Strategy
How to play ultimate tic-tac-toe well
This guide is for players who know the rules but still lose games they thought they had under control. Most super tic-tac-toe tips stop at the first move; this guide covers everything that comes after. Six principles, ordered by how many games they help you win, each with the reasoning behind it.
Never give away a free board
This is the game’s most costly mistake, and by far the most common. The cell you play does more than place a symbol: it names the small board where your opponent must respond. If that board is already won, drawn or full, the restriction disappears and they can choose any board still open — just like on the very first move of the game.
A free move is no small concession. Your opponent uses it to close the board where they already have two symbols in a line, to take the centre, or to redirect you somewhere you have built nothing. You have just played a move for them, with nothing in return.
The trap closes at the very moment you think you are doing well. You win the central board by playing its central cell: the cell says ‘central board’, the central board has just closed, and your opponent is free. Before each move, ask one simple question — can the board this cell points to still accept a move?
Why the centre is worth more
A three-by-three grid has eight lines. The central cell belongs to four of them, each corner to three, and each edge to only two. The centre is therefore worth twice as much as an edge: that is not a player’s hunch, it is a count.
The game applies the same count at both scales. The nine small boards themselves form a three-by-three grid, and the central board belongs to the same four lines — half the paths to victory pass through it. These values combine: the central cell of the central board is the most valuable cell in the game.
There is still a tension, and that is what makes this interesting. Playing a central cell on any of the nine boards sends your opponent to the central board. The most valuable cell is precisely the one that offers the most valuable board. Many games turn on this dilemma: take the local centre now, or wait until the central board is decided — knowing that once it is, that same cell no longer sends your opponent anywhere specific and hands them a free choice, which the previous section warns against.
Read the next move, not your own
The habit that changes everything is simple: before placing your symbol, look at the board this move designates for your opponent.
Ask yourself what their best move there would be. A move that wins a small board but sends your opponent to one where they already have two symbols in a line is still a bad move, even if it wins something.
A won board closes immediately: its empty cells can never be played again. A local win you did not need costs you both tempo and a destination.
Sacrifice a small board
Not all small boards have the same value, and their lines tell you how much they differ. In the big grid, an edge board belongs to two of the eight lines, a corner board to three, and the central board to four. Giving up an edge means giving up a quarter of the paths to victory; giving up the centre means half.
This leads to the manoeuvre that separates a decent player from a good one: deliberately giving up a low-value board to retain control of redirection. While your opponent spends moves closing an edge, you keep choosing where they play. The board is lost; the game is not. Better still, a drawn board counts for nobody and removes its lines for both sides.
The sacrifice has a price, paid later. A closed board accepts no more moves: the cell that pointed to it gives a free move to whoever gets sent there — and you will get sent there too. It is a permanent hole in the map of redirections, open to both sides. Weigh the board’s value against the free moves that hole will give out until the end.
Two-move traps
A forced response is constrained not only in what it does, but also in where it goes. Combinations are built on this double constraint, and two moves are enough.
The starting position is simple. You have two symbols in a line on the central board, with the top-right cell completing that line; on the top-right board, you have another two symbols in a line with the third cell empty. First move: play the central cell of the board you are on. Your opponent lands on the central board, sees the threat and must block it or let you take the board that carries four of the eight lines. But the only blocking cell is the top-right one — which redirects you to the top-right board, the one you prepared. Second move: complete your line there.
If they refuse to block, you take the centre. Both branches suit you: that is what separates a trap from a hope. The rule needs no diagram: build the threat so that the cell that blocks it points to a board where you are already ready.
Practise against bots
These principles take ten minutes to read and months to put into practice. Reading is not what makes them stick: losing a game because of one of them and seeing exactly where it turned is.
The site’s training ladder has twenty-five rungs, with Elo targets spaced a hundred points apart, from 100 to 2500. The most common mistake is staying at a level where you win comfortably: an opponent who does not punish a free board teaches you nothing. Move up a rung once you win more than half your games, even if you need to move back down later. As you climb the ladder, search simulations increase and randomness decreases: the lower rungs let free moves slip by, while higher up the bot explores more moves before choosing instead of picking one at random.
Play both sides. The first player freely chooses their opening board and sets the shape of the game; the second starts with a redirection they did not choose. These are two distinct problems, and practising one does not improve the other.