Capture go problem

[mitsun]

I agree that seki is a misleading concept in capture go. In chess and maybe other games the German term "zugzwang" is used for this situation.

By the way, what happens if both sides make two one-point eyes and fill all dame, so that there are no legal moves left? Is the game a draw, or does the next player lose?

[Bill Spight]

I agree that seki is a misleading concept in capture go. In chess and maybe other games the German term "zugzwang" is used for this situation.

Well, seki can be used as a ko threat in regular go, but there are no kos in capture go (capture-1, anyway). Otherwise, in a regular seki the player to move will suffer a loss. (There are other standoff situations in regular go that are not seki, such as three-points-without-capturing.)

By the way, what happens if both sides make two one-point eyes and fill all dame, so that there are no legal moves left? Is the game a draw, or does the next player lose?

It is legal to fill your own eye. But yes, there are possible positions in capture go where a player may not have a move. I would call such a position a loss for that player.

[RobertJasiek]

The term "seki" only makes sense when passing is allowed.

As Tang Rules show, seki makes sense also without passes.
viewtopic.php?f=45&t=4005

[ChradH]

The term "seki" only makes sense when passing is allowed.

As Tang Rules show, seki makes sense also without passes.
http://www.lifein19x19.com/forum/viewto ... =45&t=4005

Yes, but in capture go there is "zugzwang", you can not simply end the game by agreement. Or am I missing something?

[Bill Spight]

Yes, but in capture go there is "zugzwang", you can not simply end the game by agreement. Or am I missing something?

Actually, ending the game by agreement (instead of ending the game by passing) is an emergent feature of no pass go, as is the concept of territory.

[Bill Spight]

On seki in capture go and other forms of no pass go

There is more than one kind of seki, and the Japanese 1989 rules redefined seki in a strange way. However, unlike the Chinese term for mutual life, as far as I can tell, the traditional basic meaning of seki is a kind of standoff, where each player loses by making a local play if the opponent replies. We may have such a standoff in capture go, but, as life is not guaranteed in capture go, mutual life is not guaranteed, either.

[Bill Spight]

I have only taught one person using capture go, but the experience made me a fan. This problem illustrates a couple of things that I learned about capture go.

$$ Failure
$$ -----------
$$ | X . 2 . 1 |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------

Click Here To Show Diagram Code

[go]$$ Failure
$$ -----------
$$ | X . 2 . 1 |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------[/go]

is a mistake, as it allows to make a seki on the top side. Black's next play puts a stone in atari, and Black can resign.

My student found this kind of seki all by himself.

$$ Two eyes
$$ -----------
$$ | X . 1 3 . |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O 2 . O . |
$$ -----------

Click Here To Show Diagram Code

[go]$$ Two eyes
$$ -----------
$$ | X . 1 3 . |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O 2 . O . |
$$ -----------[/go]

Two eyes are not necessary for independent life in capture go, but in this kind of position Black must make a second eye to prevent the seki.

After White's next play puts his group in atari, and White can resign.

[Bill Spight]

Territory as an emergent property of capture go

When I first heard about capture go, I thought that it might engender the bad habit of just trying to capture stones instead of making territory. When I realized that territory is implicit in capture go, I changed my mind.

$$ Territory?
$$ -----------
$$ | X . . . . |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------

Click Here To Show Diagram Code

[go]$$ Territory?
$$ -----------
$$ | X . . . . |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------[/go]

Now, in regular go we count this as 4 points of territory for Black and 3 points for White. Black wins by 1 point. What is the concept of territory in capture go?

In this position note that a safe play by each player reduces the number of safe plays left for that player by one (assuming correct play). Therefore the players can count the number of safe plays left for each player and determine the result of the game without playing it out. That is, they can score the game.

Black has 2 safe plays left and White has 1, so the net score is 1 play for Black. Black wins by 1 play, or, IOW, by 1 point.

Note that each player has 2 fewer points in capture go than in regular go. That is because of the group tax. Each player needs 2 points for each group to keep it from being in atari.

Even though it is not in the rules, the players can agree to stop play when they reach a scorable position and count the territory. Because territory is an emergent property of capture go, ending the game by agreement is also an emergent property of capture go. (We can amend the rules to allow that possibility.) Pace Chen, it is not necessary to assume that the concept of territory arose as a means of equivalence scoring as a convenience for stone scoring. The concepts of territory and group tax emerge together from capture go.

$$ Evaluation?
$$ -----------
$$ | X . . . 1 |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------

Click Here To Show Diagram Code

[go]$$ Evaluation?
$$ -----------
$$ | X . . . 1 |
$$ | X X X X X |
$$ | X O X O X |
$$ | O O O O O |
$$ | O . . O . |
$$ -----------[/go]

How do we evaluate the position after ? White has 1 point (safe move) on the bottom side. On the top side each player has 1 safe move to a position worth 0. It is like a dame, with a count of 0. White is now 1 point ahead, and the correct play is in the "dame" on the top side to leave White still 1 point ahead.

[xed_over]

Even though it is not in the rules, the players can agree to stop play when they reach a scorable position and count the territory. Because territory is an emergent property of capture go, ending the game by agreement is also an emergent property of capture go. (We can amend the rules to allow that possibility.)

I believe its this emergent property that lead to the development of -- what we at the Seattle Go Center like to call -- Capture Plus

As far as I understand, the only difference between Capture Plus, and Capture Go is that once a scorable position is reached, we agree to stop play by passing.

[Joaz Banbeck]

I nominate this for the best thread of the year.

[Bill Spight]

Even though it is not in the rules, the players can agree to stop play when they reach a scorable position and count the territory. Because territory is an emergent property of capture go, ending the game by agreement is also an emergent property of capture go. (We can amend the rules to allow that possibility.)

I believe its this emergent property that lead to the development of -- what we at the Seattle Go Center like to call -- Capture Plus

As far as I understand, the only difference between Capture Plus, and Capture Go is that once a scorable position is reached, we agree to stop play by passing.

Interesting. I suppose that if you allow passes, you can have ties? Do you have a group tax, as well?

[Bill Spight]

I nominate this for the best thread of the year.

Gee, thanks, Joaz.

[xed_over]

Interesting. I suppose that if you allow passes, you can have ties? Do you have a group tax, as well?

ties? sure, why not? is it really necessary to have a winner? the whole point is to introduce basic concepts to beginners of how to capture, how to not get captured, and then, if no one gets captured, the one with the most open area wins.

group tax? that sounds overly complicated. because either someone gets captured, or we just count all the empty spaces (when the beginner finally realizes that its to their advantage to pass before filling in their own spaces)

I try move on to real go as soon as they start to begin to understand these basic concepts. no need to belabor it all with game theory

[gogameguru]

I try move on to real go as soon as they start to begin to understand these basic concepts. no need to belabor it all with game theory

This is my view as well. While all this stuff is really interesting to hardcore Go players like us, we have to remember other people might find it boring. One of the hardest things when teaching Go is to only say what's necessary and not go off on tangents. However, I suspect Bill's main interest is not in how to teach Go.

Bill, the reason I came back to this thread was because I've been thinking about this on and off for the last couple of weeks. I agree with Joaz, it's one of the most thought provoking threads I've seen here.

Are you inferring that Go evolved from the capture game and the emergent properties were formalised as rules later? It seems to me that that is where you were going, and it's a plausible hypothesis. Very difficult to prove of course, but it could make a fascinating historical study .

And by the way, I appreciate your didactic method.

David

[Bill Spight]

However, I suspect Bill's main interest is not in how to teach Go.

Actually, David, if I were to teach rank novices, this is the approach that I would take. Some people teach stone scoring right from the start, I would start with no pass capture go. Both have a group tax, but that can be dispensed with soon enough. (I might change my mind with experience, who knows?)

Are you inferring that Go evolved from the capture game and the emergent properties were formalised as rules later? It seems to me that that is where you were going, and it's a plausible hypothesis.

As it turns out, I came to this backwards. Some years ago I realized that territory is an emergent property of no pass go, and that no pass go with prisoner return is a form of territory go with a group tax. It was only after that that I found out about ancient go rules on SL ( http://senseis.xmp.net/?AncientChineseR ... Philosophy ), which apparently were for territory scoring with a group tax!

It is attractive to think that go began as no pass go with prisoner return, but that is no less complicated than territory scoring with a group tax. One peculiarity of territory scoring is why prisoners and dead stones are counted. I suspect that go may have evolved from a game where the object was to capture stones. Perhaps it was not no pass go where the inability to move means a loss, but maybe it was a game where each player made the same number of plays in the attempt to capture stones. (And perhaps instead of filling in the next to last eye of a group, you could simply hand a stone to your opponent as a captive.) That might explain why ancient records said that each player made the same number of moves. I think that the concept of territory could still emerge from such a game. Anyway, it is all speculative, as you say.