Billiards Follows Football: A Big Opportunity and the Unnamed Data Trap

On the second Saturday of the World Championship at the Crucible, I sat in the upper press row with a notebook ruled into columns. Below me, two players entered the eleventh safety exchange of a deciding frame. The big screen in the arena showed only three things: points, shots taken, and time remaining. There was no column for what was actually deciding the frame: the cue-ball position after each push, the quality of a long-rail cut, the psychological distance between two men when one is forced to hit a difficult ball.

Across seventeen days of the tournament, I counted fourteen occasions when the organisers put the "century breaks" statistic on screen as a measure of form. Each time, the eventual winner of that match had made fewer centuries. That proves nothing about any individual player, but it proves something about how we are looking at this sport: we are measuring the easy part, not the deciding part.

I came to billiards from football, where metrics have become an entire industry. In the summer with the German coach, I learned that pressing is not anger; it is a calculation. But that calculation only means something when you have data to feed it. In billiards, the table is empty.

CONTEXT: A STABLE SYSTEM AND A MARKET COMING APART

Professional billiards, specifically snooker, has a tournament system shaped over nearly half a century. The World Championship has been held at the Crucible Theatre in Sheffield since 2026, with a best-of-35 final spread across two days. Alongside it are the two Triple Crown events — the UK Championship and the Masters — forming the most prestigious trio. The whole system is run by World Snooker Tour under the oversight of the professional association. It is a stable, tidy, and very English structure.

But billiards is no longer only an English story. China has become the second-largest market, with millions of players, dozens of ranking events, and an entire variant of its own — Chinese 8-ball — with different rules and a different table. In the United States, 9-ball retains its own place in the professional tour system. Three rule sets, three scoring systems, three playing philosophies, and three data sets that almost nobody can join together.

The golden generation — the cohort born in 2026, comprising Ronnie O'Sullivan, John Higgins and Mark Williams — is still in the top group after more than thirty years as professionals. O'Sullivan has won seven world titles and holds the record for career century breaks. Higgins has won four, Williams three. Three men, three styles: O'Sullivan favours speed and instinct, Higgins balance and control, Williams relies on arguably the best single shot in history — the left-handed cueing.

These numbers raise a question the available data cannot answer: is this the greatness of three individuals, or the sign of a system slow to change? And if it is the second, which data set is being collected to prove it?

I follow this sport with a method learned from football. Across seven months in Liverpool, I manually counted 312 turnovers in the opposition's defensive third to build six pressing-trigger signal types. The principle is simple: to analyse, you must first code. To code, you must have something to code. With billiards, I found I was coding from memory, not from data.

CORE: THE LOUDEST METRIC IS THE MOST MISLEADING ONE

In every billiards statistics panel, the first thing to appear is the century-break count. It is pretty, it is easy to understand, it suits television audiences. And it is the most misunderstood metric in the sport.

A century break tells you that a player scored one hundred unanswered points in a frame. It does not tell you how many safety shots he had to play to earn the right to the table. It does not tell you the quality of his cue-ball position on the twentieth shot, when the frame was still hanging. It does not tell you when the opponent was pushed into a position where he had to gamble.

A player's real value lies in the shots that never reach the scoreboard: the safety that forces an opponent onto a difficult ball, the push that leaves the cue ball dead, and the decision to sacrifice one frame to keep control of the match.

I started counting by hand at a ranking event last season. Across three quarter-finals, I logged every safety shot and sorted them into two groups: survival safeties and control safeties. The second group — the pushes that leave the cue ball tight to the cushion and force the opponent into a worse return — accounted for under a third of all safeties, yet determined who got to the table first in the next exchange. No official statistics panel separates these two groups.

That is the first gap. And it is not a small gap. It is the gap between what the sport shows off and what the sport actually consists of.

BREAK QUALITY VERSUS BREAK QUANTITY

There is a structural confusion between two concepts that billiards analysts routinely merge: the number of breaks and the quality of breaks.

Break count is a cumulative metric. It grows with time, with frames, with opportunities. A player who plays more frames will have more breaks, regardless of efficiency. It is a metric of opportunity, not of ability.

Break quality is an entirely different metric. It measures whether a player retains cue-ball control after opening the pack, whether he pushes the cue ball into a favourable zone, whether he forces the opponent to sit still. It does not accumulate over time. It fluctuates frame by frame, table by table, night by night.

In football, I learned a similar distinction when analysing passing. A team that completes seven hundred passes is not stronger than a team that completes four hundred. What matters is where those four hundred passes go, at what speed, and in what situations. Russia 2026 taught me one thing: data does not win matches, it only wins before kick-off. Billiards is the same. Break count does not win frames. Break quality does.

THE FORMAT PROBLEM: WHEN DATA CANNOT BE COMPARED WITH ITSELF

This is the point I consider the most underrated in the entire debate about billiards data.

A best-of-7 match and a best-of-35 match are not the same sport. In terms of rules, they are identical. In terms of psychological and tactical structure, they are two different disciplines. In best-of-7, errors are punished immediately. In best-of-35, errors can be nurtured, hidden, and offset by four hours of patience.

So when you pool every match into one dataset and take an average, you do not have a dataset about billiards. You have a dataset about some average that does not exist in reality. It is a methodological error, and it is happening at system scale.

A century break in a best-of-7 qualifier and a century break in a best-of-35 Crucible final are two different statistical objects, and their sitting in the same column in a dataset is a failure of design, not of the player.

I once made this error in football. In 2026, analysing 40 matches at the World Cup, I pooled group-stage and knockout matches into a single sample. A colleague in Europe pointed out that knockout matches have a completely different risk structure, and that pooling them blurred exactly what I was trying to find. He was right. I had to split the entire dataset and rewrite my conclusions.

Billiards is at that exact point, but nobody has stepped forward to split the dataset.

THE SAFETY PROBLEM: THE SUBMERGED PART OF THE ICEBERG

If there is one place where billiards analysis could overtake football, it is the safety game. And if there is one place where billiards analysis is failing most spectacularly, it is also the safety game.

In any elite frame, the majority of the time the balls are in safety exchanges. That is where the positional war is fought, where the cue ball is moved centimetre by centimetre, where a slightly heavy shot opens the door for the opponent. And it is where no standard metric exists.

I tried to build such a metric myself. Over about two months, I logged every safety exchange at ranking events, classified by three criteria: cue-ball distance to the cushion, difficulty of the target ball, and frame situation. Preliminary results showed a trend I believe is real: the player who wins the safety exchange — not the player who scores more — is the one who controls the tempo of the next three frames.

But this is one man's data, logged by hand, over two months. It is not proof. It is only a signal strong enough to keep me going.

The problem is that this signal is not being collected at industrial scale. Football has dozens of companies tagging every phase of play in every major league. Billiards has no company doing the same at the level of the safety shot. People tag break shots because they are visible on television. People do not tag safety shots because that requires an expert to classify them, and experts cost more than algorithms.

THE GOLDEN GENERATION AND THE AGE PARADOX

My years of match-watching lead me to an observation hard to explain with any existing data model: the cohort born in 2026 is still in the top group after more than three decades.

In most speed-based sports, the age curve is a parabola. Athletes peak around twenty-five to thirty, then decline. In billiards, that curve is nearly flat. Three men in their late forties to early fifties still beat opponents twenty years younger over long matches.

Several explanations are offered. That billiards does not demand physical fitness. That experience in billiards compounds exponentially. That the table does not change.

I do not fully believe those explanations. Billiards demands sustained concentration over many hours, and concentration is a function of fitness. Experience in billiards has value, but it also creates habit, and habit is the enemy of adaptation. The table does not change, but table conditions — humidity, temperature, cushion wear — change daily.

The explanation I consider more accurate is a systemic one: the pace of innovation in billiards is slower than the pace of human ageing. If it takes a player fifteen years to learn the sport's tactical patterns, then a thirty-five-year-old player still being in the top group is not a paradox. It is the inevitable consequence of a slow-innovating system.

And that leads to an uncomfortable question: if the technical base of this sport innovates slowly, where is the incentive for a young generation to catch the golden generation?

THE CHINESE MARKET AND THE TALENT PIPELINE

If there is one region changing the face of this sport, it is China. The number of players there is larger than anywhere else, and more importantly, the number of tables there is larger than anywhere else. Professional playing infrastructure depends on the number of tables, and the number of tables depends on the sport's popularity as mass recreation.

That means China's billiards talent pipeline operates on a different logic from Britain's. In Britain, players grow out of local clubs, amateur events in small halls, and family or regional tradition. In China, players grow out of academies where children are trained systematically from an early age.

The academy system generates data. It records every session, every drill, every progress metric. That is exactly what the British club system cannot do systematically. In other words, the country that may dominate billiards for the next thirty years is not the country with the best current players, but the country that collects the most data about the process of developing players.

That is a claim that can be wrong. But it can be tested: if over the next decade the number of young Chinese-trained players entering the top 32 rises faster than the number of young British-trained players, this argument holds. If not, I am wrong, and I will be the first to say so.

CONTRARIAN VIEW: THE TRAP OF COPYING FOOTBALL

Here I have to be blunt, even if it goes against the general direction of an entire industry.

Billiards is trying to copy football. Analysis conferences, long-form articles, proposals to build data platforms — all take football as the model. I understand why. Football has the most mature data machinery of any sport, and anyone wanting to modernise another sport looks there.

But there is a structural difference that the copying ignores.

Billiards has the appearance of a Markov process: each shot is a state, and the next state depends only on the current state. This makes people think billiards can be modelled as a sequence of independent decisions, and from there build shot-by-shot predictive models. That is a foundational error.

Table conditions are not a background variable. They are a state variable: humidity shifts within an evening, cushions wear down across tournament days, and a player's sense of pace depends on things recorded nowhere. A Markov model that ignores these variables will predict correctly under standard conditions and incorrectly under real ones.

In football, modern models learned this long ago. They do not model each touch independently; they model match state — territory, pressure, time remaining, score. If billiards copies football at the passing-model level while ignoring the state-model level, it will copy exactly what football left behind a decade ago.

The second trap is the sample-size trap. In a season, a professional plays a few hundred frames. Split by format, by event, by table, by opponent, and each data cell becomes too small to be statistically meaningful. Football has hundreds of phases per match to fill the cells. Billiards has dozens of shots per frame. This sample shortage cannot be fixed with better algorithms. It must be fixed with better research design — and better research design requires accepting that some questions will have no statistical answer, and must be answered by qualitative observation.

I know this sounds like a step back. But I have followed this sport long enough to know that sometimes the right thing is not the most modern thing.

THE THIRD TRAP: AN EMPTY FIELD IS NOT A CLEAN BILL OF HEALTH

This is the point I most want to make clear, because it affects how people decide.

When a dataset is empty, people tend to read it as "no problem". No injury report means the player is fit. No safety metric means safety does not matter. No match-fixing data means no match-fixing.

None of those inferences is valid.

In epidemiology this is called a "false negative". A test that fails to detect a disease does not mean the disease is absent; it means the test was not sensitive enough. In sports analysis we rarely use this concept, and that is a shortcoming.

In billiards, the absence of data about a phenomenon is not evidence of the absence of that phenomenon. It is evidence of the absence of a measuring instrument.

This has practical consequences. If a governing body reads a report saying "no anomalies detected" and concludes the sport is clean, it is reading a report that could not measure what needed measuring. Football went through this lesson with match-fixing scandals, when betting-monitoring systems proved insufficiently sensitive to abnormal betting patterns in smaller competitions. Billiards, with a more fragmented tournament system and lower data transparency, faces higher risk, not lower.

I am not saying there is match-fixing in billiards. I am saying we do not have the tools to know, and the lack of tools should not be presented as a guarantee.

A LESSON FROM THE WINTER WITHOUT CROWDS

There is a period in my career I return to whenever I think about billiards.

In 2026, when football was suspended by the pandemic, I did not write articles. I went into the data archive and analysed around one thousand matches from 2026 to 2026 to find spatial patterns in earlier crowdless matches. I found that home teams lost on average about twelve percent of their home advantage. And I predicted Liverpool could drop around seven points from losing the Anfield advantage. That prediction came true in reality.

The lesson I drew was not "data predicts correctly". The lesson was: when an important variable disappears — crowd noise, stadium pressure — models based on old data become systematically biased, and that bias is measurable.

Billiards is in the opposite situation. The crowd variable does not disappear; it remains. But the data variable about the crowd never existed. Nobody records the sound of the Crucible arena in a deciding frame, and nobody measures its effect on the next shot.

When the pitch fell silent, I could hear clearly the crack of a football system refusing to change. In billiards, the arena has never been silent, but data about it has been silent from the start.

FROM ONE MAN TO AN INDUSTRY

I ask myself whether I should write this piece. I am fifty-nine, and I no longer need to prove I am right.

But I spent seven months counting 312 phases by hand in Liverpool, and I spent many nights redrawing the map of the Qatar final. I know the value of collecting data with your own hands. And I know the frustration of realising your work cannot be scaled, because nobody else will sit down and count with you.

What I propose is not to build a vast data platform for billiards. What I propose is to change the order of priorities: before building the platform, decide what to measure. And what should be measured is not century breaks, but safety quality and break quality — two things that decide frames but never appear on screen.

I do not watch the match from the stands. I count the gaps between two positions. And in billiards, the biggest gap today is not on the table. It is inside our own notebook.

At this point, if you ask me what to believe, I would say: do not trust the statistics panel. Trust the statistics panel that does not yet exist. Trust the gap. And be prepared for the possibility that what you are looking for is not a correct number, but a correct question.

TAKEAWAY: WHAT TO TEST NEXT SEASON

My observation offers a concrete test for next season.

If my argument is right — that the player who wins the safety exchange, not the one who scores more, controls match tempo — then at upcoming events you will see a repeating pattern: in long matches, the player with the higher safety-exchange win rate will win the next frame with higher probability, regardless of how many centuries he has made.

You can verify this yourself without any data platform. Just a notebook, a pen, and enough patience to sit and count.

And if you find me wrong — if century breaks remain a better predictor than safety quality — then say so. I have lived long enough to know that a refuted hypothesis is a hypothesis that has done its job. The only unacceptable thing is the silence of data presented as a guarantee.

Do not ask me what my philosophy is. Ask me how many safety exchanges that player won in the final fifteen minutes.