A game is solved when we know who wins it with perfect play from both sides. Tic-Tac-Toe is solved: a draw. Four in a Row is solved: the first player wins. English checkers is solved: a draw. Chess is not, and may never be.

Knowing the answer does not end a game. It just changes what winning means. Here is what is known about the games you can play on this site, and what is left to prove when you sit down to one.

Three levels of solved

Mathematicians separate three strengths of "solved":

LevelWhat is knownExample
Ultra-weakly solvedWho wins from the start, but not howHex on large boards: the first player wins, nobody knows how
Weakly solvedA perfect strategy from the starting positionEnglish checkers
Strongly solvedThe best move from every possible positionTic-Tac-Toe, Four in a Row

A strongly solved game can be played perfectly by a computer from any position, however strange. A weakly solved one only from the start.

Tic-Tac-Toe: a draw

The smallest solved game there is. There are 255,168 possible games of Tic-Tac-Toe on a 3×3 board, and with correct play from both sides every one of them ends in a draw. X wins more of all possible games than O, because moving first is an advantage, but every win depends on a mistake.

That is why the computer in Tic-Tac-Toe cannot be beaten on 3×3. It knows every position, and a draw is the best you can do. On 4×4 and 5×5, where you need four in a row, there is room to outplay it.

More: how to win at Tic-Tac-Toe.

Four in a Row: the first player wins

Solved in 1988, independently and within weeks of each other, by James Dow Allen and Victor Allis. The answer is a forced win for the first player, but only by starting in the centre column. Start in the column next to it and the second player can hold a draw. Start in one of the four outer columns and the second player can win.

The board looks small, but it allows 4,531,985,219,092 different positions. The Perfect level in Four in a Row is an exact solver: it knows the result of every one of them. When it moves first, it always wins. When you move first, the win is there for you, if you never make a single mistake.

More: how to win at Four in a Row.

English checkers: a draw

Solved in 2007 by Jonathan Schaeffer's team at the University of Alberta, after a computation that started in 1989. Their program, Chinook, proved that with perfect play from both sides, English checkers is a draw.

Checkers has about 500 billion billion possible positions, so the proof did not look at all of them. It showed that from the starting position, neither side can force a win against perfect defence. That makes checkers weakly solved: the perfect line from the start is known, but not the best move from every position a real game can reach.

Russian checkers, with backward captures and flying kings, is a different game, and as far as we know it has not been solved in the same way. In Checkers you can play both.

More: how to win at checkers.

Rock Paper Scissors: solved, and still unbeatable

Rock Paper Scissors has a solution of a different kind. Because each pick beats one and loses to one, the only strategy that cannot be exploited is to choose completely at random: a third each. Against that, nothing you do wins more than a third of the time.

That is exactly how the computer in Rock Paper Scissors plays, so it is the one game on this site that luck decides over a few rounds and nobody wins in the long run. People are another matter: they are not random, and their patterns can be read.

More: rock paper scissors strategy.

Solitaire and sudoku: the deal decides first

Single-player games can be "solved" too, deal by deal. For Klondike solitaire, a program called Solvitaire found that about 82% of Turn 3 deals can be won with perfect play, and the rest cannot be won by anyone. For FreeCell almost every deal can be won: out of Microsoft's original 32,000 numbered deals, only one, #11982, has no solution.

A proper sudoku is solved by construction: it has exactly one solution, and it can always be reached by logic.

More: is every game of solitaire winnable? and does every sudoku have one solution?

Dominoes: a different kind of game

Dominoes has never been solved in the way checkers or Four in a Row have, and it may not make sense to try. Those are games of perfect information: both players see everything. In dominoes, half the tiles are hidden in your opponent's hand and the boneyard, and the deal itself is luck.

What you can do is play the odds well. Each number appears on seven tiles, so counting what has been played tells you what your opponent can still hold. That is what the Hard computer in Dominoes does: it sees only its own tiles and guesses yours from what you drew and passed on.

And chess

Chess is not solved, and may never be. Estimates put the number of legal chess positions at around 10 to the power of 44, far beyond anything a computer can search. Every game of chess between people is still an open question.

So what is left to prove?

Quite a lot. A solved game tells you what perfect play leads to. It says nothing about whether you can play perfectly, or whether your opponent can. Most games between people are decided by the first mistake, and the fun is in not making it.

That is the idea behind this site. Draw the computer on 3×3. Beat Perfect at Four in a Row when you move first. Take a draw off Hard at checkers. Then send a friend a link and find out which of you blinks first.

Start in Duel.