University of Tokyo Rock Paper Scissors Robot
It wins every round, but not because it solved the psychology of the game. It solved the timing problem by being faster than human perception.
The Direct Answer
The University of Tokyo robot wins Rock Paper Scissors by reading the human hand during the throw and responding fast enough to finish with the winning gesture before the reveal completes. That makes it a remarkable robotics demo and, under normal WRPSA rules, an illegal reactive opponent.
How It Actually Works
The key is not prediction. The system uses high-speed vision to detect the hand as it forms, classifies the incoming gesture in real time, and drives a robotic hand quickly enough to answer with the winner. Humans cannot perceive and counter that fast. The robot can.
Why Randomizing Does Not Save You
Against a predictive model, better randomness can help. Against this robot, it does not. The machine is not guessing your move ahead of time. It is reading the actual move you are making and finishing second while appearing simultaneous. That is why its win rate can stay perfect even if your own throw selection is truly random.
Why It Is Not Fair Under Official Play
Official Rock Paper Scissors requires simultaneous reveal. A throw that reacts to the opponent after the commitment window has effectively begun is a late throw. The Tokyo robot is therefore not evidence that a legal machine opponent solved the sport. It is evidence that very fast perception and actuation can break fairness if the reveal standard is not enforced.
Why The Demo Still Matters
The project is useful precisely because it separates two different questions. One is engineering speed. The other is strategic play. The robot dominates the first question. Human competition still lives in the second. That is why the demo belongs on a science page, not on a rules page declaring the game solved.
Where To Go Next
If you want the broader research context around AI, psychology, and non-transitive systems, continue to The Science of Rock Paper Scissors. If you want the best-known human-behavior paper instead of the robotics demo, go next to The Rock Paper Scissors Study from Zhejiang University.
The Useful Short Version
If someone asks what the University of Tokyo robot proved, the clean answer is this: a machine with enough visual speed and motor speed can react during the throw and win every round, but that is a timing exploit, not fair competitive Rock Paper Scissors.