What about passing?Mike Novack wrote:
In go, black and white stones are placed alternately and once placed, not moved unless removed by capture.
White passes
White passes
What about passing?Mike Novack wrote:
In go, black and white stones are placed alternately and once placed, not moved unless removed by capture.
Mike Novack wrote:Question:
Assuming the original definition of legal state (all groups have at least one liberty) there is a question of "reachable" states. Take this example:
In go, black and white stones are placed alternately and once placed, not moved unless removed by capture. Only one white stone could have been captured (the number of black stones placed on the board can never be less than the number of white stones placed) so there are too few white stones present in the diagram.
So that would be a legal state but not a reachable legal state?
Such as? This doesn't seem obvious to me at all. I think we can construct arbitrarily long sequences of capturing and under-the-stones tesujis to create any (valid) board state we want. Even a single white stone, assuming black started and never passed, is possible.HermanHiddema wrote: (but of course there exist actual unreachable positions)
uPWarrior wrote:Such as? This doesn't seem obvious to me at all. I think we can construct arbitrarily long sequences of capturing and under-the-stones tesujis to create any (valid) board state we want. Even a single white stone, assuming black started and never passed, is possible.HermanHiddema wrote: (but of course there exist actual unreachable positions)
Oh, that's a handicap game.HermanHiddema wrote:uPWarrior wrote:Such as? This doesn't seem obvious to me at all. I think we can construct arbitrarily long sequences of capturing and under-the-stones tesujis to create any (valid) board state we want. Even a single white stone, assuming black started and never passed, is possible.HermanHiddema wrote: (but of course there exist actual unreachable positions)
Great. The negative has 2 eyes.HermanHiddema wrote:uPWarrior wrote:Such as? This doesn't seem obvious to me at all. I think we can construct arbitrarily long sequences of capturing and under-the-stones tesujis to create any (valid) board state we want. Even a single white stone, assuming black started and never passed, is possible.HermanHiddema wrote: (but of course there exist actual unreachable positions)
HermanHiddema wrote:
There are no unreachable legal states, unless you're willing to impose certain assumptions on theMike Novack wrote:Question:
Assuming the original definition of legal state (all groups have at least one liberty) there is a question of "reachable" states. Take this example:
In go, black and white stones are placed alternately and once placed, not moved unless removed by capture. Only one white stone could have been captured (the number of black stones placed on the board can never be less than the number of white stones placed) so there are too few white stones present in the diagram.
So that would be a legal state but not a reachable legal state?
I think one takeaway from this is that it exposes how much and aggressively we as humans prune the game tree when evaluating moves and positions. Presumably, as we get stronger, we do it even more. This is demonstrated both by beginning players who worry that their position might collapse if their opponent plays at A, B, and C, as well as by professionals who overlook a sequence that completely destroys their opponent in a game because it requires an unforced empty triangle, or similar "bad" moves.dohduhdah wrote: There are no unreachable legal states, unless you're willing to impose certain assumptions on the
behavior of the players, like acting in accordance with the goal of the game and not just playing
moves that are allowed according to the rules.
For instance, it would be allowed for one player to keep passing from the start as the other player
keeps playing moves until they feel they have enough stones on the goban that the opponent will no
longer be able to create any living groups.
It would be a nonsensical game, but it's still a valid game, as both players play valid moves.
The assumption here, as introduced by Mike, is no passes.dohduhdah wrote:I don't see how there are states that are unreachable, except under certain assumptions
(like players not doing weird things like filling up their own eyes, which would usually
be unreasonable given the goal of the game).
For the concept of theoretical games, where the only assumption is that both sides play
in accordance with the rules, any valid state can obviously be reached.
Now this is a great picture. Before, I was too busy (read: lazy) to try to figure out what your diagrams are actually all about, while this shows it very clearly with (almost) one glance.