Billiards and the Data Gap: A Sport That Cannot Identify Itself

In the 2026/25 tracking file I maintained for seven months, there is a column named discipline. Its job is to answer the simplest question in the trade: which branch of billiards does this record belong to. By week thirty, that column returned an empty value on most of the rows I entered from secondary sources. The reason does not lie with the person entering data. It lies with the sources themselves, which cannot distinguish snooker from American 9-ball, from Chinese 8-ball, from three-cushion carom, or from Russian pyramid.

On the night of 5 May 2026, at the Crucible Theatre in Sheffield, Zhao Xintong beat Mark Williams 18-12 to become the first Asian world snooker champion, barely a year after completing a twenty-month suspension. It was the most recorded final in the history of Asian snooker. To reconstruct it as a set of verifiable metrics, I had to build three bridges: CueTracker for break-building and century data, the official World Snooker Tour site for the calendar and prize money, and the WPBSA ranking list for position. Three bridges, three data formats, one match.

My trade began in football, where every event carries a shared identifier. A Premier League match is logged by Opta, StatsBomb and Wyscout; you can open two sources and compare shot by shot. Billiards has no such luck. WST runs the ranking system for snooker. Matchroom Sport runs the 9-ball ecosystem around the Mosconi Cup. Chinese domestic promoters run Chinese 8-ball, where the Joy Masters is among the richest events in that branch. The UMB governs three-cushion carom. Four governing systems, four rulebooks, four data pipelines.

For a data journalist, the first consequence is purely technical: no cross-branch metric can be built, because there is no shared primary key to join the tables. An empty discipline column is not an input error. It is a miniature portrait of the entire infrastructure.

Cross-checking is a step I do not skip. In Zhao Xintong's case I ran three layers. The first was results and scorelines, compared between WST and CueTracker, matching perfectly across qualifying and all four Crucible matches. The second was the century-break sequence, where the two sources diverge on the counting threshold — one includes centuries made in qualifying, the other only the main stage. The third was prize money, where the 500,000 pounds paid to the champion is the only anchor nobody disputes.

It is the second layer that held me longest. A threshold discrepancy sounds trivial, but it breaks every historical comparison. Ronnie O'Sullivan passed 1,000 career centuries in 2026; if the counting standard is not consistent across eras, placing him beside a player from the 1990s is a structurally flawed comparison. The medal does not sit on the scoreboard, it sits in the xG table — and in billiards, that xG table has not even been defined consistently.

I tried a simpler measurement frame, borrowed from snooker itself: the frame-win rate immediately after losing a frame. That metric captures the ability to restore a state after a small shock. At the level of public data, it can only be computed for matches inside the WST system. In Chinese 8-ball, no equivalent frame unit exists, so the frame collapses at the variable-definition stage.

Thirty dead-ball rhythms, one release clause, and an entire market shifts. I wrote that line for football, but it holds harder in billiards. Here the unit of the market is not a transfer fee but a two-year tour card. Q School is the narrowest gate: hundreds of amateurs chase a handful of places, and one week of results decides two years of income. No body publishes income structure by tour card. Agents hold that information, and they release it only when negotiating.

The familiar explanation is that billiards is small, poor, and therefore data-light. I think causality runs the other way most of the time. A sport split into five rule branches cannot accumulate measurement infrastructure, because the cost of building it must always be divided by five. Sponsors do not pour money into an undefined object. And when no trustworthy data layer exists, oversight weakens precisely in the least illuminated segments.

The June 2026 match-fixing case is structural evidence, not personal evidence. Ten Chinese players were sanctioned, with Liang Wenbo and Li Hang receiving lifetime bans. The violations clustered in events with low prize money and low broadcast reach. In that segment, absent data means absent anomalies. Here I must be extremely careful: the correlation between data opacity and fixing risk is not enough to conclude causation. An alternative hypothesis stands — thin data may simply follow from thin budgets, and both may stem from market size.

The signal I am tracking in the next cycle is not a trophy. It is whether the parties accept a shared player identifier and a shared discipline identifier. The transfer market is at heart a regression model, but everyone keeps calling it a race. Billiards is at a stage before variables are even set. While the discipline column in my file still returns empty, every conclusion about a player, an event or a season is only a politely presented guess.

Data limits: this piece draws on public sources including CueTracker, the official World Snooker Tour site, the WPBSA ranking list and the WPBSA sanction notice of June 2026. The sample size for cross-branch comparison is zero, because no common dataset exists. Any claim about the link between data opacity and integrity risk needs retesting at match level.