Billiards and the Data Gap: Twelve Statistical Columns, Eleven of Them Meaningless

A twelve-column spreadsheet and a match that refused to fit

In March 2026, on Lach Tray Street in Hai Phong, I sat in the third row of a billiards club with a notebook and a spreadsheet open on screen. Two three-cushion players were walking into the nineteenth inning. I recorded twelve columns: scoring rate on the opening inning, number of times I left my opponent at the table from a dead position, average cushion contacts per inning, average time per inning, success rate from the long-rail position, and seven more columns I have since forgotten, simply because they never predicted anything.

The player who won that night was below average in nine of the twelve columns. He won with three long innings and one push that forced his opponent to break. By the twenty-eighth inning I closed the notebook, paid for my coffee, and walked home with a feeling I had met exactly two years earlier on a football pitch: I was measuring the wrong thing.

That was the second time in my career I had to rewrite the definition of my own job. The first was in 2026, when my xG model collapsed in front of seven saves by a goalkeeper. The second was at a billiards table, where I realised the sport displays more numbers than football but holds almost no data in the sense an analyst needs.

This piece is not an argument that billiards cannot be analysed. It is an argument that we are analysing it with a toolkit designed for another sport, and that the data gap is not a technical accident.

Three disciplines, three measurement systems, one missing standard

Billiards is an umbrella word. When a Vietnamese spectator says they are going to watch billiards, they may mean three entirely different disciplines, with different rules, tables, scoring and ways of measuring achievement.

The first is carom, most commonly three-cushion and one-cushion. This is the discipline with the oldest official data system, governed by the Union Mondiale de Billard (UMB). The base unit is moyenne générale — average points per inning. A player at 1.500 averages 1.5 points every time they come to the table. Alongside it sits best run, the longest scoring sequence in a match or tournament, and the average number of innings needed to close a game.

The second is snooker. This is the discipline with the best data infrastructure of the three, largely thanks to CueTracker and the World Snooker Tour statistics. Pot success, long-pot success, safety success, century breaks, 147 maximums and average shot time are all recorded.

The third is pool — 8-ball, 9-ball, 10-ball. It has the youngest audience and the fastest-growing prize money, and the weakest standard data system of the three. Most of what we know about a pool player comes from television and from the memory of viewers, not from a queryable database.

These three systems do not talk to each other. There is no shared unit for comparing a three-cushion player with a snooker player, or a 9-ball cueist with a 10-ball cueist. Football has xG as a reference unit for almost every league on the planet. Billiards has no equivalent.

The absence of a shared unit is the root problem, and it is not purely technical — it is a question of power.

Whoever controls the definition of a metric controls the narrative. The UMB defines success through moyenne. The World Snooker Tour defines it through ranking titles and centuries. Matchroom defines it through championships and break-and-run rates. These definitions cannot be converted into one another, and every organisation has a reason to keep it that way.

| Category | Carom (three-cushion) | Snooker | Pool (9/10-ball) | |---|---|---|---| | Main governing body | UMB | WPBSA / WST | WPA / Matchroom | | Base unit | Points per inning | Frame | Rack | | Most-cited metric | Moyenne générale | Century, pot success | Break-and-run | | Public data | Moderate | Good | Weak | | Historical queryability | Limited | Reasonable | Poor | | International standardisation | Partial | Partial | Almost none |

The table above is what I compiled after four years of cross-checking three different sources. It is not a quality ranking. It is a map of what exists and what does not.

Which metrics actually measure technique

When I started recording roughly twenty consecutive billiards matches by hand in 2026, the goal was simple: find out which metrics have predictive power. The result forced me to discard nearly half the columns I had built.

In three-cushion, two metrics held their predictive power in my sample: moyenne générale and success rate on the opening inning. The second is rarely discussed but matters. A player who scores on the first inning tends to control the tempo, because the opponent is pushed into chase mode very early. Across roughly three hundred innings I recorded, the group that scored on the opening inning had a higher win rate than the group that did not, but the confidence interval was wide enough that I would not call it causation.

In snooker, long-pot success is the metric I trust most, followed by safety success. Average shot time sits in the cautious category. It measures rhythm, not quality. A slow player may be calculating carefully, or may be hesitating. Same metric, two opposite causes.

In pool, I have to admit I have not built a stable metric set. Break success is the easiest to record and the most contaminated by table conditions, ball sets and venue climate. At some events the same player breaks at a high rate on day one and collapses on day three with everything else unchanged. That makes me suspect the variable that actually moved was not on the table.

A metric without its measurement conditions is a statement, not a data point.

This sounds obvious, but in practice it is not. When a broadcast displays a pot success rate of 94%, viewers assume it measures the same thing as 94% at another event. It does not. The difficulty threshold of the shots differs, the cloth speed differs, the pressure differs, and sometimes the counting method differs.

Four years ago I argued with a friend who coaches pool in Hai Phong about whether a young cueist was improving. I produced a dataset showing his win rate climbing from 41% to 48% over six months. My friend said one sentence: watch his last three matches before you say anything. I watched. He was right about something my spreadsheet could not see: the player had changed his stance, and that change cost him about two months of instability. The rising win rate was the result of shedding an old habit, not of learning a new skill.

Data never lies, but I have misheard it before.

Player data and the age curve

If I had to name the single biggest weakness in billiards data today, I would pick the age curve.

In football we have thousands of players at the same age to compare. In billiards, a top player competes in a few dozen matches a year. The sample is so small that each match becomes an observation with abnormal weight. When a 40-year-old wins a major, we immediately talk about form at forty. When a 22-year-old wins, we talk about a new generation. Both conclusions rest on a single observation.

In snooker, the group known as the Class of '92 is an example I reuse in talks. Three cueists born in the same year, 2026, stayed competitive for more than three decades. Ronnie O'Sullivan passed twelve hundred career centuries and holds more than fifteen official 147 maximums. Mark Williams and John Higgins also held top-tier places. The remarkable thing is not that they were good. It is that no biological metric in existing billiards data explains why they held form while most of their generation fell away.

We have outcome data. We do not have cause data.

In carom the problem is harder. Vietnam's leading players — Tran Quyet Chien, Nguyen Tran Thanh Luc, Chiem Hong Thai, Nguyen Duc Anh Chien, Bao Phuong Vinh — have all had periods at world level. But when I tried to build a form curve for each from public data, I hit two walls. First, data exists for majors, not for domestic events. Second, formats change so often that comparing periods becomes meaningless.

| Metric group | Snooker | Three-cushion | Pool | |---|---|---|---| | Ranking titles | Complete | Incomplete | Incomplete | | High-scoring shots | Century, 147 | Best run | Not standardised | | Head-to-head | Good | Moderate | Weak | | Form over time | Reasonable | Limited | Poor | | Fitness, injury data | Very little | Very little | Almost none |

I once wrote a short piece saying a young Vietnamese cueist was at the peak of his age curve. Three months later he lost four straight qualifying matches. My mistake was not a wrong prediction. My mistake was applying a metric built for a sport with high observation density to a sport with low observation density.

Tournament structure and the shape of the money

Tournaments are where data is generated, and tournament structure determines the quality of that data. In billiards, two structural features make analysis harder.

The first is short format. Most open events across all three disciplines use short matches — a few racks, a few frames, or a race to a low number. Short format means high variance. A weaker player can beat a stronger player over a short match with a probability that is far from negligible. That is good for audiences and good for television. It is bad for anyone trying to draw conclusions from a single match.

The second is a steep prize structure. At many events, the gap between champion and runner-up is larger than the gap between runner-up and quarter-finalist. That structure creates a clear incentive: top players concentrate on a handful of majors, skip smaller events, and the data at smaller events thins out.

When money concentrates at the top, data concentrates at the top — and most of the truth about a sport lives where the data does not.

In Vietnam, the domestic event system carries far more weight than it is credited with. National championships, open events and club-level competitions are where most cueists earn a living and accumulate experience. They are also where data is most abandoned: no detailed score sheets, no quality video, no post-event reports.

One concrete example I recorded: at a national-level open event I followed, the organisers did not publish players' moyennes after the group stage. Fans knew who advanced; they did not know who had played well. The story of the tournament was reduced to a list of winners, and the most interesting part — the process — vanished.

Qualifying is a separate problem. Qualifying systems at international events tend to be narrow, Asian quotas are limited, and wildcards are allocated on criteria that are not always public. This affects data directly: a player with potential who never gets into majors produces no data, and without data there is no invitation. The loop feeds itself.

The power map: rules in Britain, money in China, rhythm in Asia

Over four years of tracking, I have come to see the billiards power map as different from football's. It stratifies by function, not by strength.

Britain and the wider United Kingdom shape rules and tradition. This is where snooker was born, where the oldest ranking system lives, and where the traditional audience is stable. Britain's pipeline is still steady, but the rate of renewal has slowed.

China holds the money. The number of events staged in China grew fast over more than a decade, and the arrival of a Chinese world snooker champion changed how the whole industry views the market. China's youth pipeline is deep and starts early, but internal competition is so dense that many never clear national qualifying.

Asia outside China holds the rhythm. Vietnam, South Korea, Japan, the Philippines, Thailand — each has its own tradition. The Philippines is tied to pool and produced a generation that left a deep mark. South Korea and Vietnam are tied to three-cushion carom and are currently at their most productive internationally. Japan holds a stable position in both carom and pool.

| Region | Main role | Strength | Data weakness | |---|---|---|---| | United Kingdom | Rules, tradition | Durable event system | Slow renewal | | China | Money, market | Scale, investment | Closed internal data | | Vietnam | Three-cushion, training | International results | Poor domestic records | | South Korea | Three-cushion, organising | Strong event structure | Limited international publishing | | Philippines | Pool | Technical tradition | Data near empty | | Europe | Multi-discipline | Historical depth | Fragmented federations |

One point I want to make clearly, because it is often misread in Vietnamese social media debates. A Vietnamese cueist winning an international event does not automatically mean Vietnamese billiards has reached the corresponding level. It means that individual reached that level. Those are two different sentences. In the data I cross-checked, the number of Vietnamese players regularly clearing group stages at international events has grown far more slowly than the number of Vietnamese players who have had one good tournament.

A title is an event. A generation is a data series. We routinely confuse the first with the second.

Rules, governance and the grey zone of betting

This is the hardest section to write, and the one I want to write most directly.

Billiards is among the most structurally vulnerable sports to integrity risks around match outcomes. There are three structural reasons.

The first is individualism. There is no team to share responsibility with and no team-mate to accidentally expose an anomaly. A player can change an outcome alone.

The second is precision. At elite level, the distance between a good shot and a bad one is tiny. A positional shot that is off by a few millimetres looks identical to a positional shot deliberately off by a few millimetres. That makes visual detection of wrongdoing close to impossible.

The third is money. The amount wagered on billiards events, especially across Asia, is far larger than the level of monitoring currently in place.

| Check item | Status | Risk level | |---|---|---| | Transparency of match results | Moderate | Moderate | | Betting monitoring | Weak at many events | High | | Playing-rule disputes | Rare but serious | Moderate | | Eligibility, wildcards | Inconsistent | Moderate | | Player contracts and discipline | Low disclosure | Moderate |

I have no evidence of any specific case in Vietnam, and I will not write as though I do. What I have is a structural observation: when a sport lacks sufficiently dense public data, its ability to detect anomalies falls. A system cannot detect what is abnormal if it does not know what normal looks like.

Three goalkeepers fumbling in one round is a signal. One cueist suddenly missing a single inning is not, because we have no baseline to compare against.

That is why building public data for billiards is not an academic project. It is a protective measure for the sport.

The professional ecosystem and pressure away from the table

One common mistake in billiards analysis is assuming professional cueists have a career ecosystem similar to athletes in major team sports.

The reality differs. In most countries, a professional billiards player organises their entire career structure alone: schedule, travel costs, accommodation at events, coaching and sometimes income outside competition. There is no salary system, no off-season protocol, no injury insurance to team-sport standards.

Income is sharply polarised. A small group earns from prize money, equipment sponsorship and commercial activity. The rest survive on small events, teaching, or unrelated work.

| Dimension | Assessment | Note | |---|---|---| | Income structure | Highly polarised | Dependent on majors and sponsorship | | Coaching team | Almost none | Mostly self-coached | | Playing rhythm | Uneven | Concentrated in major seasons | | Psychological support | Almost none | No professional standard | | Post-career path | No defined route | Teaching or business |

The row I care about most is playing rhythm. In many sports, athletes follow a designed rest cycle. In billiards, the calendar forms spontaneously. A player might compete in three events across three weeks, then rest two months, then enter a major immediately on return. That pattern produces a type of variance no current metric captures.

When home stops being a fortress, I learned to listen to the empty stands.

I wrote that line about football in the summer of 2026, but it applies to billiards differently. In billiards, empty stands stand for playing conditions without spectators, without atmosphere, without a background hum. Many cueists play better in silence and worse with a crowd. We call this psychology, but it is a variable we could measure if we chose to.

Risk matrix

When assessing risk for a billiards player or competition system, I use six groups. The grouping is not meant to make analysis look sophisticated. It is meant to prevent risks of different natures being collapsed into one judgement.

| Risk group | Main content | Level | |---|---|---| | Competitive | Opponent dependence, short-format variance | Moderate to high | | Career and income | Polarisation, financial insecurity | High | | Integrity and reputation | Betting grey zone, weak monitoring | High | | Rules and organisation | Disputes, inconsistent wildcards | Moderate | | Psychological | Deciding-inning pressure, no support | High | | Systemic | Playing rhythm, data gaps | Moderate to high |

The group I watch most closely is psychological, because it is the only one I can observe directly with my eyes and still cannot prove with data.

I once watched a national final in which the leader stepped into the last inning. He had a position that was not technically difficult — a position my spreadsheets said should succeed above 80% of the time. He missed. He then lost. I spent two weeks looking for the metric that predicted that moment, and the answer was that none existed. My dataset knew he could make that shot. It did not know whether he believed he could.

Media, expectations and the trust gap

Immediately after a major title, the media narrative usually follows a familiar arc: arrival, peak, long-term expectation. The arc is appealing because it is tidy. It is also frequently wrong.

Cross-checking four years of Vietnamese billiards coverage against actual results, I found a pattern that repeated three times.

First, after an international title, coverage in the first two weeks tends to describe a historic turning point. From the third week, frequency drops sharply.

Second, expectation is placed on a single individual rather than on a cohort. This is a data problem, not an emotional one. When expectation concentrates on one person, the observation sample narrows and error rises.

Third, when results do not arrive, the story shifts from praise to a search for personal causes rather than structural ones. Nobody asks whether there are enough events. Nobody asks whether data was recorded. People ask why that player underperformed.

| Dimension | Media expectation | Data check | Gap | |---|---|---|---| | Individual results | Continuous peak | Cyclical ups and downs | Large | | Event appeal | Sustainable growth | Growth tied to majors | Moderate | | Discipline development | Rapid breakthrough | Slow, cumulative | Large |

The majority laughed. The data did not. A year later, I copied that article over again.

I do not write to convince anyone. I write so that the data has a witness.

The billiards industry transmission chain

Billiards does not exist independently of other sectors. It has a clear transmission chain, from local clubs to the equipment market and broadcast rights.

The chain starts with the hall and club ecosystem. This is where most players first encounter the sport. Club density, table rental prices, table and ball quality all directly affect the quality of players produced. In Vietnam this network is dense and widely distributed, but data about it barely exists.

The middle layer is players, events and media. This layer is affected directly by the first, but with a lag. A player who starts at fifteen needs close to a decade to reach international level.

The final layer is sponsorship, derivative products and the equipment market. This is the fastest-moving financial layer and the most dependent on the image of the middle layer.

| Segment | Direction | Magnitude | Time horizon | |---|---|---|---| | Clubs and halls | Foundation | Moderate | Long term | | Asian market | Rising | High | Medium term | | Equipment and gear | Star-driven | Moderate | Short term | | Broadcast and sponsorship | Volatile | High | Short term | | Talent development | Slow growth | Moderate | Long term | | Derivatives | Unstable | Low to moderate | Short term |

What I want to stress is the time mismatch. When a title happens in the middle layer, the final layer reacts within weeks. The first layer reacts years later. So when we assess the impact of a title on the whole industry, we usually measure the wrong layer.

The model knew in October. I only had the courage to believe it in May.

The counterintuitive angle: correlation is not causation

This is the part I use to correct myself.

When a cueist wins a major, three data groups rise together: titles, media recognition and search volume. The three curves move almost in lockstep. It is tempting to conclude that results create attention. But the causal order may be partly reversed: attention creates invitations to majors, and invitations to majors create results.

In billiards this causal path matters more than usual, because major-event slots per region are very limited. A player who is invited to more events has more chances to accumulate points than an equally strong player who is invited less. When we look at a ranking table, we are looking at the combined output of ability and opportunity, but there is only one column to write it in.

This leads to an uncomfortable conclusion. Most of the “player A is better than player B” comparisons we read online compare things that are not the same kind of thing. Comparing title counts between two players in two different event systems is methodologically meaningless, however enjoyable it is as an argument.

One goalkeeper fumbling is a mistake. Three goalkeepers fumbling is a signal.

In billiards I have never seen three top cueists collapse in one event in the same way. But I have seen the equivalent at system level: multiple young players struggling at the same stage, after the same format, under the same conditions. That is when an analyst should stop talking about individuals and start talking about the environment.

Three thousand matches taught me that one match can teach more than all of them.

Data still needed

I always close analysis with a list of what is missing, so my conclusions are not read as certainties.

First, inning-level data at domestic event level. Without it, any analysis of Vietnamese development is guesswork.

Second, a unified definition for basic metrics — at minimum opening-inning success rate and safety success rate.

Third, data on playing conditions: cloth type, humidity, room temperature. I suspect this is the most undervalued variable in the entire sport.

Fourth, structured psychological data — at minimum, outcomes in deciding innings. This is the hardest data to collect and may be the most valuable.

Fifth, larger samples. Every conclusion in this article should be read with a sample-size warning attached.

Signals for the next cycle

Three signals I will track over the next twelve months.

The first is the emergence of an independent billiards data platform not owned by any federation. If that happens, the sport's power structure will shift faster than organisers expect.

The second is the number of young Vietnamese players clearing international qualifying, not the number of titles. This is a leading indicator and will show up two to three years before final results.

The third is the number of domestic events publishing detailed data after conclusion. If that number rises, the entire analytical layer above it changes.

I do not know what happens next. But I know exactly what I will record.

Data never lies, but I have misheard it before — and I want my next mishearing to have a witness.