Three-cushion: the data gap Vietnamese billiards has never measured

Three-cushion: the data gap Vietnamese billiards has never measured

The final of the three-cushion world championship in Ankara in September 2026 put two Vietnamese players in the last match of the sport's biggest event. Bao Phuong Vinh and Tran Quyet Chien. The hall looked up at the scoreboard and waited for exactly one thing: the number 50. Whoever reached 50 points first lifted the trophy.

I was sitting in a coffee shop on Lach Tray Street in Hai Phong, rewatching the recording at one in the morning with a squared notebook open at its three hundredth page. I was not waiting for the 50. I was waiting for the innings count.

Because 50 points divided by innings gives you the average. If the match ends on the twenty-eighth inning, the champion's average is 1.78. If it drags to the thirty-sixth, the same trophy and the same name on the honours board come with an average of just 1.38. One result, two completely different stories about form. In domestic tournaments, no scoreboard in Vietnam puts that number on screen.

I keep the notebook for a reason that has followed me through four years in this trade: most conclusions about Vietnamese billiards are written with the eye, not on paper. Spectators remember the beautiful shot. Nobody remembers the twenty-third inning.

A sport played with the hand, judged by the eye

Carom three-cushion in Vietnam runs on a clear paradox. The base is huge: thousands of tables in clubs from Hai Phong and Hanoi to Da Nang and Ho Chi Minh City, open tournaments almost every month, and a generation of players who have climbed into the world's top group. The results are real: the World Cup has been staged for several consecutive years at Nguyen Du Stadium, and in September 2026 a Vietnamese player won the world title for the first time. The data infrastructure, meanwhile, is close to zero.

At international events, the Kozoom system records every inning: points, innings, longest run, and a figure called TSR, the tournament shot success rate. Open the live scoring page and you have numbers you can compare across events and seasons. At the national championship and domestic open events, the only thing usually preserved is the final score and the winner's name. No innings. No runs. No table conditions.

That gap has concrete consequences. Without inning data there is no baseline, which means nobody can answer a simple question: is this player rising or fading over the past six months. A player who wins four events in a row at an average of 1.05 and a player who loses in the semi-final at 1.32 can live identically in the crowd's memory, because memory only stores results. Without a baseline, every argument about form ends in feeling.

There is a technical reason why billiards data is harder to collect than football data. Football runs on a 90-minute clock; every event attaches to a fixed timestamp. Billiards has no shared clock. A three-cushion match can last twenty minutes or two and a half hours, and its real unit of time is the inning, a unit the players themselves create. To get data, somebody has to sit and record inning by inning, or enough tournaments have to run electronic scoring. Both require investment, and investment in data is the kind nobody can see with the naked eye.

Before the numbers, four definitions. Average is points divided by innings. An inning is one visit to the table, ending in a score or in the handover. A run is the number of consecutive points within one inning. TSR is successful shots as a share of total shots, usually recorded only where an international system is in place. Full notes are at the end.

Average is the laziest number in the game

Two players averaging 1.20 can be playing two different sports. The first has innings of 8, 6 and 5, separated by blanks. The second grinds out two or three points every visit, never exploding, never collapsing. In a 40-point match, the first finishes faster, meaning his opponent gets fewer innings to recover, but he is also the man most likely to stall at 34. The second travels the long road steadily but owns no weapon for opening a gap.

Average does not tell those two apart. That is why I added a column beside it: the number of innings with five points or more, and the longest run of each match. Across roughly forty domestic matches I recorded by hand, the group with at least three runs of five or more won the majority of their matches, even when their average was not higher than the opponent's. The group with the same average but a longest run of three won noticeably less.

Forty matches is a small sample. I am not using it to prove anything; I am using it to ask a question. If the spread of runs matters as much as the average, why does nobody print it on the scoreboard? One extra column on an electronic board costs less than an advertising banner. But it requires organisers to believe the run figure means something, and that belief only arrives when somebody shows that it does.

There is a simple check I still run whenever I rewatch a recording. I split the match into three stretches by innings: the first ten, the middle ten, the rest. For most young players, the average in the opening stretch sits below their match average, sometimes by as much as 0.3. For players who have come through international events, the gap is smaller and has no fixed direction. With only the final score, we would call both men the same kind of player.

The table never makes it into the record

This is the most neglected part, and the one that destroyed my faith in comparing averages across tournaments.

A three-cushion table is not a constant. New cloth and cloth three months old roll completely differently. Cushion rubber loses rebound over time, and it does not lose it evenly across the four cushions. Ball sets from different makers carry different weight and bounce. Then there is humidity. Hai Phong in July often sits at 80 to 90 percent humidity, the cloth absorbs moisture, the balls run short, and the stroke needs a different force from a dry, air-conditioned hall in Ankara.

I once logged twelve practice sessions by the same player, six on his familiar practice table and six on a tournament table. The average on the practice table was higher, by roughly 0.2 to 0.3 points per inning. That margin is enough to move a man from the seeded eight into the qualifying rounds.

This leads to an uncomfortable conclusion: any average comparison between tournaments that does not record table conditions is meaningless, like measuring height with a ruler that stretches with temperature. I am not saying abandon the average. I am saying that every time you use it, you should record at least four things: cloth type and cloth age, cushion condition, hall humidity, and the ball set. Those four lines need no expensive equipment. They need somebody willing to sit down and fill them in.

One missed shot is an error. Three missed shots from the same position on the cushion, across three different innings, is a signal. But the signal only appears if somebody records the position. No tournament in Vietnam does this, not even roughly.

Inning number twenty-five

There is a moment in three-cushion when every average becomes useless: the deciding inning, when both players reach match point at the same time. The rules call it the tie-break. First to score wins.

Across roughly twenty-five tie-breaks I have watched live or on recordings, players under 25 tend to stretch their preparation before the decisive shot to 20 seconds or more, against 12 to 14 seconds in ordinary innings of the same match. Their conversion rate in that specific inning runs below their own match average.

I call this measurement the rhythm index: the average time between shots within an inning, read off video timestamps. The index does not judge anyone. It shows when a player's rhythm has been broken. A man cruising at a 12-second rhythm who suddenly jumps to 22 seconds in the twenty-fifth inning of a semi-final is telling a story the scoreboard cannot carry.

This is where I deliberately put a qualitative element in the middle of the numbers. For years I measured everything measurable, and at some point I forgot that a player is a person, not a data series. The feel of the cushion, the breath before the shot, the noise of the hall, the experience of sitting and watching an opponent run eight points: all of it shapes the result, and none of it enters the record. The rhythm index is how I force the spreadsheet to hold at least one trace of the human being.

Format compresses form

Another variable rarely discussed: the format. A 40-point match and a 50-point match are different tests by nature. The shorter the match, the more the steady player's edge is compressed, and a single seven-point run from the weaker man can swing the whole thing. The longer the match, the more time the gap in class has to surface.

In my notes, the higher-average group wins most 50-point matches, but in 40-point matches that rate falls closer to a coin flip. This is no revelation to anyone who works in the sport, but it is rarely said out loud, because saying it invites the question of why organisers choose 40.

Then there is the number of matches per day. Group stages often force a player into two, sometimes three matches in a single day, each lasting hours on his feet around the table. On the days I tracked a full session, most players' averages in the third match sat below their first. Nobody records it, nobody publishes it, and nobody asks why a seed went out in his third match of the day.

On the World Cup circuit the calendar spans continents in a season, long flights, shifted time zones, disrupted eating and sleeping. I have never seen a dataset with a column for hours flown before an event. For a sport where the feel in the hand is the entire profession, that is a variable left completely blank.

When the data is right and the conclusion is still wrong

Now the part I consider most important, and least discussed.

TSR is the most quoted figure in elite billiards. But a player who chooses only safe shots will post a high TSR, because a safety is easier to execute than a scoring shot. A high TSR can be a sign of caution rather than class. Inverted, late in a match the leader often gets comfortable innings on easy layouts, so his TSR climbs. Read only the final figure and you conclude the winner had a higher TSR because he played better, when in part it is higher because he was winning.

That is correlation read as causation, and it happens constantly in every sport with data. In billiards the error is more dangerous because the recording is thin. A single column with no cross-check easily becomes truth simply because it is the only thing available.

Data never lies, but I have misheard it. In 2026, at an open event in Hai Phong, I used group-stage averages to declare that a certain player would reach the final. He lost in the quarter-final to an opponent averaging nearly 0.2 lower. Reviewing the record, I saw that opponent had three runs of six or more in his previous two matches. The column I needed was right there in my notebook. I simply had not looked at it.

There is one more problem few want to hear: the data analyst is walking into the practice room with conclusions detached from real rhythm. Telling a player his average drops 0.3 in the third match of the day is a correct and useless sentence if nobody answers why. The coach standing by the table knows it might be that he skipped lunch, or that the hall's air conditioning runs too cold. A column describes; it does not explain. The analyst's value lies in bringing the right question, not a full spreadsheet.

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

Signals for the next cycle

Three things I will track next season. One: whether the national championship adds an innings column beside the points column, even in the early rounds. Two: whether anyone fills in the four-line table-conditions sheet. Three: how many players under 25 reach the last sixteen, an indicator of squad depth.

None of it needs a large budget. It needs one person who believes measurement is part of the sport rather than a side ritual. I still keep the notebook, and if next season the organisers print an innings column beside the points column, I will be the first to copy the data down, and to joke that I doubted for rather a long time. The crowd laughed. The numbers did not.

Metric notes

Average: total points divided by total innings in a match or tournament. It is the standard three-cushion measure and the basis of ranking in most international systems.

Inning: one visit to the table, ending when the player scores or hands over.

Run: consecutive points within a single inning. In this article, a run of five or more is treated as an explosive inning.

TSR: successful shots as a percentage of total shots, usually recorded only where an international data system operates. It does not distinguish a safety from a scoring attempt unless shot type is logged, which is rare.

Tie-break inning: when both players reach match point simultaneously, a separate inning decides the winner; first to score wins.

Rhythm index: the average time between shots within an inning, read from video timestamps. This is a measure built by the author for reference, not an official figure of any governing body.

Hand-recorded data in this article comes from domestic events since 2026, on a small sample, and is used only to illustrate how to frame the question, not to assert statistical conclusions. Sports analysis, not betting advice.

Three thousand innings taught me that one inning can teach more than all of them.