World Billiards and the Lesson of an Empty Data Column

On a Friday night in a club on the outskirts of Liverpool, I sat behind a nineteen-year-old player for four hours of a practice match. He won 9-7. What made me open my notebook was not the score. In the eleventh frame, across fourteen consecutive safety exchanges, he kept pushing the cue ball back into the same half of the table, even though his opponent had three times left a gap open in the middle area. He did not see it. Or he saw it and did not believe what he was seeing.

Back home I opened my own tracking sheet. The column for successful safety shots had numbers. The column for how many times an opponent left a gap sat empty, because I had never defined what "leaving a gap" means. The column for deliberation time before choosing a line sat empty, because I had never started a stopwatch for that concept. A sheet with four numeric columns and six empty ones still looks very much like a complete sheet. That is the most dangerous thing in this job.

A few days later I received an analysis file in which every input field had been left blank: no tournament name, no source, no time marker, no player named. The document was still long, still had a title, still had clearly divided sections, and every conclusion inside read "insufficient information". I read it twice. The second time I realised it was more honest than most reports I have ever received, because it refused to invent a player, a tournament or a record just to look complete.

Error is where reality signs its name. But an empty cell is not error. An empty cell is where reality has not yet written anything, and how we handle it decides whether we are analysts or storytellers.

Billiards is not one sport; it is six sports sharing a name

Before anything else, one point has to be settled, and most coverage in Vietnam skips it: the word "billiards" does not identify a discipline. Snooker, nine-ball pool, American eight-ball, Chinese eight-ball, three-cushion carom and Russian pyramid are six rule systems, six table geometries and six coaching cultures. Without the step of identifying the discipline, every analysis that follows is a house built on sand.

The difference is physics before it is technique. A competition snooker table is 12 feet long, the balls are 52.5 mm in diameter, and the pockets are cut tight. An American pool table is 9 feet, the balls are 57.15 mm, and the pockets are markedly more forgiving. A carom table has no pockets at all, measuring 2.84 m by 1.42 m in match play, with 61.5 mm balls, and every scoring shot must contact at least three cushions. The Chinese eight-ball table borrows the 9-foot pool footprint but tightens the pockets close to snooker levels, making it a hybrid with a far lower clean-run rate than American pool.

That physics produces six data languages that cannot be translated directly into each other. In snooker the unit of achievement is the frame and the points scored in a single visit, with the century mark at 100 points and the absolute ceiling of 147 built from fifteen reds, fifteen blacks and the final 27 points of colours. In nine-ball the unit is the run-out from the break, and analysts count consecutive rack wins rather than points. In three-cushion the unit is the moyenne, the average points per inning, a figure that sits below 2.0 for most of the world's leading players.

Placing those three units side by side in one ranking table is a logic error, not a presentation error. A snooker century requires roughly fifteen consecutive acts of cue-ball control on a tight surface. A nine-ball run-out requires nine acts on a wider surface with far more open pockets. A ten-point run in three-cushion requires ten calculations of three-cushion trajectories. Those three things do not measure the same ability, and calling all of them "strong form" only blurs what should be clarified.

I once saw a report describing a pool player as having "made a century" at a tournament. In reality his results were counted in rack wins, and the number 100 the reporter had seen was a cumulative points total for the whole event, an index that exists in no rulebook. That mistake is not simple carelessness. It comes from never having been taught that the first step in any billiards analysis is to establish which discipline is being discussed.

Fragmented governance and what it costs

At the governance layer the fragmentation is even clearer. Professional snooker is disciplined by the World Professional Billiards and Snooker Association, while World Snooker Tour runs the commercial circuit. Pool sits under the World Pool-Billiard Association. Carom sits under the Union Mondiale de Billard, which has existed since 2026. At amateur level there is the International Billiards and Snooker Federation, and in Asia the Asian Confederation of Billiards Sports.

Four bodies, four record books, four definitions of what a record is. This produces a very concrete consequence for readers: when a report says "world record", the next question must be which book. A record run of rack wins in pool is not ratified the way a carom moyenne record is ratified by the UMB, and no body sits above all four to adjudicate.

For an analyst based in England, this fragmentation is a permanent variable. It means every cross-discipline comparison needs a source note, and every "world champion" claim needs the name of the awarding body. In Vietnam, where three-cushion is the flagship discipline and has produced world champions, that lack of source transparency creates a paradox: fans know exactly who the champion is but rarely know which system the title belongs to or how heavy it really is against the rest of the world.

How accurate does a ranking table really look

This is the core of the problem, and it is where billiards data models are failing quietly.

In professional pool, the most widely used composite metric is the TPA, the total performance average, built by a long-standing statistics system in the American pool world. It aggregates shot-making, cue-ball position play, kicking, break quality, safety play and a component labelled "mental". That last component is where I stop.

A metric with a "mental" component means the person scoring decides for themselves what a brave shot looks like. There is no operational definition, no threshold, no cross-check. Across eight years of watching events in England I have compared TPA sheets for the same match scored by two different people, and the gap between them is often wide enough to reverse the standings of two players. Nobody is wrong. The ruler simply is not defined.

Snooker looks stricter, because most of its metrics are binary outcomes: the ball goes in or it does not. Pot success, safety success, centuries, 147s, average shot time. But even those apparently objective numbers depend on a human decision about what counts as a successful safety. A push that leaves your opponent a difficult shot is a success. A push that leaves your opponent a difficult shot which he then pots is what.

Three-cushion has the most theoretically elegant metric: the moyenne, the average points per inning. It spreads evenly across every inning, with no distinction between decisive and dead innings. That is precisely why it hides the most important thing. A player holding a 1.4 average who lifts to 2.2 in the decisive inning is not the same type of player as one holding 1.4 steadily and dropping to 0.8 when behind. An average cannot tell those two apart. In billiards, telling them apart is the entire story.

Prize money shapes behaviour before it shapes technique

Tournament structure is part of data analysis, not decoration. In snooker the tiers are clear: the Triple Crown group of the World Championship, the UK Championship and the Masters; then ranking events; then invitationals; then the seniors circuit. The World Championship is staged at the Crucible in Sheffield, with a final of up to 35 frames across two days and a winner's prize in the region of half a million pounds through the mid-2020s.

That structure produces highly predictable behaviour: most players ranked outside the top 32 spend the bulk of the season at low-purse, high-ranking-point events, because next season's entry to the big events depends on ranking rather than form. In other words, the schedule of the world number 40 reflects a points optimisation problem, not his form. Any model that forecasts results from recent form while ignoring that variable will be systematically wrong.

In Asia, money creates the opposite pressure. Chinese eight-ball events carry purses large enough to pull players across from other systems, and that forces some snooker players to weigh a ranking event in England against a flight to Asia for a bigger prize in a shorter format. This is reverse talent flow that official rankings never record, because rankings record results, not reasons for absence.

Generational transition: what cohort data says

The most discussed generation in snooker is the group born in 2026: Ronnie O'Sullivan, John Higgins and Mark Williams, all of whom turned professional in 2026 and were still winning titles in their late forties. Three players from one birth year dominating for three decades is a far more interesting statistical phenomenon than the way it is usually told, because it does not prove that generation was greater; it proves later generations were not deep enough to take the space.

When I count major titles by birth cohort across the decade from 2026 to 2026, the striking thing is not that the 2026 group kept winning, but the delay in handover. Players born after 2026 won major titles at a markedly lower rate than the speed at which media called them a golden generation. Media called them golden because they were young. Data called them unripe because they lost in the deep rounds.

The most striking case of the period is Zhao Xintong, who won the World Snooker Championship in 2026 as a qualifier, after having served a suspension in the WPBSA's major disciplinary case. That result broke almost every forecasting model built on ranking, age and form trajectory. It also delivered a concrete lesson for anyone building indices: the decisive variable was not in the match data at all, but in how the player spent the period during which he was not allowed to compete.

The most neglected dataset in billiards

In 2026 the WPBSA announced disciplinary sanctions against ten Chinese players over match-fixing and betting, including lifetime bans. It was one of the sport's largest governance cases in decades, and it has a feature analysts routinely overlook: the sanctions themselves are a dataset.

How a governing body handles such a case reveals its transparency, its decision speed, its cooperation with outside investigators, and how it protects whistleblowers. For an analyst this is structural data, not merely personal data. A system that can process ten cases within months, publish detail and offer a clear appeal route is a resilient system. A system that processes matters in silence is a system sending the opposite signal to those on the outside.

This is also the area where I am strictest with myself. Match-fixing risk may only be raised when a real signal exists, because raising it without evidence is not analysis, it is accusation. On a blank input file, naming this theme as a warning about any specific party would be a professional failure, however cautious the wording.

The blind spot is not missing data, it is the reflex to fill the gap

This is where I want to push against the industry's instinct.

The default assumption in most sports analysis is that where data has a hole, reason should fill it. From one result, infer form. From form, infer psychology. From psychology, infer next season's outlook. The chain sounds coherent, and it is the source of nearly every wrong forecast in this sport.

I call it the prodigy filter. When a young player wins a big match, the story is written immediately and never verified again. Three years later, if that player has won nothing else, nobody returns to correct the old story. They simply write a new one, headlined decline. Both stories rest on the same sample: one match.

A safety plan does not collapse because it is wrong, but because the player believes in it so completely that he stops watching the table. I have seen this repeat at every level. A player builds an effective safety system, wins with it, then applies it even when the opponent has read it and is waiting for exactly that line. Belief in the system does not fade when the system stops working. It grows, because at that moment the player needs the system to hold him together. That is the moment data becomes decoration.

And this is the variable no model can encode. The no-crowd season erased a variable no model can encode: noise. The 2026 World Snooker Championship at the Crucible was one of the UK government's pilot events, admitting a limited crowd on the opening day before moving behind closed doors when guidance changed. Ronnie O'Sullivan won it, beating Kyren Wilson 18-8.

What matters is not the winner. It is that in a silent arena, a missed safety no longer drew an immediate crowd reaction, and a fine shot no longer earned applause. Both directions of feedback vanished. Models trained on pre-2026 data were applied unchanged to the period after, even though they had been trained in an environment where a variable had disappeared. Nobody removed that variable from the model, because nobody had measured it in order to remove it.

Finally there is a subtler trap: using error as an excuse. When a forecast fails, the easiest reflex is to say billiards is a game of luck, cushions throw the ball, a fine contact changes everything. That is physically true. But it does not explain why the same model fails in the same direction every time it meets the same type of player. Error is where reality signs its name, while error that repeats systematically is where the model signs its name. Telling those two signatures apart is the line between an analyst and an advocate.

The value of a player is only a story the market repeats until it believes it. And when the match ends, the number lies more elegantly than the player ever could.

What to track for the rest of the season

If I had to choose three signals to watch, I would choose those verifiable by data rather than by feeling.

The first is the gap between fame index and output index, the difference between how often a player appears in coverage and how often he wins in the deep rounds. In three-cushion this is measurable as moyenne split by match phase. In snooker it is decider win rate. In pool it is run-out rate when trailing. In all three, a player whose fame index outruns his output index for two consecutive seasons is usually a player about to be repriced.

The second is progress towards a unified record book. If governing bodies begin cross-ratifying records between systems, the quality of every comparison table changes within a few years. If they do not, readers will keep receiving "world champion" titles that cannot be compared with one another.

The third is prize-money flow between regions. When one market pays more for a shorter format, player schedules change before rankings change. That is the earliest signal an analyst can catch, and it appears before the headline does.

In an annual season, most of the real story sits beneath the table. My job is not to retell the table. It is to point out which column is empty.