dispute resumption rules and ko

[dmwit]

Are there any rulesets which distinguish between a declaration that the game is over and a move which places no stones (that is, has two different kinds of pass)? It seems like this would solve a lot of weird game resumption rules, since we can allow the second "game-is-over pass" player to resume without worrying that the first player only passed to avoid an illegal ko move. (The canonical way to avoid an illegal ko move is to make a "placing-no-stones pass"; the two players can alternate playing this kind of move as long as they want without ending the game.)

[Joaz Banbeck]

I'm confused. Could you perhaps restate the question?

[oren]

:scratch: I'm confused. Could you perhaps restate the question?

With an example board would help too.

[tchan001]

I think he's talking about how passing twice ends the game. He is thinking about passing instead of finding a ko threat but he fears that the other player would also pass and end the game. Thus he wants to know how to resume the game in case both players pass during a ko fight.

[Bill Spight]

I think he's talking about how passing twice ends the game. He is thinking about passing instead of finding a ko threat but he fears that the other player would also pass and end the game. Thus he wants to know how to resume the game in case both players pass during a ko fight.

That's my guess, too. Ending play by two passes can allow final positions that are similar to Moonshine Life, with an unfinished ko that one player would like to take. AGA Rules and Tromp-Taylor Rules allow such final positions. Ing Rules, Japanese Rules, Korean Rules, Chinese Rules, and Spight Rules do not, all in different ways.

[dmwit]

Sometimes, after a game ends, the two players can't agree on which stones are dead or alive. Many rulesets have some allowances for the game to resume to help resolve such disputes. There's a couple of choices to make when writing such rules; for example: who plays first? may the first player retake a ko, if one is on the board? etc.

It seems like these questions might be easier to answer if we distinguish two different kinds of "pass" move. One is a move which definitely will not end the game, but which transfers play back to the other player; in particular, you can use this move if there is nothing useful for you to do on the board other than retake a ko, but you are not (yet) allowed to retake the ko. The other move is a declaration that you think the game is over; two of these in a row end the game (as two of any pass do in many rulesets). Of course this still gives the other player a chance to announce that they think the game is unfinished (by playing a stone).

I suppose I really have a few questions, like: are there any current rulesets that make this distinction (or one like it)? is my assertion that such a distinction would be useful actually true? are there any obvious problems with making such a distinction?

[Tryss]

Just use 3 pass to end the game, and you don't need to do a specific distinction.

[Bill Spight]

Sometimes, after a game ends, the two players can't agree on which stones are dead or alive. Many rulesets have some allowances for the game to resume to help resolve such disputes. There's a couple of choices to make when writing such rules; for example: who plays first? may the first player retake a ko, if one is on the board? etc.

It seems like these questions might be easier to answer if we distinguish two different kinds of "pass" move. One is a move which definitely will not end the game, but which transfers play back to the other player; in particular, you can use this move if there is nothing useful for you to do on the board other than retake a ko, but you are not (yet) allowed to retake the ko. The other move is a declaration that you think the game is over; two of these in a row end the game (as two of any pass do in many rulesets). Of course this still gives the other player a chance to announce that they think the game is unfinished (by playing a stone).

I suppose I really have a few questions, like: are there any current rulesets that make this distinction (or one like it)? is my assertion that such a distinction would be useful actually true? are there any obvious problems with making such a distinction?

Pass is a modern concept in go. Traditionally, games did not end with a succession of passes. Ing was the first, I believe, to designate pass as a move. Under Ing rules it takes four passes to end play, and that takes care of the problem of unfinished kos or superkos. Earlier, in one of the first sets of written rules, Yasunaga had the game end after three of what today we would call passes. That allows the first passer to take a ko if the opponent then passes.

Let me add that when I learned go we did not have passes. Games ended by agreement. We did not have any problems with that, but in the history of go problems occasionally arose.

[RobertJasiek]

Ing was the first, I believe, to designate pass as a move.

Robinson / Olmsted might have been the first, some five decades earlier.

[christian freeling]

Let me add that when I learned go we did not have passes. Games ended by agreement. We did not have any problems with that, but in the history of go problems occasionally arose.

With fair play in the spirit of the game, players need little formalization. But it's not always like that. In the mindsports arena Go is introduced thus:

Go is arguably the quintessential territory game. However, even Occam's Razor cannot prevent the ambiguities arising from cycles, "ko" being the most basic one. Go's rules appear simple and concise, but the devil is in the details. Yet Go, as a concept, has an almost universal appeal, and there is a broad concensus about all but the most eccentric positions.

As a result we received some guestbook entries stating in various degrees of eloquence that our stupidity could hardly be overestimated .

[RobertJasiek]

the ambiguities arising from cycles

Which ambiguities? Since my general definition, none remains.
http://home.snafu.de/jasiek/ko.pdf
On the rules level, avoiding ambiguity can be even simpler, using, e.g., positional superko. In terms of strategy, it is more appropriate to speak of "complexity".

[Bill Spight]

Ing was the first, I believe, to designate pass as a move.

Robinson / Olmsted might have been the first, some five decades earlier.

Thanks. I have not seen their writing.

I just did web search, and found a couple of interesting things. Thorp and Walden ("A Computer Assisted Study of Go on MxN Boards", 1972) have pass as a move. They base their rules on Robinson and Olmstead ("A Treatise on the Rules of Go", 1964). That's a decade before Ing. Ing's four pass rule can be traced back to Shimada ("The Mathematical Theory of Go", 1958). OC, Shimada does not consider a pass to be a play.

[christian freeling]

On the rules level, avoiding ambiguity can be even simpler, using, e.g., positional superko. In terms of strategy, it is more appropriate to speak of "complexity".

You got it .
the arena

[RobertJasiek]

I have not seen their writing.

Ask Terry for "Rationalization of Go" and "Structure of Go".
The latter is a refinement of the former.

[Harleqin]

Under Ing rules it takes four passes to end play, and that takes care of the problem of unfinished kos or superkos.

This is a bit misleading. Yes, there are four passes, but they do not have the same meaning. First, there are two consecutive passes that put the game into "removal mode" (my wording). Then, after all removals are done, the players use passes again to put the game into "scoring mode", in which no further removals are allowed.

(The inaccurate communication of "four passes" has led to one incident where, in a certain tournament that was announced as using Ing rules, a certain player removed all dead stones during play, then his opponent passed, then he passed, then his opponent passed (which was a mistake), then he passed.)

So, the first two passes mean "I have no worthwhile moves", while the second two passes mean "There are no dead stones left, we can score now."

This is in a way an answer to the original question.

In my personal opinion, no amount of consecutive passes should have automatic consequences on the game state.