I don’t think 6x6 has even been conclusively solved for regular Go yet. There are have been some claims regarding even larger boards, like 7x7, but the rigor of those results have been questioned.
Much of that work was done well over a decade ago, so it’s possible that revisiting these problems with current hardware could expand the frontier of knowledge. I imagine that applying similar methods, and even adapting the code, could be done to extend these solution searches from standard Go to TMG.
The complexity should be roughly equal, since the number of moves stays roughly equal for both players, so disregarding that TMG is a lot less popular, I’d expect a similar amount of time to solve TMG as compared to regular Go
If you pass on a “double” move, it will skip to your opponent’s turn. The reason is not very interesting - I built this on top of the standard Baduk class, and it will end the game if you pass twice (even if it’s the same player).
I guess this raises the question - how important to this game is passing (without the intention of ending the game)? Is the behavior for passing even well defined?
Whose move is it anyway?
I’d like to add a widget that tells what moves are coming up to help with planning. Something simple like:
…WBBWBWW…
For now, you have to keep track of the sequence yourself if you want to plan ahead. Of course, the server will prevent you from playing when it’s not your turn.
When one has a double-move, it seems reasonable to allow a player to either:
Play two stones (typical case).
Fully pass (playing zero stones).
Partially pass (play only one stone).
The second case would represent what one would do if the game is essentially over, but the third case could also reasonably arise at the end of a game, if there is just one more worthwhile stone that should be played, but playing another stone might be disadvantage (e.g., only filling in one’s own eye).
The question of how many passes should end the game is a bit a tricky. It could be possible that a player how no more effective single moves, but does have an effective double move. Thus, they might want to pass at a single move, waiting for their next double move in order to launch into an attack to finish something.
I guess one way to handle such situations is through the players resuming the game to settle such positions.
On the govariants.com version, it is still possible to make a single move if you make the move before the pass. (For example: B at A1; B pass)
But there is still one scenario missed by skipping the second move- you want to skip a move, but don’t want your opponent to end the game. For this example consider the move sequence BBW, if passes didn’t skip a move:
B pass; B A1; W pass (game continues?)
B A1; B pass; W pass (game enters scoring phase)
This might be it. Unfortunately… we haven’t added a scoring phase to any variant and so resumption isn’t possible (yet)
Actually before going to one player always black, they would do the 1-rank handicap difference of the weaker player taking black in 2 games out of 3, called sen-ai-sen.
The part handling scoring is probably trickier than the implementation of the rest of the game.
My understanding is that under all typical area scoring rules, the game resumption does not depend on who requests the resumptions or raises a dispute. However, under Japanese rules, the first to play after resuming is the opponent of whoever requested the resumption. Further, I should emphasize that life/death disputes under Japanese rules usually should not be settled by resuming the game (as that may alter the score and outcome).
Note: while AGA has the rule of “white passes last”, this convention should not be applied until all resumptions have ceased (due to all life/death disputes being resolved), and thus it is basically only just applied at the point when counting is performed (essentially White gives an extra pass stone). Further, this convention (along with the pass stones) is unnecessary if using area counting, since it is only an accounting trick used to make territory counting agree with area counting.
Is there any number of passes where we can end the game with confidence? I fail to see how we can guarantee that there will never be a time point in the move-colour sequence s.t. continuing is profitable.
I think in practice, a pass by both players in a row should suffice. If you think there’s some advantage to be had by passing but not ending the game, you can always play dame, because area scoring.
Coming back to the original idea of using this instead of komi. I think that could work if you use the sequence not on a move by move basis, but to decide who gets the next big move after a Sente sequence. I.e. if the Thue-Morse sequence says you get to move twice, you get to go again after you make a Gote move.
That would obviously require a more precise way to determine which moves are Sente/Gote, so that it can be part of the rules. One way to achieve that is Coupon Go, where you can take a decreasing number of points instead of making a move on the board. In Coupon Go the strategy is to play a Sente sequence and then start taking coupons until the next big move is more valuable than the current largest coupon. So the idea would be to use the Thue Morse sequence to determine who gets to play after a coupon taking session, but I’m not sure if that works out.
