Nine Analysis Sections, Not a Single Number: When Volleyball Data Chooses Silence

The sheet of paper sat on my desk all night. Nine sections. Every one had a heading, a table, a conclusion line in bold at the bottom. And every one repeated the same phrase: insufficient information. I flipped through it a third time, folded it in four, and set it beside a cup of coffee gone cold. In twenty years of reading volleyball data, I had never held an analysis that looked so complete while being so hollow inside.

A draft does not have that structure. This was the output of a process that had finished running, been numbered, and been carefully formatted. Only the raw material — the list of input facts — was left blank. It should have stopped, thrown an error, printed a single line saying there was nothing to say. It did not stop. It still produced nine handsome sections, each marked “insufficient information,” as if that emptiness were itself a finding.

A flawless document with no information is more dangerous than a blank page, because it makes the reader believe someone actually did the work.

I write this article in exactly my trade: volleyball. And with exactly the habit that has become instinct — numbers first, stories second.

Why volleyball is ground where numbers are easy to distort

Volleyball has no xG like football. It offers no single composite metric to compress a match into one figure. So people measure through four main families: perfect-pass rate, attack efficiency, blocks and service errors per set, and dig rate. All of it is logged by the industry-standard scouting software, Data Volley, phase by phase, ball by ball.

Precisely because every phase must be charted by hand, volleyball is a sport where the capacity to misread statistics is far higher than in football. I have seen enough to believe that.

The first time I realised this was not in volleyball but in football, in April 2026. I sat before three screens, rewatching Leicester 2–4 Everton. The press praised a striker, but xG data from Understat showed Leicester created only 1.2 xG against the opponent’s 3.8. One miss dragged wide of the goal was worth 0.65 xG. From that day I stopped trusting live commentary. 2026 taught me how to listen to what the model cannot measure.

Volleyball has a pair of metrics that mislead in exactly the same way: attack success rate and attack efficiency.

Attack success rate takes the number of attack points and divides by total attempts. It sounds simple. But it ignores the two most important things: the number of balls hit out and the number of times blocked. An attacker can post a 45% success rate, looking very handsome, while in reality every point he wins he pays back with nearly two dead balls or stuffs. Attack efficiency is the real number: points minus errors and blocked attempts, divided by total attempts. The same attacker, on efficiency, may be down to 18%.

In my notes, this is the most common error in volleyball media. The statistics do not lie; people simply choose which figure to print.

Where the gap sits

When I say an analysis pipeline has two stages, this is what I mean. Stage one extracts facts: who, playing where, in which competition, what figure, from which source, on what date. Stage two takes those facts and only then analyses — tactics, data, opponents, risk. The entire strength of stage two rests on being allowed to speak only about what stage one delivers.

And stage one that day returned empty.

No title. No source. No list of facts. The “related entities” field did still contain text — but that text was an instruction: “identify from the facts above,” pointing at a list that did not exist. That is no longer missing data. It is a structural defect. A cell in the table pointing at itself, then coming back empty-handed.

And yet stage two still ran.

If stage one delivers facts, stage two can conclude. If stage one stays silent, the only way for stage two to still “have work” is to invent facts. That is the greatest temptation of this trade. I have witnessed it at larger scale: analyses sprouting in neat rows after every tournament, each with a confident voice, when the facts are nothing more than a few figures copied from a table anyone can open.

International volleyball has a competition system thick enough to generate mountains of facts each year. The FIVB’s VNL is an annual commercial competition, counting for world ranking points, running over many weeks, one cluster of matches per week. Then the continental championships, national leagues like Italy’s Serie A1, the Turkish league, Poland’s PlusLiga, Japan’s SV.League, and Vietnam’s V-League. Every match is a sea of numbers, but that sea is worth something only when it is charted carefully and read correctly.

Without facts, all analysis is just storytelling.

The trap: correlation is not causation

This is where I have to say the hardest thing.

People prefer a story to a table. When Vietnam’s women’s team wins in a row, an ode to “the spirit of Vietnamese women’s volleyball” appears at once — something beautiful, something unmeasurable, and therefore something unfalsifiable. It is exactly what I ran into in 2026, when I backed Croatia: Croatia is not a fairy-tale, they are a problem that must be solved again from the start.

Readers imagine a team wins on desire, while I must say plainly: most surprise winning streaks reduce to numbers that can be calculated. Croatia’s run in 2026 reduced to distance covered, progressive passes, and a measurable pressing index. Matches in the VNL or in domestic leagues are the same. A team winning three straight may simply have met three weak opponents, with no explosion at all.

And here is the irony. I accuse the media of distorting numbers, yet my own model has its holes. 2026 taught me that a model has gaps, and that the skill is knowing when numbers fall silent. Today, I hold an analysis with nine sections and not one figure. The most honest thing I can do is not to fill it with story, but to say that it is empty.

A few years ago, during the pandemic, I had a habit of sitting and recounting history. In the middle of the pandemic, I counted history again and saw every cycle wearing a familiar face. How a surprise streak begins, how it gets blown up by the media, how it collapses. I saw nothing new. Looking closer, every time the data went silent, someone talked loudly in its place.

Why I will not fill the gap

In analysis there is a fine line between inference and invention.

Inference needs at least one verifiable anchor. Without that anchor, every word is merely a flourish of the pen. I could easily write a piece on Vietnam’s men’s volleyball team at some tournament, assign them a few plausible-sounding figures, add a tactical remark, and readers could not tell which numbers were real and which I had just thought up over coffee. I could do that in thirty minutes.

I refuse.

Not out of some grand morality. But because it contradicts the very principle I have built over twenty years. If I invent a table of figures today, then tomorrow the real tables I publish will be worth no more than the invented one. The voice of a data person is built and lost bit by bit.

People in the trade call this a null result — a kind of conclusion everyone hates, because it gives no one anything to cling to. Saying “I don’t know” is not weakness; it is the only sign that the speaker can still tell knowing from not knowing.

What is worth noting is that those nine sections, formally, still looked good. They had complete structure, tables, bold conclusion lines. Placed beside a real analysis, to the naked eye, a reader could hardly tell them apart. That is why I call it the greatest danger in my reading room: an empty analysis looks exactly like a full one.

Consequences seen from the long cycle

I believe in long cycles, not in short-term illusions. In volleyball, that means I care more about a season, an Olympic cycle, than about one high-profile match.

Seen at that scale, a pipeline returning empty is no small thing. It exposes several things at once.

The sports-analysis industry is running faster than its own capacity to verify. People pour out content that is thicker, more frequent, more automated, while the quality-control system upstream is thin. An empty list of facts should have stopped the whole chain. It did not. Which means that gate does not yet exist.

Readers are being placed in a position where they cannot tell the difference. When every analysis shares the same form — full sections, full tables, full conclusions — credibility shifts from the form to the byline. That, in the end, is a good thing: readers are forced to remember who wrote a piece, rather than trusting a template.

And, most important to me: it reminds me that a model has limits. A good model is not one that always answers. A good model is one that knows to stay silent when there is not enough data — and knows to say loudly that it is staying silent.

What I am waiting for in the next round

I cracked the door and looked at the street the next morning, while the sheet still lay folded in four on the desk.

What I am waiting for is not a new full analysis. What I am waiting for is a gate that knows how to stop at stage one: if there are not at least three verifiable facts, do not run stage two. If there is no title, no source, no date, then do not call it an analysis. For volleyball, that gate would mean: at least one named team, one competition, one metric with a source and a scope — a match, a leg, or a full season.

Only then will I allow myself to write a conclusion line in bold.

Until then, the empty analysis will stay on my desk as a reminder: this trade does not die from a lack of data. It dies from replacing data with a story, because that is easier.