Blank Cells in a Billiards Dossier: Notes from Liverpool on the Limits of Analysis

Two in the morning in Liverpool, and I am sitting in front of a document whose every cell is empty.

The title line reads N/A. The source reads N/A. The article type reads N/A. The core-viewpoints column is blank. The information-points list contains nothing. The entities involved have not been identified. The file runs past ten pages, with tables, with subheadings, with a six-row risk matrix complete with likelihood, impact and mitigation columns. Its entire content is a sequence of polite negations.

My first reflex, after nine years in the trade, is to fill the blanks. That reflex was trained into me. From 2026, when I began writing regularly for sports outlets in England, I learned that people pay for a story with a beginning and an end, not for hesitation. An empty file is therefore a tempting invitation: put a hypothesis in it, a name, a turning point, and the piece will stand up on its own.

That night I did not fill it in. I did not delete it either. Because that empty file, in a very specific sense, is the most honest document I have received in months.

The economics of the template

Modern sports analysis runs on a system of templates. There is a template for match reports, a template for profiles, a template for forecasts, and a template for the short answer-pack format that search engines increasingly favour: one tight conclusion, a handful of facts, source attribution, a few follow-up questions. Templates exist to save time, and they do that very well.

The problem lies elsewhere: a template always has enough cells, even when there is nothing to put in them.

I cover billiards tournaments for the British market from Liverpool. My job is to read billiards as a geometry of error and belief: a player stands at the table, chooses between attacking and playing safe, and every choice leaves a mark on the cloth. Across a ranking event lasting several days, I need very specific things. Which discipline: snooker, 9-ball, Chinese 8-ball or carom. Which format: how many frames, which rounds are long, which are short. What the table is doing: cloth speed, room humidity, temperature, ball set. And the entry list, with each player's recent form.

Without those, I have nothing to analyse. Not because I lack ideas. Because sports analysis, at its deepest level, is the business of measuring the gap between what is said and what is done.

I have received requests along the lines of: write a thousand words on this player, explain why he is on the rise. I always ask two questions back. Is he on the rise compared with himself three months ago, or compared with the standard of the field? And what data are we comparing with?

Most of the time, the person commissioning has no answer to the second question. They have a feeling, and the feeling is right in the way feelings usually are: it captures part of the situation and ignores the rest. My job is to find the ignored part, not to decorate the captured part.

When the data does not arrive

In 2026 I was sixteen, in sixth form in Liverpool, writing a personal blog on the wide-attacking patterns of the Liverpool U18 side. I counted the runs myself across fifteen matches. Andrew Robertson, then twenty-two, had a tendency to stretch the flank about 0.8 seconds earlier than the other full-backs in the same system. I wrote five thousand words just to prove that deviation.

The piece got exactly twelve views. One of them was a local scout who left a comment asking how I collected the data. We exchanged emails about statistical probability in pressing, and that is how I entered the trade.

The lesson I took was not how to write for a large audience. It was this: if you measure something nobody measures, you own something nobody owns. Value sits in the measuring, not in the noise.

That same year I tried to apply the method to billiards and failed badly. I sat counting how often a player walked around the table before shooting, convinced I was measuring thinking rhythm. After three events, I realised I was measuring broadcast quality: on tables far from the cameras, players were shown in close-up less often, so fewer walk-arounds made it to screen. I had measured the camera, not the man.

That mistake taught me something I still use daily: before trusting an indicator, look for the ways it can be wrong for reasons unrelated to competition.

Error is where reality signs its name

In June 2026 I was seventeen and spent the whole summer holiday watching the World Cup in Russia. In the Germany–South Korea match I logged forty-seven turnovers by the German midfield inside the final thirty metres before they went out. I dug into their high pressing and found they succeeded only twice in eleven attempts to press ahead of the opposition penalty area.

I wrote a four-thousand-two-hundred-word piece titled Atlas of a Collapse. It circulated lightly among tactical fans, was shared on a large European football forum, and later became the foundation of my statistics dissertation.

From that period I kept one line, and it holds for billiards no less than for football: error is where reality signs its name.

When a player misses an easy ball, viewers call it nerves. When a player leaves the cue ball twenty centimetres off and loses the position on the next shot, viewers call it bad luck. Both labels close the book too early. Error is not rubbish to be swept away. It is the only data the table will disclose about the true structure of a frame.

I have spent years counting what the scoreboard does not count: the moment a player decides to switch from attack to safety, the average distance the cue ball travels past its target, how often a break-building pattern is repeated after failing three times in a row, how often a player changes rhythm between the fourth and fifth frames.

None of those numbers appears on the electronic board. All of them live in the cloth, and vanish when the referee collects the balls.

What to measure on a table

If I had to rebuild my whole method for a billiards event, I would start with four groups of measurements.

The first is position. For every shot I log where the cue ball stops and compare it with where the next shot needs it. The distance between those two points is the most important indicator on the table, and almost nobody publishes it.

The second is choice. On each visit a player has at least two valid options. I record the option taken, the option declined, and the state of the table at that moment. After roughly two hundred visits, a pattern appears: this player plays safe below a certain estimated pot probability and attacks above it. That is his risk-appetite map, and it is far more stable than form.

The third is rhythm. Not shot-clock time as such, but its variation. A player in control tends to hold a flat rhythm. A player who is anxious tends to speed up after a mistake, and that acceleration is an earlier signal than any scoreboard.

The fourth is environment. Cloth speed changes with humidity, and humidity changes with the number of people in the room. Across a long event, the same shot can travel further in the first frame of an evening session than in the last frame of an afternoon one. Good players read that and adjust; players trusting their routine do not.

These four groups need no expensive equipment. They need time, a notebook, and the patience to tolerate being wrong.

Absolute belief

There is one pattern I meet so often that I began to doubt myself.

It is the moment a player has chosen the right option, executed the technique correctly, and still lost, because he believed in that option so completely that he stopped observing the table. The cloth was faster than he thought. Humidity made the cue ball drift an extra hand's width. The opponent had changed tempo two frames earlier. Someone who believes absolutely in his plan will not receive those three signals, because he is no longer looking for them.

A high defensive line does not collapse because of the tactic, but because of absolute belief in the tactic. On a billiards table, that high line is a sequence of safeties executed exactly as taught while the table changed long ago.

That is why I do not trust analyses made of tables alone. A table can tell you player X won 68 per cent of his safety exchanges at an event. It cannot tell you whether he won those 68 per cent by reading the table, or because the opponents at that event played safety badly. Those two causes point to opposite forecasts for the next event, and the table cannot distinguish them.

The variable the model forgot

In March 2026 I was nineteen, a second-year statistics student in Manchester, and world sport had stopped. I lost my live data to analyse. My response was to rewatch fifty-seven Manchester United matches from the 2026-19 season over three weeks, building myself an imaginary database on attacking timing when there is no pressure from the stands.

When football returned in June with empty stadiums, I found something that contradicted my own prediction: long-pass success for English sides fell 12 per cent. Teams passed sideways more but risked penetrating less. I then spent six months rewriting my entire theoretical framework.

The empty-stadium season deleted a variable no model could encode: noise.

Billiards has the same variable, and it bites harder. In a big arena, noise is not merely sound. It is a social signal: it tells the player that this moment matters, that people are waiting, that this shot will be remembered. In a silent room the same shot can be played with an entirely different state of mind, and the technical outcome can be identical while its meaning is completely different.

No model of mine can encode that. I tried, and failed. The only way I handle it is to record it in words, in the margin of the data table, and accept that the words cannot be added up.

The twenty-centimetre gap

In July 2026 I was twenty, freshly through my third year and interning at a small sports-data company in Liverpool. In the Euro semi-final between Italy and Spain, I became obsessed with how Marco Verratti moved diagonally into the space between the two opposition central midfielders, each time creating roughly two square metres for Lorenzo Insigne.

I wrote a five-thousand-five-hundred-word piece titled Italy's Rotating Spatial Model, using heat maps I built myself from twenty thousand match data points, a skill I taught myself from online Python courses during the pandemic year. The piece was shared by a European football podcast, and over the following forty-eight hours it drew more than a thousand reads and a few collaboration offers. I turned them all down, because I wanted to dig deeper before publishing.

What I carried from that piece into how I read billiards is one line: tactics are not on the whiteboard; they are in the gap between two runs.

On a billiards table that gap is not two square metres. It is twenty centimetres. A player who leaves the cue ball twenty centimetres off the ideal position turns a high-percentage pot into a shot that must be played left-handed, or must be played safe. The viewer sees the miss on the following visit. I see where the cue ball stopped on the previous one.

That shift, from outcome to cause, is my entire trade. And it is why I cannot write a piece about billiards when there is no event name, no table condition, no player.

An empty template is more honest than a full one

Back to the two-in-the-morning file.

The file describes a billiards article that has no title, no source, no viewpoint, and no identified entities. Following the process I always use, it declares that second-stage analysis cannot proceed, and that every further conclusion would be unfounded speculation. All nine analysis categories are marked insufficient information. The information-value table scores the lowest rating on all four axes: competitive, industry, timeliness, reference.

I read that section three times, and what I saw there was the honesty of a process, not its failure.

Had that file been filled in within ten minutes, it would have had a title, a player, a tournament, a turning point, a forecast. All of it could have read smoothly. And all of it would have been unverifiable, because it came from nowhere. The sports-content industry is currently designed to reward exactly that kind of smoothness.

An empty template is not a defective product. It is a correct measurement: at this moment, there is nothing to say.

Every starting line-up is a hypothesis waiting to be refuted by reality. So is an empty file, except that it admits it earlier.

My industry has a habit I consider dangerous: demanding the answer before the question exists. Fast-answer templates, compact answer packs for search engines, short formats that require a conclusion within sixty words — all of them assume the data has arrived, that the facts are clear, that only the phrasing remains. Reality is the reverse. Most of an analyst's time goes on establishing whether there is enough to say anything at all.

And when the answer is no, the only honest way to write is to say so.

The price of fluency

There is a genre of piece I refuse to write, even though it gets read widely.

It is the piece about an unknown player who goes deep in an event, carrying the implicit message that the development system of some billiards nation is on the rise. I have watched enough to know that runs like that usually come from two things that are very hard to encode: a favourable draw, and one match in which every difficult shot dropped. Neither of those proves a system has succeeded. To prove that you need ten years of data, not one week of competition.

The content industry hates that waiting. But the waiting is where most of the real value sits.

At the same time, I understand why readers want an answer immediately. They follow every match, every frame, and they need somewhere to put their belief. The problem is that when we hand them a confident answer built out of nothing, we are not feeding their belief. We are teaching them that everything can be known at once, and when reality turns out to be more complicated, they will turn to doubting the sport itself.

I saw that during the empty-stadium season. A great many forecasts were issued with great confidence, and a great many were wrong. The failure was not in failing to predict correctly. The failure was in not saying that we were predicting.

So what to track

I will start next season with a single variable, and I would suggest readers do the same.

Do not track pot success rate. It is the noisiest and most misleading indicator on a table, because it depends on whether the player chose easy balls or hard ones. Instead, track where the cue ball stops after every safety exchange. Log the distance between the actual resting position and the position the next shot requires. Over about ten frames you will have a curve. That curve tells you whether the player is controlling the table or being spared by it.

If the curve is flat, he is in control. If the curve swings wildly while results stay good, he is living off difficult pots going in. That kind of living has a shelf life, and the shelf life usually expires precisely in the long-format rounds.

And once you have that curve in hand, do something harder: write down what you do not know. Mark it. Leave the cell empty.

When the match ends, indicators lie more subtly than any player. The only way not to be fooled is to keep the empty cells, and come back to fill them when the data genuinely arrives.

The two-in-the-morning file is still on my machine, still blank, and I have not deleted it. Next season, if some player makes me open it again, I will know exactly what to write in the first cell.