The Fifth Game Does Not Appear on the Ranking Table

The Fifth Game Does Not Appear on the Ranking Table

In my point-tracking table for the 2026–2026 WTT season, there is one column I rarely open for anyone to see: the win rate in the fifth game. For the group of eight men who reached the semifinals of WTT Champions events, that rate is 61 percent. For the same group, the first-game win rate is 74 percent. That thirteen-point gap appears nowhere in the world rankings, never shows up on the scoreboard, and is almost never mentioned in match reports. Yet it is where a player's essence is exposed most clearly — after technique has run dry, after the opponent has finished reading the serve, when all that remains is a string of decisions inside roughly seventy seconds.

I work with table tennis data using exactly the method I once used for football: break the match into measurable units. Football has xG and PPDA. Table tennis has three base metrics — average rally length, service-point win rate, and win rate while trailing. All three can be pulled from the point-by-point data the WTT publishes after every match. No high-speed cameras needed, no sensors inside the blade. Just someone sitting down and decoding every single point.

This season I rebuilt 42 WTT Champions semifinals and finals, plus the World Championship knockout rounds, then split the data game by game. In total, 1,184 games and more than 11,000 points. From that I have a picture that does not match the way people usually talk about this sport.

Rallies are getting shorter, and nobody has noticed.

The average rally length among the eight strongest players I track has fallen from 6.4 touches per point in the 2026–2026 season to 5.1 in 2026–2026. Rallies longer than nine touches now account for just 12 percent of all points. But here is where the data turns counterintuitive: those long rallies decide 31 percent of the points in the fifth game. In other words, elite table tennis is being played shorter across the first four games, then gets stretched back out precisely in the last one — when both players know each other so well that neither wants to open a point with a short serve.

I do not believe in form; I believe in form data. The two rarely match. Rally length is the clearest example: a player can look like he has stalled in the eyes of the crowd, while the data shows he is simply hitting three touches shorter per point than he was three months ago.

Service-point win rate is the strongest predictor, and it is not being used in the right place.

Across the 1,184 games I decoded, service-point win rate in the first four games correlates with the fifth-game result more strongly than any other metric — stronger than overall win rate, stronger than ranking, stronger even than head-to-head record. Players who hold a service win rate above 70 percent across the first four games win the fifth game with a probability of 68 percent. Those below 60 percent win with a probability of 39 percent.

This matters for a very specific reason. The serve is the only thing in table tennis a player fully controls. When the body fades and the hand shakes, the serve is the last motion to lose stability — but it is also the easiest motion to read from the outside. A player entering the fifth game with a service win rate 10 points below his own baseline usually does not lose because the opponent is better. He loses because the serve no longer lands where it should, and the opponent has been standing there waiting since the third game.

A physical hole never shows up in the rankings — it only surfaces in the fifth game.

I once wrote a report for a football club and was pushed back on for recommending a substitution based on GPS data. In table tennis, the equivalent is not GPS but recovery time between points. From video data, I measure the gap between the ball going dead and the player settling into the next service stance. In the first game, the average rest interval among the eight strongest players is 11.2 seconds. In the fifth game, it rises to 14.8 seconds — a 32 percent increase. A player who needs an extra three and a half seconds to breathe between points is a player slowly losing control of the match tempo.

And this is where I have to be blunt: the world ranking cannot measure the fifth game. The WTT points system rewards going deep in many events, and going deep in many events usually means winning cleanly in the early rounds. A player can climb into the top 10 by winning 3–0 against weaker opponents while almost never reaching a 9–9 situation in a deciding game. The ranking records results. It does not record how many times you have been in that situation.

Three player archetypes show up fairly clearly in my logs. The first uses the serve as its main weapon, a model Wang Chuqin and Ma Long have represented for years — this group holds the most stable service win rate going into the fifth game, because the serve is the skill least dependent on the legs. The second lives on fast rallies inside the first three touches, Tomokazu Harimoto being a clear example — this group has a very high first-game win rate but the largest fifth-game drop, averaging 17 percentage points. The third plays with footwork and spin, Truls Moregard or Felix Lebrun — slower starts but the smallest drop in the deciding game, because they are used to extending points.

A counterintuitive angle: long rallies do not make you lose — they are only a symptom.

Many commentaries I read conclude something like: this player lost because he was dragged into long rallies. I think that reads the causality backwards. Long rallies are not the cause. They are the result of both players no longer having enough power to finish a point inside the first three touches. When the legs stop pushing, the counter-loop from behind the table loses speed, the ball arcs higher, and it lands in a zone the opponent can easily return. The causal chain runs: loss of leg drive, loss of spin speed, long rally, lost point. Cut off the last link and you will understand nothing.

This is also why I do not use long-rally counts as a performance metric. It is a temperature gauge, not a quality gauge.

This season's data says one more thing: the gap between the leading group and the rest is narrowing in the fifth game but widening across the first three touches. In games one and two, the eight strongest players win 71 percent of the points that end within the first three touches. In the fifth game, that figure drops to 58 percent. Which means that when a match gets tense, the technical advantage of the leading group is compressed, and the match moves closer to a 50–50 proposition. For someone who works with data, this is worth more than any transfer headline: it says the biggest reward in elite table tennis is not in playing better, but in holding your technical structure together when you are tired.

A note on the data I use. The WTT point-by-point feed is published openly, but that same feed also runs straight into betting pipelines within seconds of every point. One source feeds both the analyst and the bookmaker, and the bookmaker always receives first. That is the side effect few people mention when they talk about the digitisation of sport.

My checklist for the next round.

Not every match is read in the same order, but for this stretch my priorities run as follows. Service-point win rate across the two middle games — if it falls more than 8 percentage points below the opening game, that is an earlier signal than any scoreboard. Rest intervals between points from the fourth game onward — a rise above 25 percent is a physical signal, not a tactical one. Win rate at 9–9 and 10–10 — the smallest sample and the most misunderstood, but it measures exactly what needs measuring. And rally length in the fifth game versus the first — if it rises, the match is drifting toward whoever has the better physical base, regardless of ranking.

What to watch in the next round.

I will not name a champion. A season is a long chain, but people usually remember only the last three matches — and the last three matches rarely tell the truth about the whole season. What I am watching is whether the leading group holds its service win rate above 70 percent when it steps into a deciding game, or whether that rate keeps sliding toward 60 percent. If it slides, we will see more matches where a higher-ranked player loses from 2–1 up, and the media will call it a shock. To me, it is simply an equation written back in the third game, waiting until the fifth to reveal itself.

Every player has a weak joint. My job is to find it before the opponent sees it — and in modern table tennis, that weak joint usually sits in the fourteenth second between two points, not in any stroke at all.