Game, Set, Code: How Yuexi Brings Computational Thinking to the Tennis Court
When Computer Science student Song Yuexi steps onto the NUS tennis courts, she’s not just thinking about her next serve. She’s thinking about strategy — patterns, probabilities, and precision — the same way she approaches a complex computing problem.
That mindset recently earned her international recognition.
Under the supervision of Associate Professor Liang Zhenkai from the NUS School of Computing, Yuexi developed a new language for tennis strategy analysis, mapping tennis games to chess games based on their structural similarities. Her paper, “TANS, A Chess-Inspired Notation System for System Analysis of Tennis Games”, received the Distinguished Paper Award at the 2nd International Sports Analytics Conference and Exhibition 2025 (ISACE’25) held in Shanghai.
“We wanted to create a notation system that could describe the flow of a tennis match as clearly as chess notation describes a game of chess,” explains Yuexi. “It’s about capturing the logic behind every rally.”
According to one reviewer at ISACE, “By defining an unambiguous grammar plus PGN-style metadata, TANS fills the long-recognised gap of a standard rally-level notation, a prerequisite for reproducible tennis data science and for training generative AI on raw match text.”
Bridging Code and the Court
An avid tennis player herself, Yuexi has long been fascinated by how logic and instinct intertwine in sport. “I’ve always loved both STEM and tennis,” she says. “But after coming to NUS Computing, I realised I could actually connect the two.”
That connection took shape — quite literally — on court.
During a friendly match with her supervisor, Liang, their casual discussion about strategy evolved into an idea for a formal research project.
“We started talking about how chess has a language to record every move and counter-move,” she recalls. “By the end of the session, we thought — why not build something similar for tennis?”
Turning Play into a Language
Over the next few months, Yuexi and Liang designed the Tennis Algebraic Notation System (TANS) — a notation that uses coordinates and compact strings to represent every rally.
“TANS defines an unambiguous grammar and metadata, allowing computers to process tennis strategies systematically,” Yuexi explains. “It’s semantically rich — with just a few tokens, you can capture how the point was played, the players’ positions, and tactical intent.”
To test their model, they manually charted three full doubles matches and even tried out the findings on court.
“It was surreal,” Yuexi laughs. “We won a point using one of the strategies we’d identified in our analysis — and we celebrated like it was a research breakthrough.”
That moment confirmed their intuition: the language worked. When Yuexi later compared their data with larger studies, the same patterns emerged. “That’s when I realised TANS could truly capture meaningful insights with much less data,” she says.
The Thinking Behind the Win
The project reflects SoC’s belief in the versatility of computational thinking — how the ability to reason systematically and distract problems can spark innovation in unexpected places.
For Yuexi, this was exactly what her education made possible. “At NUS Computing, professors really emphasise how to think, not just what to code,” she says. “After my first year, I found myself looking at everything differently — from how I plan my day to how I play tennis. Every rally feels like running an algorithm: analysing inputs, executing decisions, and adjusting when something fails.”
Sometimes, she even jokes about debugging her matches. “After playing, I replay the points in my head and look for logical errors,” she smiles. “It’s like reviewing code — only more fun.”
Learning Beyond the Classroom
This independent project also marked Yuexi’s first foray into academic research. “When we started, the idea felt overwhelming,” she admits. “A/Prof Liang helped me break it down into smaller, manageable parts and guided me through each step. That process taught me how to think systematically — and gave me the confidence to explore beyond my comfort zone.”
For her, the experience embodies the best of SoC’s learning culture — where mentorship and curiosity interact. “Professors here are open-minded,” she says. “They encourage you to explore ideas, even if they fall outside traditional areas. That freedom makes learning feel personal and creative.”
The Bigger Picture
Although Yuexi doesn’t plan to pursue sports analytics as a career, the project has changed how she views computing — not just as a technical skill, but as a lens for understanding the world.
“This project showed me that computing connects to everything,” she says. “It’s not just about building systems — it’s about using that mindset to see patterns, reason clearly, and create impact in any field.”
When asked about who inspires her most, she doesn’t hesitate: “Carlos Alcaraz,” she grins. “His game is so dynamic and strategic — a combination of all possible shots. It reminds me of good code: efficient, elegant, and creative.”
Looking ahead, Yuexi hopes to build on her work by integrating AI into TANS and exploring robotics — perhaps one day creating a robot that can play tennis and analyse strategies in real time.
For now, she’s content knowing that her project bridges the two worlds she loves.
“When I first joined NUS, I thought computing was just about programming,” she reflects. “Now, I see it as a way of thinking — one that helps me connect my passions and make sense of the world around me.”
