End of game under AGA rules vs .sgf files

[herzbube]

After studying the AGA rules (external link to PDF) I have a few questions which all revolve around rule 10:

10) Disputes: If the players disagree about the status of a group of stones left on the board after both have passed, play is resumed [...] The game is over when the players agree on the status of all groups on the board, or, failing such agreement, if both players pass twice in succession. [...]

With my knowledge of the English language, the expression "both players pass twice in succession" can mean two things:

• 4 pass moves in a row
• 2 pass moves, followed by 1-n stone-placing moves, followed by another 2 pass moves

Can anyone enlighten me as to which one of the two meanings it is?

In the PDF, the text in curly braces that adds explanation to the boldfaced rule text certainly seems to indicate that meaning 1 (4 pass moves in a row) is intended. If that is indeed the case, then what about rule 11 which states "white must make the last move"? Because with 4 pass moves in a row we could have Wpass-Bpass-Wpass-Bpass, i.e. black moving last. As far as I understand, such a sequence would also defeat the purpose of pass stones under territory scoring (rule 7).

Or am I reading the text too literal here and I should understand rule 10 to mean this?

• 2 or 3 passes (depending on who starts passing)
• followed by scoring + dispute
• followed by Black passing and White passing

Yours confusedly
Patrick

[Bill Spight]

Since I am one of the people who came up with the idea of pass stones, I can assure you that rule 11) always applies. If play ends after a pass by Black, White must pass. The point is, in an even game, to have the same number of stones by each player on the board during counting, so that the area score and territory score are the same. (The rules make an adjustment for handicap games.)

In rule 9), two consecutive passes do not end the game, they signal the end of the game. If the players agree about which stones are alive or dead, rule 11) applies. If they do not agree, they resume play.

IIUC, the four pass situation arises under rule 10 in the following scenario. There are two passes. The players are unable to agree about which stones are alive or dead. Play resumes, but each player now passes. That makes four passes in a row, and play stops with all stones on the board considered to be alive. Then rule 11) applies.

[skydyr]

As I read the document, this procedure only occurs after both players have passed once in succession, moving on to the stage where they agree on what is alive and what is dead, and then disagree about the status of the stones on the board.

If the game has come to this point, and there is a disagreement, play resumes. To end this resumed play, once they have removed all the dead stones on the board by capturing them through play, both players pass twice in succession, meaning to me that A passes, B passes, A passes again, and B passes again, without intervening moves. At that point, any stones still on the board are considered alive for scoring purposes. There is some commentary here: http://www.cs.cmu.edu/~wjh/go/rules/AGA.commentary.html suggesting that this is to prevent a player from keeping the game from ending indefinitely by refusing to agree on the status of groups, requiring that someone who claims stones to be dead be forced to remove them or have them deemed alive. I think the reason to require the double passing instead of single passing is so that a pass can be used as a ko threat in the event that one side can't actually fill a ko, but I'm not certain.

After all of this has occurred, if black played the last pass, white then passes as well to maintain parity and make the territory and area counts equivalent.

[xed_over]

I generally like AGA rules, but that's just plain stupid and has no useful value, that I can see, for each player to pass twice in a row -- again.

... [dispute settled after resumed play] ...
A: "I think the game is really over now, so I pass."
[hands over pass stone]
B: "I agree, I pass too."
[hands over pass stone]
A: "But since we resumed play earlier, the rules say we now have to pass again, even though we both already agreed the game was really over now. So I pass again."
[hands over another pass stone]
B: "Ok, me too, ... again."
[hands over yet another pass stone (which doesn't change the score at all from any of the previous passes)]

ugh, what is the point??!?!
really!?

(and regardless of the 4-pass rule, ultimately white must still play last -- so if you already passed a 3rd time before the dispute, and then the resumed play results in another odd-parity, then white will have to pass yet again, and lose a point due to the dispute.)

[DrStraw]

Pass stones arules re the reason I lost interest in playing any more AGA tournaments. They just make the end of the game confusing, as this thread shows.

[Bill Spight]

I generally like AGA rules, but that's just plain stupid and has no useful value, that I can see, for each player to pass twice in a row -- again.

... [dispute settled after resumed play] ...
A: "I think the game is really over now, so I pass."
[hands over pass stone]
B: "I agree, I pass too."
[hands over pass stone]
A: "But since we resumed play earlier, the rules say we now have to pass again, even though we both already agreed the game was really over now. So I pass again."
[hands over another pass stone]
B: "Ok, me too, ... again."
[hands over yet another pass stone (which doesn't change the score at all from any of the previous passes)]

ugh, what is the point??!?!
really!?

No, not really.

The game is over when the players agree on the status of all groups on the board, or, failing such agreement, if both players pass twice in succession. In this case any stones remaining on the board are deemed alive.

Emphasis added.

So you have this:

[Players disagree about which stones are alive or dead and resume play until the matter appears to be resolved.]
A: "I think the game is really over now, so I pass."
[hands over pass stone]
B: "I agree, I pass too."
[hands over pass stone]
[Players agree about which stones are dead. White makes last pass if necessary.]

Or this:

[Players disagree about which stones are alive or dead and resume play until the matter appears to be resolved.]
A: "I think the game is really over now, so I pass."
[hands over pass stone]
B: "I agree, I pass too."
[hands over pass stone]
[Players disagree about which stones are alive or dead and resume play again.]
A: "Pass."
[hands over pass stone]
B: "Pass."
[hands over pass stone]
[Now the 4 pass rule applies. Play stops and all stones on the board are deemed to be alive. White passes if necessary.]

[Bill Spight]

(and regardless of the 4-pass rule, ultimately white must still play last -- so if you already passed a 3rd time before the dispute, and then the resumed play results in another odd-parity, then white will have to pass yet again, and lose a point due to the dispute.)

No. First, rule 11 does not apply before agreement. (I know that is not clear.) But suppose that White made the third pass anyway, before resumption. Now play resumes with Black playing first. Therefore, for the number of stones played to remain equal, White must pass last. White can only lose a point if White passes last before the disagreement, and then illegally plays first in the resumption. (And the TD, if made aware of that, should give White the point back, IMO.)

[herzbube]

Thanks @skydyr for the link to the commentary document - had I seen this before I probably would not have started this thread. Although Bill's clarifications are still needed to make everything crystal-clear.

Thank you @Bill for your explanations. The second example in your first reply to xed_over makes it clear that a game can never have more than 2 scoring phases, to prevent infinite play. This is something that even the commentary document fails to explicitly state.

@DrStraw I have never played using the AGA rules, but from a theoretical point of view they look very elegant, especially because of pass stones.

[DrStraw]

@DrStraw I have never played using the AGA rules, but from a theoretical point of view they look very elegant, especially because of pass stones.

Actually, they are anything but elegant, exactly because of the pass stone.

[Bill Spight]

Thank you @Bill for your explanations. The second example in your first reply to xed_over makes it clear that a game can never have more than 2 scoring phases, to prevent infinite play. This is something that even the commentary document fails to explicitly state.

I am not sure that I was completely clear. I interpret the four pass rule to refer to four consecutive passes. And the only way that can happen is to have two passes to stop play, followed by disagreement about which stones are alive or dead, followed by resumption of play, followed by two immediate passes.

[xed_over]

@DrStraw I have never played using the AGA rules, but from a theoretical point of view they look very elegant, especially because of pass stones.

Actually, they are anything but elegant, exactly because of the pass stone.

except for the 4 pass thing (that I'm still a little confused by), I think the rules are completely elegant because of the pass stones

AGA rules are at their heart area based rules, which I like to recommend to beginners, because they can just keep playing until they know all the life and death situations are resolved, yet they can still count using Japanese style counting without worrying about changing the score because of the pass stones. The rules were designed so that any two lower level players could play and resolve their own issues without requiring a stronger player tell them how to end the game. I think that's rather clever. And mathematically solid to boot.

Ing tried to do the same thing with his rules -- make them beginner friendly. But I think the AGA rules are better.

But, after gaining a little skill, its probably not necessary to insist on using them. So, maybe you're right afterall.

[ez4u]

"If you are out to describe the truth, leave elegance to the tailor."
- Albert Einstein

I think the same sentiment applies to disputes. You can have elegant rules or you can have rules that will plod their way through all the arcane disputes that can (and, occasionally at least, do) arise IRL. However, the latter won't be elegant and the former won't make it past RJ's devious mind.

[skydyr]

(and regardless of the 4-pass rule, ultimately white must still play last -- so if you already passed a 3rd time before the dispute, and then the resumed play results in another odd-parity, then white will have to pass yet again, and lose a point due to the dispute.)

You're not supposed to do the 'white passes last' thing until the status of groups has been agreed upon or determined through rule 10.

[hyperpape]

Pass stones arules re the reason I lost interest in playing any more AGA tournaments. They just make the end of the game confusing, as this thread shows.

I hope you didn't enjoy tournaments much to begin with.

[DrStraw]

Pass stones arules re the reason I lost interest in playing any more AGA tournaments. They just make the end of the game confusing, as this thread shows.

I hope you didn't enjoy tournaments much to begin with.

Why would you say that? Of course I did otherwise I would not have gone. I'm not a masochist.