Life In 19x19
http://www.lifein19x19.com/

dead stones and japanese scoring
http://www.lifein19x19.com/viewtopic.php?f=11&t=6373		 Page 1 of 1
Author:  Ma2ReeD [ Tue Jul 24, 2012 2:47 pm ]
Post subject:  dead stones and japanese scoring
I'm confused.

I was looking at the Japanese scoring rules(http://www.online-go.com/faq.php?name=rules) and the example board leaves me wondering.

Image

if black placed stones at places like G1, J3, C9, etc. where it would be in white's territory, but still have >1 liberty, it would cost white two territory to capture the black stone, and would end up with one less point after the entire exchange is over. Why is this considered a finished game, when black (or white) could both fight for a few extra point? If there were to be a dead (black) stone at J3, why would it only count for 1 point, when white would have to lose 3 territory to capture it? Would black not pass until his dead stone at J3 were captured, costing white 2 points?
Author:  palapiku [ Tue Jul 24, 2012 2:56 pm ]
Post subject:  Re: dead stones and japanese scoring
Suppose there's a black stone at G1. Then white doesn't have to capture it because it's already dead. If black disagrees, and claims that it's not dead, they can play it out, in which case White does have to spend two moves to capture it.

But - and this is a very important part of the Japanese rules - once they play it out and prove that the stone at G1 really was dead, the game is reverted to the original position after the two passes. So White gets his two points back and G1 is still dead.

(Note that if there's a black stone on G1, White actually has to play F3 or lose a point if Black plays there next. But G1 is still dead, and the exchange G1-F3 still loses a point for Black, so Black wouldn't play G1. This is unrelated to your actual question of having to capture dead stones.)
Author:  Bill Spight [ Tue Jul 24, 2012 3:15 pm ]
Post subject:  Re: dead stones and japanese scoring
Ma2ReeD wrote:
I'm confused.

I was looking at the Japanese scoring rules(http://www.online-go.com/faq.php?name=rules) and the example board leaves me wondering.

Image

if black placed stones at places like G1, J3, C9, etc. where it would be in white's territory, but still have >1 liberty, it would cost white two territory to capture the black stone,


Yes, it would cost points to capture dead stones. :)

Quote:
and would end up with one less point after the entire exchange is over.


But White does not have to capture a dead stone. At the end of the game, after the dame are filled, White simply removes dead Black stones from the board. (The rules for pros are slightly different.)

Quote:
Why is this considered a finished game, when black (or white) could both fight for a few extra point?


They cannot fight for those points. At least, they should not.

Quote:
If there were to be a dead (black) stone at J3, why would it only count for 1 point, when white would have to lose 3 territory to capture it? Would black not pass until his dead stone at J3 were captured, costing white 2 points?


The rule is that dead stones are removed as prisoners after play is over. You are already able to see that those stones are dead. What is the problem? :)

Now, it is true that sometimes players make mistakes about whether stones are alive or dead. If you or your opponents are sometimes unsure about whether stones are dead or alive, why not play by area scoring, and play on until the only stones left on the board are alive? :) AGA rules use area scoring.
Author:  Mike Novack [ Wed Jul 25, 2012 6:35 pm ]
Post subject:  Re: dead stones and japanese scoring
I think I understand what is causing the confusion.

Some configurations of stones are unconditionally alive (cannot be killed) even if surrounded. Stones that are surrounded may also be alive because the surrounding group lacks external dame and so they cannot be killed (seki). Otherwise surrounded stones are dead.

But what about when there is disagreement? The mistake here is in placing the onus of proof on the wrong party. In the example given it isn't up to the white player to prove the stone can be captured (by playing two stones while black passes). The White player simply repeats "you're dead" until/unless the black player can place stones to achieve a live group or seki with the white player either responding by playing stones or passing. In which case white will not have played more stones inside than black.
Author:  EdLee [ Thu Jul 26, 2012 5:59 pm ]
Post subject: 
Ma2ReeD, let's play it out step and step so we can see the situation exactly.
If B plays :b1: at G1 in your example, W connects at :w2:.
If B passes now, W will also pass -- W claims :b1: is dead so B just gave W an extra point with :b1:.
If you are B and you disagree with W here -- you have to prove :b1: is not dead -- what is your move now?
Click Here To Show Diagram Code
[go]$$Bc
$$ ---------------------
$$ | . . . . O O X . . |
$$ | . O . O O X . X X |
$$ | . . O X X X , X O |
$$ | O O O O X . . X O |
$$ | O X O X X X X O O |
$$ | X X X . X O . O O |
$$ | . . , . X 2 O O . |
$$ | . . X . X O . O . |
$$ | . . . X O O 1 . . |
$$ ---------------------[/go]
Author:  lightvector [ Fri Jul 27, 2012 10:53 pm ]
Post subject:  Re: dead stones and japanese scoring
Ma2ReeD, the short answer is that in rulesets that use territory scoring, players are simply not required to capture dead stones and lose points for doing so. If you prefer, you can just take this as an axiom and consider it true by fiat.

Basically, for one reason or another, people first decided that they wanted this property to be true and only then wrote the rules so that it was true, rather than the other way around. But different rulesets in different places all chose their own ways to write the rules to make this true, so despite all of them having this property in common, the actual precise ways that disputes and scoring are handled actually differ quite a bit, and can potentially be very complicated. So complicated, in fact, that many online go servers don't bother trying and don't provide a way to resolve disputes. Fortunately, in practice, in almost every game between players with at least a little experience, disputes over what is dead or not don't happen.

---

Now, to quibble with some of the other posters here: In my opinion, a ruleset that puts the burden of proof on the player claiming a group is alive is a bit perverse.
Edlee, Mike, what is the actual statement of the dispute procedure that you are advocating here? Just as an example, something like:

"If the players disagree on whether one or more groups are alive or dead at the end of the game, play continues until:
1. Agreement is reached, in which case the board is scored appropriately,
2. Or all disputed groups have two actual eyes or are recognized by an experienced player to be a 'seki', in which case they are considered alive and the board is scored appropriately.
3. Or both players refuse to continue (by passing twice again and then refusing to resume play again) yet still do not agree, in which case all disputed groups that don't have two eyes and are not in seki are considered dead and the board is scored appropriately."

...might be problematic because it doesn't give a way to resolve some disputes without an experienced player (it's nice if a board game has precise rules that don't leave things up to interpretation), and because there's nothing in this example procedure to stop a player from disputing the status of a group that's alive but doesn't physically have two eyes, in order to force the other player either to play (and lose points filling their own territory) or else have the group scored as dead. Although just technicalities, I don't see a trivial way to fix these two issues in a ruleset that assumes that a disputed group is dead until proven alive.

Rather, I was under the impression that many (most?) territory rulesets place the burden of proof on the player claiming the group is dead to actually capture the group, except that for scoring purposes, these additional moves don't lose points (the moves are undone and the original position is scored, or the player is compensated for making additional moves, etc).
Author:  Mike Novack [ Sat Jul 28, 2012 4:21 am ]
Post subject:  Re: dead stones and japanese scoring
lightvector wrote:

Now, to quibble with some of the other posters here: In my opinion, a ruleset that puts the burden of proof on the player claiming a group is alive is a bit perverse.
Edlee, Mike, what is the actual statement of the dispute procedure that you are advocating here? Just as an example, something like:

"If the players disagree on whether one or more groups are alive or dead at the end of the game, play continues until:
1. Agreement is reached, in which case the board is scored appropriately,
2. Or all disputed groups have two actual eyes or are recognized by an experienced player to be a 'seki', in which case they are considered alive and the board is scored appropriately.
3. Or both players refuse to continue (by passing twice again and then refusing to resume play again) yet still do not agree, in which case all disputed groups that don't have two eyes and are not in seki are considered dead and the board is scored appropriately."


What I meant was that there are minimal configurations necessary for a live group, for a seki. etc. It is apparently your contention that because the rule sets do not bother to specify what these are that the matter is totally undefined. The way the game is played is that at the end (by agreement, no more useful moves) "dead stones" are removed. The player with these "dead stones" need not agree that they are dead, may insist on playing on in an attempt to achieve one of these minimal configurations (or to kill the surrounding group)

You gave an example of one stone played inside? Which of these has been demonstrated?
a) The remaining space (including this stone) represents a "dead shape" for the surrounding group, lacking a separate eye, so the surrounding group will die. The rules do not bother specifying which shapes live or die but a little experience playing go teaches you them.
b) The remaining space between is two dame and either the surrounding group has no separate eye or both do and "a" doesn't apply. That's the condition for seki, again a consequence of the rules.
c) The surrounded group* can achieve two separate eyes which is the condition for life. Again a consequence of the rules. You might want to investigate the minimum number of stones required for such a group.

With one stone "a" might apply but "b" or "c" are impossible**. However note that in the case of "a" the player isn't going to be claiming "I'm alive" but "you are dead". Beginning players may need to play out to the bitter end situations in dispute (and might make a crucial mistake in doing so) but with a little experience you'll know "alive", "seki", "dead".


* Need not be a totally connected group as long as each of its connected pieces touches two disjoint eyes.

** Do you see why seki is impossible?
Author:  EdLee [ Sat Jul 28, 2012 4:34 am ]
Post subject: 
lightvector, this: :)
lightvector wrote:
Fortunately, in practice, in almost every game between players with at least a little experience, disputes over what is dead or not don't happen.
Page 1 of 1 All times are UTC - 8 hours [ DST ]
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group
http://www.phpbb.com/