When Data Stays Silent: The Trap of Blank Cells in Billiards Analysis

He pushed the sheet of paper across the table. On it was the name of a player he would meet in the qualifiers for the national 9-ball tournament in Hai Phong next month. I opened my spreadsheet, typed the name, and sat still.

The spreadsheet returned four rows across eighteen months. Four matches. No average shot time. No rate of retaining the table after the break. Nothing on how he handles safety play when pushed into trouble. The rest of the sheet was blank.

He asked: "So he's weak, right?"

I closed the laptop. That question was not about the other player. It was about me — about a reflex that took me nearly four years in this trade to name: turning a blank cell into a zero.

In billiards analysis, people are taught to fear a wrong number. Very few are taught to fear a missing one.

To be fair: Vietnamese billiards does not lack talent. It lacks data infrastructure. A single Premier League football match generates thousands of data points per minute. A 9-ball match at an open domestic event, if nobody sits down to record it, leaves behind only a scoreline, a champion's name, and a few photos on social media.

I do not say this to complain. I say it to frame the problem correctly. When data infrastructure is thin, what is missing is not analytical capability. What is missing is raw material. And when raw material is missing, an analyst has two honest choices: say he does not know, or go find another source.

Choosing the third option — filling the blank with a story — is the easiest, most common, and most expensive path.

I have done it myself. In 2026, at seventeen, I confidently applied expected-goals figures to a V.League round 18 match between Hai Phong and Sanna Khanh Hoa. The numbers showed Hai Phong generating far more chances, and I concluded they would win comfortably. The match ended the other way, and goalkeeper Tran Buu Ngoc on the Khanh Hoa side made seven saves.

My model was not wrong in its arithmetic. It was wrong because I let a blank cell — goalkeeper form, a factor outside the formula — become a zero, meaning "negligible."

That was the first time I understood: the most dangerous thing in analysis is not a wrong number, but a missing number quietly filled in with an assumption.

A year later, in the summer of 2026, I wrote a piece on Mexico's win over Germany in the World Cup group stage. Germany had 66 percent possession and made 613 passes. But Mexico's PPDA was 8.4 — meaning Germany was allowed an average of only 8.4 passes before losing the ball. I concluded Germany would exit early.

The piece was laughed at. Two weeks later Germany lost to South Korea and went home. Twelve emails arrived afterwards, mostly conceding I had been right. But the lesson I kept was not "I was right." It was: if I was right by reading a metric most people ignore, then there are countless cases where I will be wrong by reading a metric nobody is around to check for me.

Two years later, in the summer of 2026, I met that lesson again at a larger scale. The Bundesliga returned without crowds. I collected all 81 matches from the final nine rounds of 2026/20. The home-win rate fell from 44.7 percent to 33.3 percent. Away teams' average expected goals rose from 1.15 to 1.32.

A forum moderator pointed at the sample size and said I was inflating random noise. He was not wrong to be suspicious. He was wrong in his conclusion. He turned a region of data that was not yet thick enough into a confident claim that "nothing happened," when the correct thing to say was "not enough to conclude."

I ran the test, got p = 0.045, and published the result along with every caveat I had about my own limits. Not to show off. But so readers would know exactly where I stood on the map of certainty.

Two years after that, at the 2026 World Cup in Qatar, I watched Japan beat Germany. Japan had far less possession and created fewer chances by volume, yet won. Look only at possession — a complete, pretty, easy number — and you conclude wrongly. The truth lay in another metric, one less often shown on television.

In billiards the story repeats almost verbatim. Pot success on the break is the number shown on the scoreboard. The quality of the position left after the break — whether the opponent gets a real chance to respond — usually goes unrecorded. A player can pot a ball on every break and still lose the set, because his break creates no "dead" table. The scoreboard records "pot made," not "break that hurts."

That is a systematic data gap: measurement systems are built to count, not to explain.

For any dataset handed to me, I keep a fixed checklist. One, who recorded the source, and were they at the table or reading from a report. Two, how long does the sample run, and what changed for the player over that span — cue, fitness, coach. Three, how is this metric calculated, what gets counted and what falls outside the count. Four, what is the blank part of the sheet hiding.

The fourth is the question I ask most and the one answered least.

Back to that four-row spreadsheet. I told my student to write by hand. Over two weeks he logged every match of the opponent he could find video for, including friendly matches with no results online. He timed the gap between turns at the table when the other player was pushed into defence. He noted by eye: the rhythm of standing up, of bending down, the pause before a decisive shot.

Three weeks later, his spreadsheet held eleven matches and thirty-four self-recorded metrics.

That opponent was not weak. He had simply appeared rarely. And what the four-row sheet actually said was not "weak" but "not enough data to know."

That is a difference that can change how my student prepares for an entire match.

There is a paradox I have to state plainly, even though it complicates my own trade: data analysts are moving deeper and deeper into the practice room, and they are not always right to be there.

When a model says Player A is stronger than Player B, it is comparing two numbers. But a billiards match does not happen between two numbers. It happens between two people, on one evening, on a table with its own speed, under the pressure of a single qualification spot.

Some things are not on the spreadsheet. The breath before a decisive break. The way a player bends down for two seconds longer when he begins to lose confidence. The applause that falls off the beat when the stands are full. I cannot measure those with a formula. But I can note them down.

And here is the part where I must argue against myself: if I admit data is missing, I risk turning every unfavourable conclusion into "not enough sample." That is the reverse trap of the cautious person. Transparency about limits is a strength, but if it is used to dodge every conclusion, it becomes a kind of intellectual cowardice.

I have to set myself a deadline. If the data is not sufficient within two weeks, I write the conclusion as a hypothesis, publish the unknown part, and accept the risk of being wrong. Because in betting and in journalism, standing still and waiting for data is not neutral — it is also a decision, just one that goes unrecorded.

Another mistake I see often in billiards analysis is reading correlation as causation. A player has a high win rate when he breaks first. Sounds plausible. But if that player is always drawn against weak opponents in the first round — where most matches are broken first — then the rate says nothing about break skill. It says something about the draw.

The same set of numbers, two readings, two opposite conclusions. What separates them is not arithmetic. It is a question about context.

I lost many rounds of misreading numbers to learn this. Not because I calculated wrongly, but because I asked the wrong question. I asked "how big is the number" when I should have asked "under what conditions was this number produced."

There is one more variable that Vietnamese billiards data almost never records: the home factor, understood as familiarity with the table. A player competing in his own city, on the type of table he practises on daily, usually has an edge in table speed and cushion rebound. That edge appears in no scoreboard. When events are moved or held without crowds, the edge disappears, and its disappearance leaves no trace in the data — exactly as home advantage in football vanished without many people noticing until they looked at the aggregate numbers.

That is why I always keep a separate "table" section in every billiards report I write: table type, cloth speed, hall humidity where known. These things are not as pretty as a percentage table, but they explain more anomalous results than blaming "form" ever does.

Seen more broadly, the transmission chain of Vietnamese billiards has the same problem. Upstream are practice halls and clubs, where data is close to zero. In the middle are players and events, where data exists but is not standardised. Downstream are media, sponsors and derivative markets, where people need numbers to make decisions but receive only stories.

When the first link is empty, every link after it has to guess. And when guessing, people tend to pick the easiest story to hear: the rising newcomer, the king's return, the end of a dynasty. Those three labels sell better than a spreadsheet that is thirty percent blank.

I do not write to persuade anyone. I write so that data has a witness.

Now, every time I receive a billiards dataset, I add one more line at the bottom: "What this sheet does not say." That line is always longer than I would like. But it is the most honest part of the report.

I learned this from that four-row sheet. What is frightening is not whether that player was strong or weak. What is frightening is that my student could walk to the table with a belief built on blank space.

Three thousand matches taught me that one match can teach more than all of them. And four rows of data reminded me that sometimes the thing that teaches most is what never got written down.

Vietnamese billiards is growing fast. There will be more events, more players, more people watching. And there will also be more blank cells, because data infrastructure always lags behind the speed of enthusiasm.

The question for next season is not who will win. The question is: when a spreadsheet returns four rows, do we have the courage to say we do not know?