I’m on a bus, so let’s talk about solving Go, by which I mean komiless Go. The way I see it, there are six gradually more rigorous levels of solution.
-
Partial ultra-weak. This excludes one or more outcomes. There is a partial ultra-weak solution to all symmetrical games, which is the strategy-stealing argument. This states that whatever strategy is used by the second player can be used by the first player first, and that therefore the second player can’t win.
-
Complete ultra-weak. This proves the outcome of the game. Go on large boards, like 19x19, can be shown to be a win for Black. This is because Black moves first and the value of the optimal move decreases throughout the game.
-
First weak solution. This is a complete game record in which Black wins and White cannot be proven to have made a critical mistake. Middlegame and opening mistakes can be very hard to actually prove, which is why we have the next level.
-
Second weak solution. This is a level 3 solution in which no observer can identify a critical mistake by White (in their opinion).
-
Third weak solution. This is the first grindy solution, in which it must actually be proven that no possible defence by White can alter the outcome. A huge amount of positions have to be explored and this is well beyond current computational capability.
-
Strong solution. The strong solution provides an optimal path to the correct outcome from any possible Go position.