In tic-tac-toe, does the player who goes first have a secret edge?
After you watchIn tic-tac-toe, does the player who goes first have a secret edge?
The short answer
If both players play perfectly, going first is no edge at all — tic-tac-toe is always a tie. The first-move edge is real only against an imperfect player: because X moves first, X gets the first chance to punish a mistake, so going first wins far more often against a real person.
Try this next
- What if O slips on almost every move? Drag O's care all the way to 'very sloppy' in the one-game lab and play a few — does X start winning nearly every game?
- Does the edge disappear if O is perfect? Set the toggle to 'Both perfect' and run the race a few times — watch the draws bar swallow everything, with no X or O wins at all.
The whole story
How it works
A perfect player always blocks every three-in-a-row threat and grabs every winning square. When both X and O play perfectly, every threat is blocked and the game is a forced draw — neither side can force a win. But real people slip and leave a winning square open. Since X moves first, X usually gets the earliest chance to spot and punish that slip, which is why first-move wins pile up against humans even though they vanish against a flawless opponent.
What people get wrong
Lots of kids assume going first is an automatic win, and others assume tic-tac-toe is purely random. Both are wrong. Going first cannot force a win against perfect play — the best X can guarantee is a tie. The edge people feel is not magic in the first move itself; it is that the first move gives the earliest chance to take advantage of the other player's mistakes.
The catch
Going first feels powerful, but it buys you nothing against someone who never slips: it is a guaranteed draw, not a win. And the second player isn't doomed either — a perfect O can always force the tie too. The whole 'edge' only exists in the gap between perfect play and human play.
Questions kids ask
Who wins tic-tac-toe if both players are perfect?
Nobody — it's always a tie. With perfect play on both sides, every winning threat gets blocked, so the game is a forced draw. Going first cannot break that.
So is going first useless?
Not against real people. Humans slip and leave winning squares open, and because the first player moves earlier, they usually get the first chance to punish a mistake. That's why going first wins far more often against a human.
Can the first player force a win?
No. The best the first player can guarantee against perfect play is a draw. Tic-tac-toe is 'solved' — neither side can force a win if both play correctly.
How does a computer never lose at tic-tac-toe?
It looks ahead at every possible continuation and always picks a move that can't lose. The board is small enough to check completely, so a perfect player only ever wins or ties, never loses.
Talk about it
- Ask them: if a perfect player can never lose, why doesn't going first guarantee a win — what's the best X can actually force?
- Ask: where does the 'going first is better' feeling really come from, if both perfect players always tie?
For grown-ups
Tic-tac-toe is a solved game: with optimal play by both sides the game-theoretic value is a draw, so the first-move advantage is exactly zero under perfect play. The felt asymmetry comes from imperfect opponents — of the 255,168 distinct complete games, far more end in an X win than an O win, and because X moves first it gets the earlier opportunity to convert a blunder. The 'perfect' player here uses minimax, which searches the entire game tree and always chooses a move that cannot lose; it's the same idea behind computer chess and Go, just on a board small enough to solve completely.
Keep going
What else makes you wonder?
- If a bigger board (like 4-in-a-row Connect Four) is also solved, who wins there — first player, second, or nobody?
- Could you ever design a 'fair' two-player game where going first truly gives no advantage even against imperfect players?
- If a computer can never lose at tic-tac-toe, why is chess still too big for any computer to solve completely?