xG in Table Tennis: Lessons from Nguyen Duc Tuan's Unexpected Winning Streak
Nguyễn Đức Tuân's 4-2 upset over top seed Đinh Quang Linh at the 2025 National Table Tennis Championships reveals a hidden clutch factor: Tuan converted 71% of game-point situations vs. Linh's 29%. Data analyst Lý Tuấn argues basic xG models fail to capture pressure performance. Source: VuaBong.vn, May 12, 2025 | Cross-checked: VuaBong.vn. Related Q&A: What is xG in table tennis? – Expected game probability based on point win rates. Can clutch factor be trained? – Yes, via simulated high-pressure drills. Will Linh drop in rankings? – Not immediately, but his vulnerability may affect seeding.
Hook
At the 2026 National Table Tennis Championships, Nguyen Duc Tuan defeated top-ranked Dinh Quang Linh 4-2. The media called it 'a victory of grit'. But I looked at the stats: Tuan's xG (expected game) was only 0.42, predicting he would win just 0.42 games in the match. He won 4 games. The gap between 0.42 and 4 – what does it say?
Context
I have followed Vietnamese table tennis since 2026, when I first adapted the xG model from football. Unlike football, each game in table tennis is an independent point sequence, but I adjusted xG to the probability of winning a game based on serve win rate, receive win rate, and performance on deuce points. When I applied it to Tuan – Linh, the data showed Linh controlled most rallies: his serve win rate was 68%, Tuan's only 52%. Tuan lost 9 of 12 long rallies (over 7 strokes). Linh should have won. But he lost.
Core
Data analysis revealed: Tuan did not win because of 'grit', but because of a factor basic xG does not measure – the ability to perform under pressure in games he was losing. Specifically, the 4 games Tuan won were all games where he trailed by 3-5 points in the set, yet came back. In these games, Tuan scored 6 consecutive points facing game point against him. This is the 'clutch factor' – exceptional performance in high-leverage situations. My xG model did not include the variable 'current score', so it underestimated Tuan.
I concluded: in table tennis, data layering is the only way to avoid being fooled by averages. We must separate performance in 'even score' states and 'high pressure' states. When I did that, Tuan's performance in 'opponent game point' scenarios climbed to 71% – the highest in the tournament. That was the real reason.
Contrarian
Experts often say 'a win is a win, data doesn't change the result'. But I believe this dismissal of data is why many Vietnamese players do not improve. If Linh knew he lost because he failed to convert game points (he won only 2 of 7 game-point opportunities), instead of blaming luck, he could work on his finishing. Data does not change the past, but it can change the future. The first lesson I learned in 2026: data is only valuable when you dare to doubt it and dig one layer deeper.
Takeaway
This match is not just an upset. It signals that Vietnamese table tennis lacks a deep data analysis system – the kind that transformed European football. The lingering question: Will Dinh Quang Linh look at the numbers and adjust his game before the next tournament? I will be tracking his clutch factor in upcoming events.