Vietnamese Table Tennis and the Data Gap After Every Match

On the scoreboard of an internal training match at the National Sports Training Centre, the score stopped at 3-2 after five games. The stands were almost empty, with only a few coaches and a group of young athletes staying behind to watch. I sat in the third row, notebook open, recording every point by rally length rather than just the score. After 118 points, a pattern emerged too clearly to ignore: the winner took 68% of points in rallies lasting more than seven touches, but only 41% in rallies that ended inside four touches. That number appears on no official statistics sheet of the tournament. It exists only in the notebook of the person who stayed behind after the lights went out. When the stands are empty, I see the truest version of a player. Vietnamese table tennis runs inside a narrow but stable tournament system. The annual national championship brings together the strongest delegations such as Ha Noi, Ho Chi Minh City, Hai Phong, Da Nang, the Army, Binh Duong, Dong Nai and Can Tho. Above that system sits the SEA Games cycle, where everything is measured in medals and continental qualification places. The familiar men's names include Nguyen Anh Tu, Dinh Quang Linh and Tran Tuan Kiet; on the women's side there are Nguyen Khoa Dieu Khanh and Mai Hoang My Trang. Regionally, Vietnam must constantly trade blows with Singapore, Thailand, Malaysia and Indonesia — countries with deeper youth development systems and more regular international calendars. The problem is that our data stops far too early. Hundreds of matches take place across the country's arenas every year, yet almost all detailed data evaporates the moment a match ends. Organisers publish the game-by-game score, and that is it. Nobody records points won by server, by rally length, or by finishing position. After years of working with statistical tables, I have realised that most arguments about Vietnamese table tennis — who deserves a SEA Games spot, who is rising, who is finished — are built on memory rather than data. Memory is selective. A ranking table is a summary; raw data is the testimony. Since 2026 I have tried a simple measurement framework on internal matches and national championship matches. The principle is borrowed from football: instead of counting goals, count the quality of chances. In table tennis the unit is expected points, abbreviated xP. Each service situation is assigned a win probability based on four variables: spin type, placement, speed into the table and the state of the score at the previous point. Add those probabilities across a match and you get a single figure — the points a player should have won if they played to the true quality of their strokes. The early results are fairly consistent. Among Vietnam's leading male players, the point-win rate inside the first three touches usually sits at 58-62%, roughly level with the regional leaders. But once a rally passes seven touches, that rate falls to around 45%. In other words, the strength of Vietnamese table tennis lies in finishing early, while the weakness lies in enduring long rallies. That gap shows most clearly when xP is placed beside actual points. A player can win a match with a lower xP than the opponent, simply because a few net-cord balls fell on the favourable side. Conversely, some defeats come with a higher xP — the sign of a player who did the right things but lacked a little composure at the decisive points. Looking only at the score, these two cases are identical. Looking at xP, they sit on opposite sides of the problem, and the remedy differs too: one calls for tactical adjustment, the other for mental training. What is striking is that the model does not demand very complex data. Four variables, one recorder and one session are enough to generate a signal. I once tried adding a fifth variable — the opponent's court position — then dropped it because it muddied the model without improving accuracy. Vietnam's table tennis problem has never been a lack of tools, but a lack of people willing to sit down after the match ends. A confession is needed here. xP is not the truth, and I would be the first to object if anyone turned it into an absolute measure. My sample is small. Internal matches take place under conditions far removed from a final with spectators, international umpires and medal pressure. The correlation between high xP and good results is an observation, not a causal relationship. Some players train superbly but collapse when the electronic board lights up, and vice versa — something no model captures unless context variables are recorded. Nor do I believe that transplanting football's statistical thinking wholesale is enough. Table tennis moves faster, has fewer points, and each point is governed by a chain of decisions made in hundredths of a second. A model that looks elegant on paper does not necessarily reflect the true rhythm of a match. If the number detaches from what happens on the table, it becomes mere decoration. The signal to watch over the coming months is not medals, but whether anyone starts recording systematically at national championship level. A sufficiently detailed notebook can change selection, coaching and player evaluation. Numbers do not lie, but the people who read them do.