It'd be neat if there were a button to undo your last move. Sometimes I click on the wrong square.

I want to do a clojurescript AI for it but want to steal some of your canvas stuff so I dont have to spend long making it pretty, would be nice to know under what conditions I can do that.

I just started this today, should have it finished tonight or tomorrow. only have to write the evalue funtion()

Great implementation! I've been playing with a few of my friends and have been having a great time on http://ken-soft.com/tictactoe/.

In your last commit 53655b0 when you made X go first, I think you forgot to switch the color line that goes through the x's and o's when you win a sub-board.

If there are no more moves to be made it should pop a js alert saying that the game was a draw

I'm not sure if this is a feature or a bug. But my initial impression is that the game should force you to play 9 separate tic tac toe games at once, randomly switching between them.

The problem comes from how the random switching works. For example, when it's X's turn to play, it picks one of 9 games randomly, rather than limiting the random game choice only to games where it's X's turn. This leads to the winner being partly determined by random luck rather than tic tac toe skill.

For example, here's the game I played against myself:

The top right corner game is the worst offender, O had 4 plays, and X and only 1.

If this is intended behavior though, please ignore me :)

The Orlin Gambit destroys this fine game -- to salvage, there is a rule variant where sending a player to an already won board is a free play..

Player 1 can play the center eight times.

ABC
DEF
GHI

So Player 1 plays EE, AE, BE, CE ... whatever order Player 2 plays the E board. The last one Player 1 plays "double", so if Player 2 finishes on EI, Player 1 continues with II. Player 1 has two goals here: win the opposite, in this case A square and play I. So player 1 plays AI, BI, CI... unless player 2 plays IE (it eventually needs to). IE is a free play to Player 1 -- it will place an A stone whenever given the chance. So Player 1 will be sent from I to A twice, once by Player 2 playing IE, once by IA, first it will put down AI (keeping Player 2 in the I board) then AA. At this point player 1 wins the A board. After this win, Player 2 will have a chance to put stones in the A board as well as the I board -- but both are already won so it doesn't matter for Player 2.

Aside from the A win, while Player 2 is pinned in I and A it needs send Player 1 to B, C, D, F, G and H twice. Player 1 wins before that finishes. The longest Player 2 can hold out is if it sends Player 1 to B, D, F, H twice and C and G once -- 20 rounds (8 rounds where player 2 played E stones, then the special IE, IA rounds and then IB,ID,IF,IH,AB,AD,AF,AH,AC,AG for example in any order). There are variants of this pattern but the number of rouns is the same.

There is no difference if Player 2 finishes the "center round" on a side one, say EF. Player 1 will strive for the opposite, the D board, with the same tactic: Player 2 needs to play FD and FE which gives two chances to play on D, so Player 1 plays DF and then DD for the win.