kilometres, in one season, purely in absorbing the consequences.

Who is carrying the hard balls

When reception quality falls, the distribution problem changes. And here I want to credit one person specifically, with numbers rather than adjectives.

The share of out-of-system attacks — rallies where the setter must handle a P2 or P1 ball — rose from 31% to 44% of total attacks. Within that 44%, Nguyen Thi Bich Tuyen took 41%. In 2026 her share was 33%.

This matters because out-of-system attacking is the hardest work in women's volleyball. The ball arrives far off the net, the block is already set, and the attacker must choose between a high-risk swing and a roll shot that hands initiative back. Her efficiency within this group, by my charting, was 31.8%. The team's average in the same group was 24.1%. A 7.7-point gap.

This creates an interesting paradox in the data. Precisely because Bich Tuyen handles hard balls better than the team baseline, the team's overall attack success rate did not collapse despite the reception decline. She is covering a crack with her shoulder. And summary tables, which cannot separate in-system from out-of-system attacks, read that as collective strength.

One entry in my log, dated 14 March 2026, an away match. Bich Tuyen took 22 out-of-system swings across four sets. She scored 8, missed 5, was blocked 3 times. Unremarkable numbers. But look at the situations: across those 22 rallies, the ball reached her at an average height of 2.3 metres relative to the net, and the opposing block had at least two players set in 19 of them. That is work that appears on no individual leaderboard.

Serving: a tax that keeps rising

Serving in my log is the story of an investment that has not paid back.

The ace-to-error ratio rose steadily across five seasons, from 0.61 to 0.79. It sounds like progress. But read it correctly: for every 100 service errors, the team scored 79 direct aces. In other words, every ace cost 1.27 service errors. In women's volleyball, where a set is usually only 25 points and the average side-out rate of strong teams sits near 55%, a service error is a free point for the opponent. An ace is only a point you earned yourself.

The trade-off is simple arithmetic. If an opponent's side-out rate on a normal reception is 55%, a serve with a 30% error risk only pays if it drags that rate below 38.5%. In my data, the team's hardest serves — the top 15% by speed — dragged opponent side-out down to 41.2%. Below the break-even threshold.

This is where I want to be blunt: Vietnam's women's serve is currently a tax on itself, not a weapon. It creates a feeling of aggression, produces spectacular rallies, draws gasps from the stands. But it has not crossed the mathematical break-even point.

The notable exception is serves aimed at the seam between zones 1 and 6. In that group the team's error rate was only 14%, while the share of receptions the opponent handled at P2 or worse reached 61%. That is a profitable direction. The problem is the team used it on only 18% of serves.

Blocking: a number fading out

Blocks per set fell from 2.4 in 2026 to 2.0 in 2026. The drop looks small, but in women's volleyball a set contains roughly 45 to 50 rallies, so 0.4 blocks is close to 1% of all rallies. Across a season, that is about 40 points.

Splitting blocks by type clarifies the picture. Read blocking — reacting to the setter's release and moving late — held roughly steady. Commit blocking, where the middle blocker must commit before the setter releases, fell sharply. In 2026 the team had 1.1 commit blocks per set. In 2026, 0.6.

This is a direct consequence of declining reception quality. When the opponent can run a quick attack at position 3, the middle blocker must choose between tracking the middle hitter and releasing to the pin. When your own reception quality falls, your setter handles more out-of-system balls, the team loses rhythm, and the middle blocker has no time to read. The loop closes: lose reception, lose the quick attack; lose the quick attack, the opposing block stands still; the block stands still, and your hitter must swing through two sets of hands.

Transition: where the season was decided

There is one indicator I compute that I have not seen in any Vietnamese volleyball stat sheet: the conversion rate of defence into points — the share of successful digs that become a point within the next three touches.

In 2026 that figure was 42.1%. In 2026, 38.6%.

The decline matters more than it looks. In women's volleyball, where raw attacking power is often insufficient to break a good block, points from transition situations — after a successful dig — account for about a third of a match's total points. I counted 34.2% of the team's 2026 points coming from this group, against 38.7% in 2026.

Reviewing video for causes, one pattern repeats. After a successful dig, the team most often attacks to the left pin rather than exploiting the gap in the middle. In 61% of transition rallies the ball went to the pin. That is above the general baseline and reflects a safe choice: when the system is unstable, you give the ball to a familiar pin hitter.

Here I want to recall something I learned during the pandemic shutdowns. In the middle of the pandemic, I recounted history and found that every cycle wears a familiar face. A team that loses its transition ability will gradually lose its ability to withstand pressure at the end of sets, and that does not show up in attack efficiency — it shows up in the win rate of sets decided by two points or fewer. In 2026 the team won 44% of those sets. In 2026, 58%.

Home and away: a gap tactics cannot explain

In my log, the perfect-reception rate at home in 2026 was 41.2%. Away, 34.4%. A gap of 6.8 percentage points.

This number does not show up in the opponent's serve or block columns. It is not because opponents served harder — I checked, and the average serve speed faced away was only 2.1% higher than at home. It is not the floor or the lighting.

Most of that gap sits in a variable my sheet files under "context" and cannot quantify: crowd noise. I measured it with a sound app across three home and three away matches. At home, average noise when the team was receiving was 86 decibels. Away, when the host was serving, average noise was 94 decibels — aimed at the visiting side. That acoustic gap does not bend the ball. It slows a receiver's reaction by a few hundredths of a second, and at this level a few hundredths of a second is the distance between P3 and P2.

2026 taught me to listen to what the model cannot measure. The 2026 season taught me that what the model cannot measure can still be counted, just not with the same ruler.

The contrarian angle: correlation is not causation, and this article could be wrong too

I have spent most of this piece dismantling a number. Now I have to partially dismantle myself.

What I demonstrated is that attack efficiency under the common convention exaggerates the team's progress. What I did not demonstrate is that the team made no progress at all. Those are different claims, and I do not want to swap them.

At least three alternative explanations survive my data.

First, 2026 may be a coaching-transition season in which the team deliberately accepts lower reception quality to build a new attack system. When a team changes its roster structure, the perfect-reception rate typically falls for one or two seasons before rising to a higher level. Judging a transition process from a single season is an error I routinely criticise in others.

Second, the fall in P0 from 11.7% to 6.9% may signal a better foundation being built. The team has learned not to lose the ball. The next step — learning to receive it in the right place — is harder and may simply take time.

Third, and most importantly: I am reading my own data. My sheet logged 41 corrections in one season. I do not know how many I failed to make because I did not review the video carefully enough. If my grading bias drifts — for instance if I was stricter early in the season and more lenient later — the downward trend in P3 could be exaggerated.

This is the silent part of the model. I write it down not to protect myself, but because a model with no silent part is a model that has not been tested.

But even subtracting all of that, one thing holds, and it is what I want readers to carry away. It is the gap between side-out on P3 and side-out on P2 — 71.4% against 48.2%. That gap does not depend on how I grade. It is too large to erase with measurement error. And it says that every dong invested in hitting harder will return less than every dong invested in receiving better.

In the other direction, I must also be careful about applying this model to the men's game. I tried; it does not run. Block height in the men's game shifts the break-even threshold for aggressive serving substantially, and the natural perfect-reception rate is lower among men. Each time I import a model from one domain into another, I force myself to find at least two structural differences before reusing a conclusion. This time I found three.

Signals to watch in the next round

If I were allowed to look at only one number going forward, I would look at the four-level reception distribution, not the attack success rate. Specifically, I would track the P2 share. If the P2 share stays above 25% while P0 remains under 8%, the team is on the right path of system transition and should be judged on the 2027 season, not 2026. If the P2 share passes 28% and P0 starts rising again, the new system is failing and the old structure should return.

I would also track the serve direction into the seam between zones 1 and 6. Its current share is 18% of all serves, while it is the highest-yield direction in the entire season's log. If that share passes 25% within a few months, the coaching staff is reading the same table I am. If it does not move, they are reading a different table — and I would like to know which columns it has.

And I will keep correcting my own notation. This season it was 41 times. If that number falls next season, it will not mean I have become a better charter. It will mean I am looking less, or trusting outcomes faster. Both are bad news.