Ultimate Tic-Tac-Toe

Nine tic-tac-toe boards in one big game – and every move decides where your opponent must play next.

Rules
  1. Win a small board with three in a row.
  2. Where you play inside a small board decides which board your opponent must play next. If that board is finished, they may play anywhere.
  3. Win three small boards in a row to win the game. If nobody does, whoever won more small boards wins.

The highlighted boards show where you may play. Your score against the computer stays in this browser.

Tic-tac-toe, but with real strategy

Ordinary tic-tac-toe ends in a draw once you know the trick. Ultimate tic-tac-toe fixes that: the big board is made of nine small boards. Win a small board with three in a row – and win the game by taking three small boards in a row.

The twist: the square you choose inside a small board sends your opponent to the matching board on the big grid. Play the top-right square, and they must play in the top-right board. So every move is attack and defence at once. Play against the computer on three levels (the hard one thinks ahead thousands of moves) or with a friend on the same device.

How to use

  1. Play anywhere on your first move.
  2. Your square decides which small board your opponent plays in next (it lights up).
  3. If that board is already won or full, they may play on any open board.
  4. Win three small boards in a row to win the game.

Ideas for using it

  • Classroom. a quick strategy game for two at break time.
  • Train your brain. the hard computer level is a real challenge.
  • Family. easier than chess, more interesting than normal tic-tac-toe.
  • Learn the trick. avoid sending your opponent to a board where they can win.

Frequently asked questions

Who invented ultimate tic-tac-toe?

It is a modern folk game played by maths fans and students; its exact origin is unknown.

What happens if nobody gets three boards in a row?

Here, the player with more small boards wins; if that is equal, it’s a draw.

How strong is the computer?

Easy only looks one move ahead; normal and hard simulate thousands of random games per move (Monte Carlo search).

More games for two?