Zugzwang

[seberle]

So, I was looking at the Zugzwang article in Sensei's Library (https://senseis.xmp.net/?Zugzwang), and I was startled by the top example.


User helopticor commented: "If white moves first, they get behind by three. If black moves first, they get ahead by only two. Both players want the other one to go first."

I've been playing for years and I'd never seen such an example. I'm assuming white at the top and black at the bottom are connected to living groups.
* What would you do in such a case?
* I guess if the game is near its end, black would capture and eventually gain two points? Better than passing and gaining no points, right?
* If both players pass, can black just claim the two points under Japanese rules?
* Has this position ever come up in real games? Did black or white make a serious mistake getting into such a position?

[RobertJasiek]

During the game, correct play depends on ko threats.

At / after the game end, assessment depends on the rules. If Japanese rules are used, assessment depends on which Japanese rules are used.

Regardless of the rules, local variations also involve a possible ko shape.

The players need not have made a mistake in creating the position if the outside is filled with further stones.

I have not heard of the local position occurring in any game.

***

You declare that you are from Niger. I wonder: do you know any other go player in your country? Have you discovered the game in the internet?

[seberle]

Thanks for the quick reply! I know you are an expert on Japanese rules and have appreciated some of your writings on rules. How would this be scored under current Japanese rules?

***

As for me, I have lived in Niger for many years, but I'm not originally from here. I first learned the game in Canada, though that was actually via the Internet. I have known only a few other go players here in Niger, all expats. The only one I play with now lives in another town, so we rarely have live games together.

[Bill Spight]

So, I was looking at the Zugzwang article in Sensei's Library (https://senseis.xmp.net/?Zugzwang), and I was startled by the top example.


User helopticor commented: "If white moves first, they get behind by three. If black moves first, they get ahead by only two. Both players want the other one to go first."

I've been playing for years and I'd never seen such an example. I'm assuming white at the top and black at the bottom are connected to living groups.
* What would you do in such a case?

As Robert says, modern rules cover this situation. They do not all agree.

* I guess if the game is near its end, black would capture and eventually gain two points? Better than passing and gaining no points, right?

That's how the Japanese 1989 rules treat this position. OC, the ko threat situation matters.

* If both players pass, can black just claim the two points under Japanese rules?

Not under the J89 rules. But Black can request reopening play, and White must grant that request, with White having the right to play first. Presumably White will not play in the corner and then Black can do so.

* Has this position ever come up in real games? Did black or white make a serious mistake getting into such a position?

To the best of my knowledge, this position has not come up in an actual game. But go has a long history, and it may well have come up in some unofficial game at some time and place.

A position arose in a game of Honinbo Shuwa's in the 19th century, from which this position could have arisen, but didn't. The story is that Shuwa declared that this position was worth "Three Points without Capturing" for Black, and that is the ruling in the Japanese 1949 rules. If Shuwa explained his reasoning, it was not passed down to us. My speculation is that he used similar reasoning to hypothetical play to capture dead stones inside your own territory. You evaluate the final position, but add 1 point for each stone played in the hypothetical play. There are other ways to evaluate this position that give the same answer. See, for instance, viewtopic.php?p=195058#p195058
and
viewtopic.php?p=194948#p194948

Edit: Also, the play under Lasker-Maas rules is interesting, as all of the newly played stones will be prisoners. In that case Black will have captured 5 prisoners and White will have captured 3, but Black will have played 5 prisoners, thus reducing White's total prisoners by 2, while White will have played 4 prisoners, thus increasing Black's total prisoners by 1. Black will have thus have gained 3 points in the net prisoner count.

[John Fairbairn]

Has this position ever come up in real games?

Yes. Once (although eventually played out). It came up in a game (which survives) between Honinbo Shuwa and Yasui Sanchi in 1839, which was remarkable not just for the torazu sanmoku but for the fact it was played over 22 sessions. One wonders how many sessions were devoted to sorting out this little mess.

Tradition has it that Shuwa gave it the name torazu sanmoku (though it is also called utazu sanmoku - same meaning) but that may be conflation with the fact that he was involved in the actual game.

There is a bigger version called torazu gomoku but that seems to have existed only in the lab.

[Bill Spight]

Here is how the position almost arose in the game (GoGoD 1839-03-15a).

$$Wcm96 Honinbo Shuwa (W) vs. Yasui Sanchi, moves 196 to 198
$$ ---------------------------------------
$$ | . X X . O . O . X . . . . . . . . . . |
$$ | O O X X . O . X X O . O O O X . . X . |
$$ | . . O X X X X X O O . . X . X O O X . |
$$ | . . O O X O . . O X O . O X X X X . . |
$$ | . O O X O O X X X X O . O O O O X X . |
$$ | . X O X . . . . . X X X . . . O O X . |
$$ | O O X X O O O O . O X . X . . O X . . |
$$ | . X . X O . X X O O O O . . . O X X X |
$$ | . . X . X X . . X 1 . . . . . O O X O |
$$ | . . . O . . . . . , . O . O . , . O O |
$$ | . X . . . X . . X X . . X O . . . . . |
$$ | . . X O O . . X . . . . X O . . O O . |
$$ | . X O O . O . . O . X . X X O O X O . |
$$ | . X O . . . O O . X X . X . X X . X O |
$$ | . X O X X O . . . X . X O X . X X X . |
$$ | . . X X O X O . X O X . O . O , O X X |
$$ | . . . X O . O . O O X O O . . O O O X |
$$ | . . X X O . . . . . O . . . . O X X W |
$$ | . . X O O . . . . . . . . . . a X 2 O |
$$ ---------------------------------------

Click Here To Show Diagram Code

[go]$$Wcm96 Honinbo Shuwa (W) vs. Yasui Sanchi, moves 196 to 198
$$ ---------------------------------------
$$ | . X X . O . O . X . . . . . . . . . . |
$$ | O O X X . O . X X O . O O O X . . X . |
$$ | . . O X X X X X O O . . X . X O O X . |
$$ | . . O O X O . . O X O . O X X X X . . |
$$ | . O O X O O X X X X O . O O O O X X . |
$$ | . X O X . . . . . X X X . . . O O X . |
$$ | O O X X O O O O . O X . X . . O X . . |
$$ | . X . X O . X X O O O O . . . O X X X |
$$ | . . X . X X . . X 1 . . . . . O O X O |
$$ | . . . O . . . . . , . O . O . , . O O |
$$ | . X . . . X . . X X . . X O . . . . . |
$$ | . . X O O . . X . . . . X O . . O O . |
$$ | . X O O . O . . O . X . X X O O X O . |
$$ | . X O . . . O O . X X . X . X X . X O |
$$ | . X O X X O . . . X . X O X . X X X . |
$$ | . . X X O X O . X O X . O . O , O X X |
$$ | . . . X O . O . O O X O O . . O O O X |
$$ | . . X X O . . . . . O . . . . O X X W |
$$ | . . X O O . . . . . . . . . . a X 2 O |
$$ ---------------------------------------[/go]

at

Later, Shuwa could have played at a to make the shape, but Sanchi played in the corner first.

[RobertJasiek]

Under current Japanese, the 1989, rules, see Example II.1

http://www.cs.cmu.edu/%7Ewjh/go/rules/Japanese.html

Their "enable" is ambiguous and must be interpreted by capturable-1 and capturable-2

http://home.snafu.de/jasiek/j2003.html
http://home.snafu.de/jasiek/j1989c.html
http://home.snafu.de/jasiek/j2003com.html

One involved string is capturable-1, the other capturable-2 so we have in-seki.

Therefore, the players should dissolve before the game end (or resume from a game stop).

[seberle]

Thanks RobertJasiek for those helpful documents.

I see that 1. Positions Related to Article 7, Clause 1, Life-and-Death Example 1 in your first document (the 1989 rules), is EXACTLY the position we are discussing, right? Therefore, by these 1989 rules, "If the game ends as shown in the diagram, the white stone and the four black stones are both alive. By Article 8, the position is a seki."

Of course, again, in a real game, the player who would gain two points would play it out rather than letting it be counted as seki.

I hope I've understood correctly.

[Gérard TAILLE]

Thanks RobertJasiek for those helpful documents.

I see that 1. Positions Related to Article 7, Clause 1, Life-and-Death Example 1 in your first document (the 1989 rules), is EXACTLY the position we are discussing, right? Therefore, by these 1989 rules, "If the game ends as shown in the diagram, the white stone and the four black stones are both alive. By Article 8, the position is a seki."

Of course, again, in a real game, the player who would gain two points would play it out rather than letting it be counted as seki.

I hope I've understood correctly.

If I understand correctly the area will be finally counted +2.
What about counting in CGT?
Ignoring the ko problem the game looks like {2||3|3} = +3. Why CGT differs from real game?

[Bill Spight]

What about counting in CGT?
Ignoring the ko problem the game looks like {2||3|3} = +3. Why CGT differs from real game?

It didn't, traditionally. And in the Japanese 1949 rules this position is scored as 3 points for Black. That is why it is called Three Points without Capturing. If Black captures, she only gets 2 points.

The story is that Honinbo Shuwa gave that as a ruling back in the 19th century. But there is no record of such a ruling. However, there is a game of Shuwa's in which the position could have arisen, and Shuwa probably expressed his opinion.

See notes #4 - #6 in this topic.

[Gérard TAILLE]

What about counting in CGT?
Ignoring the ko problem the game looks like {2||3|3} = +3. Why CGT differs from real game?

It didn't. Traditionally, and in the Japanese 1949 rules this position is scored as 3 points for Black. That is why it is called Three Points without Capturing. If Black captures, she only gets 2 points.

The story is that Honinbo Shuwa gave that as a ruling back in the 19th century. But there is no record of such a ruling. However, there is a game of Shuwa's in which the position could have arisen, and Shuwa probably expressed his opinion.

See notes #4 - #6 in this topic.

Yes Bill but according to Japonese89 rule black should play the area before the end of the game for only 2 points. Isn't it true?
Maybe it is a pity that Japonese89 rule forces black to play?

[Bill Spight]

What about counting in CGT?
Ignoring the ko problem the game looks like {2||3|3} = +3. Why CGT differs from real game?

It didn't. Traditionally, and in the Japanese 1949 rules this position is scored as 3 points for Black. That is why it is called Three Points without Capturing. If Black captures, she only gets 2 points.

The story is that Honinbo Shuwa gave that as a ruling back in the 19th century. But there is no record of such a ruling. However, there is a game of Shuwa's in which the position could have arisen, and Shuwa probably expressed his opinion.

See notes #4 - #6 in this topic.

Yes Bill but according to Japonese89 rule black should play the area before the end of the game for only 2 points. Isn't it true?

Depending on the ko threat situation, yes.

Maybe it is a pity that Japonese89 rule forces black to play?

The J49 rules were criticized for having a number of special rulings. The J89 eliminated all of those with the same results, except for this one and the Five Points without Capturing, neither of which are known to have occurred in actual play. Later it was discovered that the anti-seki idea could produce anomalies with different results than under the J49 rules. I think from their point of view the J89 rules were a success.

Rules proposals by Japanese and Western amateurs resulted in territory getting counted in seki, a potentially large difference in results, and one that occurs fairly often.

[Gérard TAILLE]

I am wondering if we have not here a problem to modelize the rule of GO game into a CGT game.
I do not know but maybe the problem comes because a player as a GO player is allowed to pass.
Let's imagine a game in which you can reach a subgame G = {+5|+12}. Now suppose the rule is the following:
1) a player as the right to pass and if the two players passes the game is finished
2) if a subgame is not completely resolved (as the subgame G = {+5|+12}) at the end of the game then the result of the subgame will be 0 (seki).

How building the corresponding CGT game allowing to get the same result as the real game, knowing that G = {+5|+12} = +6 for a CGT game?

[Bill Spight]

2) if a subgame is not completely resolved (as the subgame G = {+5|+12}) at the end of the game then the result of the subgame will be 0 (seki).

How building the corresponding CGT game allowing to get the same result as the real game, knowing that G = {+5|+12} = +6 for a CGT game?

The key is to resolve the game at temperature -1. My rules and Lasker-Maas rules do that. Lasker-Maas rules are quite straightforward. Black plays to +5 on the board with a Black captive, for a net result of +6. If White does not have Black captive to give her, they can trade stones.

[Gérard TAILLE]

2) if a subgame is not completely resolved (as the subgame G = {+5|+12}) at the end of the game then the result of the subgame will be 0 (seki).

How building the corresponding CGT game allowing to get the same result as the real game, knowing that G = {+5|+12} = +6 for a CGT game?

The key is to resolve the game at temperature -1. My rules and Lasker-Maas rules do that. Lasker-Maas rules are quite straightforward. Black plays to +5 on the board with a Black captive, for a net result of +6. If White does not have Black captive to give her, they can trade stones.

OK Bill I see but your rule or Lasker-Maas rule is not J89.
So, if I understood correctly, because in J89 rule there are no ending phase of the game allowing to resolve such subgame by playing at temperature -1, then we cannot modelize the J89 rule by a CGT game can we?