Nine boards, one winner - and every move controls your opponent
Ultimate Tic-Tac-Toe takes the game everyone solved in grade school and turns it into a genuine strategy game. The board is a 3x3 grid of 3x3 tic-tac-toe boards - nine small boards arranged inside one big board. Win a small board and you claim its cell on the big board; get three claimed cells in a row on the big board and you win the game.
Ordinary tic-tac-toe is a guaranteed draw between competent players. Ultimate Tic-Tac-Toe is nothing like that: the nested structure and the rule linking your move to your opponent's options create real depth, long-term planning, and painful blunders. It's easy to learn in one game and hard to play well for a long time after.
That second rule is the heart of the game: you are never just making a move, you are also handing your opponent their next battlefield. You can play the AI solo or challenge a friend in the multiplayer lobby.
The number one skill is checking the destination before you move. Before marking a cell, look at the board you're about to send your opponent to. If they have two in a row there, you just gifted them that board. Many otherwise fine local moves are terrible because of where they send the opponent - a mediocre cell that ships them to a harmless board beats a strong cell that hands them a win.
The center cell of the big board participates in four winning lines (row, column, and both diagonals), so the center small board is the most valuable territory in the game. By the same logic, the center cell inside each small board is the strongest local square - but note that playing it sends your opponent to the coveted center board. Weighing that trade is the classic Ultimate TTT dilemma.
You do not need to win every small board - you need three in a line. Deliberately conceding a board that doesn't fit your winning lines is often correct, especially if fighting for it would keep sending your opponent to boards where they're strong. Think of small boards as squares to be traded, not battles to be won at any cost.
Once a board is won or full, any cell that maps to it becomes a "free move" trigger. Late in the game, deliberately playing into cells that correspond to dead boards gives you a free move - or forces you to think about the free moves you're giving away. Controlling who gets free moves, and when, often decides close games.
The PlayVersusAI engine weighs both the local boards and the macro board, and on higher difficulties searches ahead through the send-chains:
| Difficulty | What to expect |
|---|---|
| Easy | Plays reasonable local moves but overlooks where it sends you - exploit it. |
| Medium | Blocks local wins and avoids the most obvious bad sends. |
| Hard | Values the macro board and steers you toward boards where you're weak. |
| AI Training | Deep search over move sequences - it will trade small boards for the global win. |
Free to play - no account needed