Order and Chaos — the Lopsided Game

Play Order and Chaos against the computer, and see why every board small enough to solve exactly goes to Order.

Both players may place either symbol on any empty square. Order wins with five in a row of a single symbol, in any direction; Chaos wins by filling the board without one. Order moves first.

Every board small enough to solve goes to Order

The two roles sound like mirror images — one builds, one blocks. They are not. Solving reduced boards exactly, every position and both players perfect:

BoardTargetWinning linesLines per squareWinner
3 × 3 3 in a row 8 0.89 Order
4 × 4 3 in a row 24 1.50 Order
4 × 4 4 in a row 10 0.63 Order
5 × 5 4 in a row 28 1.12 Order

The 4 × 4 with 4 in a row is the telling one. Only 10 winning lines exist on the entire board — 0.63 per square, barely more than half — and Order still wins. Scarcity of targets is not what would save Chaos.

The full 6 × 6 game is far too large to solve this way, so it is not claimed here. The one board that would be most revealing — 5 × 5 needing five in a row, the sparsest of all at 0.48 lines per square — did not finish inside the time budget, and I am not going to guess which way it falls.

Chaos’s real problem: there is no passing move

The asymmetry is not about who has more ways to win. It is that Chaos has to keep placing symbols, and every symbol placed is a brick Order may later build with. Chaos would dearly like to pass. The rules do not allow it.

Two facts make that concrete, and both are exact rather than approximate:

  • If any line is one square from complete and Order is to move, Order simply fills it. Order never needs to see further than one move.
  • If two such lines exist with different gaps and Chaos is to move, Order has already won — Chaos can spoil at most one square per turn.

The second is the whole game in miniature, and it is easy to check rather than take on trust: set up two threats needing different squares, then try all sixty replies Chaos has. Not one of them clears both. Order does not need a clever plan — only two threats at once, built partly out of the symbols Chaos was forced to hand over.

The board is exactly as crowded as noughts and crosses

One reason 6 × 6 with five in a row is the standard board: it packs in almost exactly the same density of winning lines as the smallest interesting game there is.

BoardTargetLinesSquaresLines per square
3 × 3 3 8 9 0.89
4 × 4 3 24 16 1.50
4 × 4 4 10 16 0.63
5 × 5 4 28 25 1.12
5 × 5 5 12 25 0.48
6 × 6 5 32 36 0.89
6 × 6 6 14 36 0.39

The standard game sits at 0.89 winning lines per square; noughts and crosses sits at 0.89. Whenever the target equals the width of the board the count collapses to just 2n + 2 — n rows, n columns and the two long diagonals — which is why those variants are the ones where Chaos has any hope at all.

How to use

  1. Choose whether to play Order or Chaos.
  2. Choose which symbol to place, then click any empty square.
  3. Order needs five in a row of one symbol, in any direction.
  4. Chaos needs the board full with no such line.

Frequently asked questions

Is Order and Chaos a fair game?

Not as even as it looks. Every board small enough to solve exactly goes to Order — 3 by 3 needing three, 4 by 4 needing three, 4 by 4 needing four, and 5 by 5 needing four. The two roles sound like mirror images but they do not behave like them.

Why does Order have the advantage?

Because Chaos cannot pass. Every turn Chaos must place a symbol somewhere, and every symbol placed is material Order may later build with. Order only has to make two threats at once, and some of what those threats are built from was handed over by Chaos.

What is the two-threat rule?

If two lines are each one square from complete and those squares are different, Order has already won. Chaos can only spoil one square per turn, so whichever it blocks, the other gets filled next move. Setting up such a position and trying all sixty Chaos replies confirms not one of them clears both.

Does having fewer winning lines help Chaos?

Less than you would expect. On a 4 by 4 board needing four in a row there are only ten winning lines on the whole board — 0.63 per square — and Order still wins with perfect play. Scarcity of targets is not what would save Chaos.

Has the standard 6 by 6 game been solved here?

No, and it is not claimed. The full board is far too large for exhaustive solving. What is shown is every reduced board that could be solved outright, and they all go to Order. The most revealing variant, 5 by 5 needing five, did not finish inside the time budget.

How many winning lines does the standard board have?

Thirty-two: rows, columns and diagonals long enough to hold five. That works out at 0.89 per square, which is almost exactly the density of noughts and crosses at 0.89 — one reason the 6 by 6 board feels like the right size for the game.

Why does the count drop so sharply when the target equals the board width?

Because only whole-board lines survive: n rows, n columns and the two long diagonals, so 2n plus 2 and nothing else. Every partial line vanishes. Those are the variants where Chaos has the most hope, and they are the ones hardest to solve.

Does this send anything anywhere?

No. The game and the opponent both run entirely in your browser.

🔒 This tool runs entirely in your browser. Nothing you enter is uploaded, logged, or stored.