Lab · experiments
The Lab
Four games you cannot win. Each one shows a different way an AI stays ahead — search, game theory, learning, and a fully solved tree. Go on, try.
The Last Move
A connection game against an AI that has already solved every line you could play. It lets you believe you're winning — then shows you the exact move where you'd already lost.
Enter the masterpiece →Tic-Tac-Toe
Minimax search · unbeatable
Your move
Nim
Game theory · the nim-sum
Your move
Take any number of matchsticks — but only from one row. Whoever takes the very last one wins.
Rock · Paper · Scissors
Online learning · it reads your patterns
Make your move
The AI builds a model of your habits as you play. The longer you go, the better it predicts your next move.
Chomp
Fully solved tree · the AI never loses
AI is solving…
Eat a square and everything to its right and below goes too. Whoever is forced to eat the poisoned corner loses. The AI moves first — and it has already solved the game.