Billiards' Data Gap: What to Read When the Stat Sheet Is Blank

BILLIARDS' DATA GAP: WHAT TO READ WHEN THE STAT SHEET IS BLANK

On 5 May 2026, at the Crucible Theatre in Sheffield, Zhao Xintong beat Mark Williams 18-12 in the final of the World Snooker Championship. He became the first Chinese player to win the world title, and he did it as an amateur, only seven months after a 20-month suspension for betting-related and match-fixing offences had ended.

That night I stayed late in my London office, opened the dataset I maintain for the billiards desk, and searched three of the simplest columns available: centuries per session, safety success rate, and average points per visit. The result set came back empty. The dataset was not broken. Those three metrics simply do not exist in a public, cross-checkable, reusable form.

A sport with almost half a century of televised coverage, with world champions from more than twenty countries, with a five-hundred-thousand-pound cheque for the Crucible winner, still has no shot-level dataset fit for a data journalist to work with. That is the first variable I want to isolate, before I say anything about anybody.

CONTEXT: A WELL-RUN SYSTEM, POORLY RECORDED

The World Snooker Championship has been staged at the Crucible since 2026, and the hosting agreement has been extended through 2027. The World Snooker Tour operates the professional circuit, the World Professional Billiards and Snooker Association acts as the governing and disciplinary body, and Matchroom Sport handles commercial promotion. The ranking list runs on a two-season cycle, with roughly 128 players on the main tour and two-year tour cards reallocated through Q School.

That structure works smoothly at the commercial layer. The problem sits at the recording layer. Every match is broadcast with dozens of cameras, every shot is captured from three angles, every session runs for hours — but what is retained is the frame score and a handful of aggregate percentages shown on screen between frames.

The current cycle of this sport is what I call transfer season. Once the World Championship ends, the tour-card list turns over, Q School opens, and players drop from the main tour into the amateur system or move toward Chinese and Middle Eastern events with larger prize money. Exhibition contracts, personal sponsorship deals, academy agreements in China — all of it is negotiated in this window.

And precisely when the money moves fastest, information is thinnest. A player changes coach, a player signs a three-week exhibition deal in Macau, a player announces he will skip ranking events to play invitationals — all of those items matter, and almost none of them arrive with enough data to verify their real value.

WHAT I CHECKED, AND WHERE IT IS EMPTY

Based on my experience following matches at the Crucible and on the ranking circuit across the 2026/25 season, I split snooker data into three layers and tested each one.

The first layer is results data. This layer is complete. Frame scores, match scores, century counts, 147 counts, rankings, cumulative prize money — all public and cross-checkable against at least two independent sources. Ronnie O'Sullivan owns seven world titles and a record 15 competitive maximum breaks; he was the first player to reach 1,000 career centuries, in 2026. John Higgins has four world titles and more than ten maximums. Mark Williams has three world titles, the 2026 one arriving a full 15 years after his 2026 title. Judd Trump won the world title in 2026 and held the world number one ranking for most of 2026-2026. Kyren Wilson won in 2026.

The second layer is match-level performance data. It exists, but it is coarse. Broadcast stat panels provide pot success rate, long-pot success rate, safety success rate and average shot time. These are useful metrics, but they are outcome metrics, not quality metrics.

The third layer is spatial and difficulty data. This layer is entirely empty. There are no ball coordinates on the table, no cueing angles, no distance from cue ball to object ball, no per-shot difficulty rating. Snooker measures whether the ball went in, but not how hard the shot was — and that gap turns every player comparison into a comparison between non-equivalent contexts.

This is why I say snooker has no xG equivalent. In football, xG normalises a shot for location, angle, pressure and body part to separate chance quality from outcome. Snooker has no equivalent mechanism. A player potting at 91% in one session may be facing far harder shots than a player potting at 94% in another — and no dataset tells us which.

Pot success rate, therefore, is a metric polluted by opponent, by table conditions, by room temperature, and most of all by shot selection. A player who plays more safety will post a higher pot success rate simply because he only comes to the table when the position is good.

CRUCIBLE 2026: WHEN THE NOISE WAS SWITCHED OFF

In 2026 the season was disrupted by the pandemic. The World Championship at the Crucible began with a pilot day admitting spectators, then moved quickly to behind-closed-doors play. I rewatched every session of that tournament to test one narrow hypothesis: whether removing crowd noise measurably changed playing tempo.

The result did not come from what increased, but from what disappeared. When the crowd left the arena, the noise variable was isolated, and what remained was a cleaner sample of players' decision-making tempo. Average shot time became more stable between sessions of the same match, with less variance than at heavily attended events in the same period.

An empty arena, the coach's voice clearer than ever, and the data too. But I have to state the limit plainly: this is an observation from a single tournament, in a single year, under conditions dominated by a global health event. It does not prove that crowds slow players down. It only shows that when you switch off one variable, the remaining variables surface more clearly — and that is an experimental condition, not a conclusion.

What I took from the summer of 2026 was not a number but a method: find windows where the playing environment changes under control, and read the sport through them.

ZHAO XINTONG AND THE SAMPLE-SIZE PROBLEM OF NINE MATCHES

Zhao Xintong's run at the 2026 Crucible consisted of four qualifying matches and five matches at the main venue. I rebuilt the data for that entire path, and the first thing I wrote on the board was the sample size: nine matches, four of them played at a different venue, under different table conditions, with no crowd.

Nine matches are not enough to assert anything about a player. They are only enough to describe a sequence. At that sample size, any conclusion of the form "Zhao has recovered his peak form" or "Zhao has completely changed his game" sits outside what the data permits. The standard deviation in pot success rate at professional level is larger than the gap between a player in form and a player out of form, when the match count is this low.

A player's journey is not an upward arrow, it is a scatter plot. Plot Zhao's path on a time axis and you see a very low point during the suspension, a cluster of recovery points on the amateur circuit, then a very high point at the Crucible. A regression line through those points would have a positive slope, but its standard error is large enough that it cannot forecast the 2026/26 season.

What I am willing to assert, after cross-checking against qualifying data, is much narrower: Zhao returned with a different shot-selection structure than before. His rate of switching to safety in difficult positions went up, and unnecessary risky shots went down. That is a change in process, not in outcome. Results are noise; process is signal.

SNOOKER'S TRANSFER SEASON: WHERE THE NOISE IS ORGANISED

Snooker has no transfer window in the football sense. There are no transfer fees, no release clauses, no registration windows. But it has a real labour market, and that market operates exactly like every other sports market: on information asymmetry.

The transfer market is fundamentally a regression model, but people keep calling it a race. The dependent variable is a player's contract value. The independent variables include ranking, age, century count, recognition in the Chinese market and — most importantly, and least measurable — ticket-selling power. The problem is that almost none of those variables are published.

Three money sources are shaping this market. The first is the World Snooker Tour prize fund, driven by broadcast revenue in the UK and Europe. The second is invitational and exhibition events in China, where personal appearance fees can far exceed the prize money of a ranking event. The third is Saudi Arabia, where the Riyadh Season World Masters of Snooker launched in 2026 with a 20-point golden ball and a one-million-dollar bonus for a 167 — a prize structure that had never existed in the sport's history.

The golden ball is a more notable institutional experiment than it appears. It creates a new statistical target nobody has reached, and turns an exhibition event into a generator of data. When one tournament invents a metric no other tournament has, it is pricing attention rather than pricing ability.

Against that backdrop, fans get very few filters. A story that player X is negotiating with a Chinese academy can come from three sources: an agent trying to create negotiating leverage, a journalist copying it from social media, or an official announcement published late. Those three sources carry entirely different reliability and are circulated in exactly the same format.

On the contract itself there are only two words: signed. Everything before that is noise.

WHAT I READ FROM THE AGENTS

In ten years following this sport I have reached a conclusion I know is unpopular: the largest hidden cost in professional snooker sits at the agent layer, not the tournament layer.

I am not making a moral argument. I am making a structural one. An agent has a clear incentive to blur the signal: tell the media his client is being courted by three events, negotiate several exhibition deals in parallel, stay silent about a wrist injury, and foreground ranking as if it were a measure of long-run ability when ranking is a lagging indicator.

The measurable consequence is that a player's market value is driven more by media frequency than by safety-conversion rate in difficult positions. And because snooker has no spatial data layer to normalise against, nobody can rebut it with numbers.

This is where the sport differs fundamentally from most team sports. In football, a club can hire an analytics firm to value a defender before buying. In snooker, there is no equivalent tool. Decisions to invest in a player — a sponsorship, an invitation, an academy place — are made largely on instinct and relationships.

At the same time, at the development layer, professionalisation is flattening individual style. Academies in the UK and China coach to standardised curricula, fitness and psychology are digitised, and average shot time is tracked as a KPI. The idiosyncratic styles that once defined this sport — the defensive grinder, the reckless pure attacker — are being sanded down into a safe median band. A young player today plays like another young player in another country, and the data to prove it does not exist.

CONTRARIAN ANGLE: CORRELATION IS NOT CAUSATION

After Zhao Xintong won, a very tidy story appeared: the sport's disciplinary system worked, the player was rehabilitated, and the reward went to the right person. That story sounds reasonable, has supporting numbers, and I do not believe it in that form.

There are at least three other explanations of the same sequence, and the available data cannot rule out any of them.

The first: a 20-month suspension creates a forced break from competition, and for a young player who had competed continuously for years, a short break can restore both physical condition and motivation. Some research on competitive cycles in individual sports supports this effect, though it is small.

The second: removal from the ranking system pushed the player back into qualifying, where media pressure is far lower and he faced weaker opponents in early rounds. That is a scheduling advantage, not an ability advantage.

The third: luck. With nine matches, the probability that an average-level player wins the title is not small enough to dismiss.

The 2026-23 match-fixing case the governing body handled reveals another paradox that public debate tends to skip. Ten Chinese players were sanctioned, two of them for life. Three years later, the world champion is a Chinese player who had just served a suspension. Both facts are true, and both belong to the same picture: the depth of Chinese billiards is real, and the sport's governance structure at that time was loose. Any report that picks only one of the two is presenting noise as if it were signal.

There is one more point I must make about my own method. The silence of the data is not a clean signal. When I say billiards' stat sheet is empty, that does not mean the sport fails to generate events. It means those events are poorly recorded. That is an experimental condition, not a conclusion. Confusing the two is the mistake I made in 2026 with an analysis of the World Championship, and it cost me months of re-reading data.

WHAT TO WATCH IN THE NEXT CYCLE

Three signals worth tracking through the 2026/26 season.

First, the Crucible contract story. The agreement to stage the championship in Sheffield runs to 2027, and pressure to expand capacity or relocate is one of the largest structural variables the sport has not resolved. Arena capacity feeds directly into ticket revenue, and ticket revenue feeds directly into the prize fund.

Second, data quality. If a broadcaster or data vendor begins collecting ball coordinates at shot level, this analytics field will change within two seasons. Player valuation models built on instinct will be replaced by models built on normalised difficulty. Every argument about who is better will then have to move to a different arena.

Third, money from the Middle East. The Riyadh Season World Masters prize structure sets a precedent: reward for a moment, not for a track record. If that model spreads, ranking — already a two-year lagging indicator — will progressively lose its role in player valuation.

The question I leave behind is not who will win the 2026 Crucible, but whether this sport has the courage to record itself before an outside data vendor does it for them.

DATA LIMITATIONS

This article draws on public results data from the professional circuit, governing-body disciplinary notices, published hosting-contract information, and the author's direct observation of televised sessions from the 2026/19 season through 2026/25.

The limits must be stated clearly. The sample size for the Zhao Xintong analysis is nine matches, far too small for causal inference; any statement about this player's long-term trajectory sits outside the data. The observation about the 2026 behind-closed-doors season rests on a single tournament and cannot be separated from its global health context. Prize-fund and bonus figures come from organiser announcements and have not been independently cross-checked against financial statements.

What I do not have enough data to assert: no conclusion in this article identifies the best player at any point in time, and no forecast is made about tournament outcomes. I am willing to assert exactly one thing: billiards' spatial data layer is empty, and that emptiness is shaping how this sport is understood, valued and sold.