Take tic tac toe and invert the goal. Three in a row now loses. Everything else stays the same - same board, same alternating turns, same nine squares.

It sounds like a trivial modification and it produces a game that is genuinely harder to play well, breaks the main theoretical tool used to analyse the original, and inverts almost every piece of advice you have ever been given about the centre square.

The centre becomes a trap

In standard tic tac toe the centre is the strongest square on the board. It participates in four of the eight winning lines - both diagonals, the middle row, the middle column - while a corner touches three and an edge only two.

In misère, that same arithmetic runs backwards. The centre participates in four lines you are trying not to complete, which makes it the most dangerous square to occupy. Taking it early hands your opponent four different ways to force you into a losing line.

So the opening advice flips completely: avoid the centre, prefer the edges. Edges touch only two lines, which makes them the safest real estate on the board - the exact inverse of the standard game, where edges are the weakest opening.

That reversal alone is worth playing a few games for. Watching your own instincts actively work against you is a strange and instructive experience.

Why an extra move can now hurt you

Here is where it gets theoretically interesting.

The standard game has a famous shortcut - the strategy-stealing argument, which proves the second player can never have a winning strategy. It works because an extra piece on the board can never hurt you: a spare mark is either useful or irrelevant, so the first player can afford to make an arbitrary move and then borrow the second player's strategy.

Misère destroys that assumption at the root. An extra mark is frequently a liability - it is one more square filling one more line you must avoid completing. Being forced to move when you would rather pass is the core difficulty of the entire game.

So the elegant argument that settles the normal game says nothing here. The analysis has to be done concretely, position by position, which is why misère versions of games are generally harder to solve than their standard counterparts. That is a broad pattern in combinatorial game theory rather than a quirk of this one game.

(The strategy-stealing argument itself, and what it does prove, is covered separately.)

Zugzwang, on a nine-square board

The mechanism that decides misère games has a name borrowed from chess: zugzwang, the obligation to move when every available move damages you.

In standard tic tac toe zugzwang essentially cannot happen, because placing a mark never hurts. In misère it is the entire game. The board fills, your safe squares run out, and eventually you must complete a line because there is nothing else left.

Which means the real skill is counting safe squares - specifically, working out who runs out first. Late in a misère game you should be asking not "what is my best move" but "how many moves do I have left that do not lose, and how many does my opponent have".

That is parity thinking, and it is the same counting that decides close games in the disc-flipping family. Anyone who plays Reversi knows this: whether a region has an odd or even number of empty squares determines who is forced to move into it last - and being compelled to move into a bad square is exactly the misère losing condition wearing different clothes.

Practical guidance

What actually works at the board:

  • Open on an edge. Two lines of exposure instead of three or four.
  • Avoid taking two squares in the same line early. Every pair you own in a line is a loaded gun pointed at yourself.
  • Spread across lines rather than clustering. The aim is maximum marks with minimum line overlap.
  • Force your opponent into clusters. When they have two in a line, the third square becomes poison for them - and you can often make it their only legal option.
  • Count safe squares from about move five. This is where games are decided, and it is arithmetic rather than intuition.

Why misère versions exist at all

Inverting a win condition is one of the oldest tricks in game design, and it is more than a novelty. It is a genuinely efficient way to produce a new game from an exhausted one - the equipment is identical, the rules take one sentence to explain, and the resulting strategy has almost nothing in common with the original.

The draughts family does exactly this. In the anti-checkers variant, losing all your pieces wins, while compulsory capture remains in force - so the rule that normally drives your attack becomes the thing your opponent exploits against you. If you play Checkers online in that form, every principle you rely on has to be rebuilt from scratch.

The disc games do it too, and the tile-matching world has its equivalents. The pattern is consistent: inverting the goal while keeping the mechanics is the fastest way to find out which of your habits were reasoning and which were reflex.

Is it solved?

Yes - misère tic tac toe on a 3x3 board is small enough to analyse exhaustively, and like the standard game it is a draw with perfect play. Both players can always avoid being forced into a line.

But it is meaningfully harder for humans, because the drawing strategy is less intuitive and the losing moves are less obvious. In the standard game a blunder is usually visible - you failed to block an obvious threat. In misère, blunders happen quietly, several moves before the loss, when you take a square that removes your own future flexibility.

That delay between mistake and consequence is what makes it feel harder despite being the same size of problem.

Give it a go

Play a few rounds of misère and deliberately open on an edge every time. Then play one where you open in the centre, just to feel how quickly it goes wrong. The contrast is stark and it teaches the whole game in about five minutes.

After that, 5x5 is the natural next stop, because a bigger board changes the safe-square arithmetic substantially. The rules cover every variant here, the maths page goes deeper on the theory, and two-player mode is where misère is most fun - it is a much better game against a human who is also fighting their own instincts.