One Name, Six Rulebooks: Where Billiards Analytics Goes Wrong at the First Data Row

One Name, Six Rulebooks: Where Billiards Analytics Goes Wrong at the First Data Row

On 31 July 2026, the Crucible Theatre in Sheffield opened its doors to a few hundred spectators, the first pilot scheme to bring crowds back to an indoor sporting event in England after months of restrictions. On the next match day, the UK government cancelled the pilot. The stands stayed empty for the rest of the tournament, and Ronnie O'Sullivan closed out the World Championship with an 18-8 win over Kyren Wilson.

On the final night I sat with my notebook. Across those three weeks I had logged the length of safety exchanges, the number of occasions when both players chose to push the cue ball towards a cushion rather than attack, and one thing that cannot be measured: the silence between shots. The silences grew longer. Nobody shouted after a difficult cut. Each player built his own tempo, and that tempo was markedly slower than at any tournament I had watched from the stands before.

The crowdless season erased a variable that no model can encode: noise.

Then I noticed a much larger problem. Every conclusion in that notebook applied to snooker alone. Paste that passage onto a 9-ball match and it becomes nonsense, because in 9-ball the break is the decisive phase rather than an opening ritual, and nobody pushes the cue ball towards a cushion in the sense I had just described. One word, billiards, two worlds of rules. That was when I began to doubt not a particular piece of analysis, but the way this industry names its own data.

One name, six rule systems

In Vietnamese, everything is called bi-a. In English, everything is called cue sports and then split into snooker, pool and carom. Both habits work fine in a corridor conversation, and both are useless the moment you try to measure something.

Beneath that shared name sit at least six competitive systems with almost nothing technical in common: snooker; American 9-ball; American 8-ball; Chinese 8-ball, also branded Heyball; three-cushion carom; and Russian pyramid. Each has its own governing body. Snooker is run by the WPBSA and the World Snooker Tour. Pool sits under the World Pool-Billiard Association. Carom is governed by the Union Mondiale de Billard. Russian pyramid has its own federation. Chinese 8-ball runs on an operator-led model, where the body staging the event also manufactures tables, opens pool-hall chains and sells franchises.

The history of the name explains part of the mess. Snooker was devised by British Army officers in India in the 1870s, initially as a blend of variants that already existed. The word pool comes from the pooled-money betting that American rooms used to run their games. From the outset, the umbrella term described a commercial space, not a sport.

The technical differences do not convert. Snooker is played on a 12-foot table with 15 reds and six colours; the core skill is building breaks and controlling the table through safety exchanges that can run for many visits. 9-ball uses nine balls, and the break unlocks the entire rack: the run-out rate off the break is the single most important metric, and it has no snooker equivalent. Three-cushion carom uses three balls and never pots anything; points come from sending the cue ball off at least three cushions before contact, so the whole technical system revolves around the path of the cue ball rather than the precision of a pot. Chinese 8-ball uses snooker-style cut pockets with pool-sized balls, plus called ball and called pocket, a variable foreign to both 9-ball and snooker. Russian pyramid uses larger balls and pockets only marginally wider than the ball, turning every shot into a gamble on weight.

At this point the useful question is which system measures what, not which system is harder.

Every rulebook is a hypothesis about how people decide under pressure.

Snooker assumes pressure accumulates over time, so it tests through length. Pool assumes pressure arrives through brevity, so it tests through variance. Carom assumes pressure lives in the imagination of a path, so it tests through geometry. Three hypotheses, three scoring philosophies, and no way to collapse them into one scale without wrecking all three.

Which metric belongs to whom

There is a simple test for whether a billiards dataset deserves trust: read the column headers.

Centuries and 147 maximums exist only in snooker. The fastest competitive 147 belongs to Ronnie O'Sullivan at 5 minutes 8 seconds, made in the first round of the 2026 World Championship against Mick Price. No other system has an equivalent, simply because no other system's table structure permits a scoring run of that length.

Run-out rate off the break exists only in pool. In Chinese 8-ball, operators measure called-ball conversion, and that number means something different: it reflects the quality of the decision taken before the shot, not just the execution of it.

Carom measures average per innings. That figure moves between roughly 1.5 and 2.0 depending on the event and the format, and it cannot be placed beside any snooker or pool number for comparison.

A mislabelled column poisons every model downstream of it. No algorithm repairs a classification error sitting in the first row.

This is where an old principle of mine earns its keep: error is where reality signs its name. But to read the signature you must first know whose paper you are holding. An error in snooker and an error in 9-ball do not share a grammar, so they cannot be pooled into a single technical-error column and fed to a model.

Skill transfer between systems is uneven too. Cue-ball control, stance, tempo and pre-shot routine travel with a player who switches disciplines. The break does not. A snooker player moving to 9-ball keeps the feel, but the break is usually the clearest weakness, because snooker contains no shot that requires scattering nine balls and holding the cue ball mid-table. In the other direction, a pool player entering snooker struggles with multi-visit safety play and with a patience that a race to nine never builds.

The translation error sits at the level of vocabulary

There is a classification failure few people notice, and it applies with particular force to Vietnamese readers.

In Vietnamese, bi-a defaults to carom and to the domestic pool-hall circuit. When a Vietnamese outlet writes about world bi-a, the general reader pictures a carom table. In English-language newsrooms, billiards on the sports desk defaults to snooker, especially when the copy originates in Britain. One word, two semantic fields.

The consequence is that every time a headline crosses from English into Vietnamese, the discipline can change without anyone intending it. A line about a 147 at the Crucible can sit beside a line about an Asian carom event inside the same bi-a roundup, and the reader takes them as two events in one sport.

For anyone working with data, this is a graver fault than a missed pot. It is not wrong at the level of results; it is wrong at the level of units. And a unit error cannot be caught by rechecking the arithmetic.

Long format and short format are separate disciplines

In snooker, the biggest matches are played across multiple sessions, with finals stretching over two days and up to 35 frames. That tests stamina, sleep quality, the ability to forget a frame just lost, and the ability to hold a line for hours. In 9-ball, most events run as races to nine or eleven, occasionally to seven.

At that length, variance dominates the outcome. A fortunate break in a deciding rack can define a player's week. That does not diminish the event; it simply means this system's ranking measures something different from the other system's ranking.

Which makes the question of who is the number one in billiards structurally malformed. A valid version of it carries three conditions: in which system, at which format, over how many matches. Drop one condition and the answer loses its value.

The case of Ding Junhui is instructive. In April 2026, aged 18, he won the China Open by beating Stephen Hendry 9-5 in the final, and the result was read as the signal of a new Chinese snooker generation. Use that same name to infer China's standing in other billiards branches, however, and the reasoning leaves the rails at the level of units.

Money does not flow down the same pipe

In Britain, snooker is a television sport. The World Championship has been tied to the Crucible Theatre since 2026, and the main revenue lines are broadcast rights, sponsorship and tickets. That structure explains why the British calendar is relatively stable and why the flagship events cling to traditional venues.

In China, 8-ball runs on the inverse model. The operator manufactures tables, opens pool-hall chains and then anchors the whole ecosystem on the tournament. Cash comes from table sales and hall franchises first, from the event second. Prize funds can therefore climb very fast without a single large broadcast contract.

Carom in Europe and parts of Asia survives on club structures and invitational events, with prize pools several orders of magnitude smaller. Russian pyramid is bound tightly to its domestic market and regional events.

An income conclusion follows, but it holds inside one system only. In snooker, the top tier can earn millions of pounds a year while most players outside the top 100 cannot cover travel and hotels from prize money. Transplant that claim onto carom and it fails, because the carom problem is not the distribution of income within a professional labour market but the very existence of a professional labour market.

An income metric cannot be applied to a system that lacks the same labour market.

The same holds for the equipment supply chain. A snooker table brand and a pool table brand do not compete along the same demand curve, do not work to the same cloth specification, and do not share a dealer network. Folding them into one billiards-equipment category to size a market is a bold and meaningless addition.

The blind spot: a label is a commercial decision

Sports-data companies handle billiards in broadly the same way everywhere: everything under one tag.

The reason is commercial more than technical. A data vendor wants a billiards package to pitch to bookmakers, broadcasters and sponsors. The broader the tag, the easier the sale. The broader the tag, the thinner the analytical value. Those two aims conflict, and the commercial aim always wins because it pays first.

The danger lies in belief, not in technical laziness. When an analyst believes a cue sport is a cue sport, he stops checking the provenance of each column. An analytical process does not collapse because the data is dirty; it collapses because of absolute faith in the label.

The heaviest consequence sits in the monitoring layer. Integrity programmes, anti-match-fixing work and betting-market surveillance all run on baselines, and baselines are built from historical data. If historical data blends six rule systems into one column, the baseline skews. An anomaly in snooker measured on a pool scale can look ordinary when it is suspicious, and suspicious when it is entirely ordinary. Both directions do damage, and both are silent.

I am not describing a specific case; I hold no case to describe. I am describing a monitoring gap, and the correct response to a gap: record it as not yet assessed rather than as clean.

The fix for the rest is not complicated. Every data record needs a minimum of five fields: discipline, ruleset version, match format, table specification and governing body. Metrics may be compared only when all five match. That is the minimum discipline the industry has declined to adopt, simply because it slows the sales pipeline.

What to verify next season

If the operator-led model keeps expanding prize money in Asia, the first signal will not appear in the records. It will appear in the calendar.

Specifically: mid-ranking players will withdraw from low-purse ranking events in the fortnight before a major Asian date. Not because their form has gone, but because they are optimising for something the ranking does not measure.

That is a falsifiable prediction, and it can only be tested by a reader who knows which rulebook he is standing in. It is the kind of work a Tactical Wizard does before drawing any chart: establish the playing field, then take your turn.