It is currently Tue Apr 25, 2017 11:29 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 42 posts ]  Go to page Previous  1, 2, 3  Next
Author Message
Offline
 Post subject: Re: An interesting new rule idea.
Post #21 Posted: Thu Dec 31, 2015 1:37 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
But do we want the dame to matter?

Consider this position.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O C C X .
$$ . O X X X .
$$ . O C X . .
$$ . O O X X X[/go]


All stones are alive. The only points of interest, including dame, are the marked points. Black to play.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O 1 . X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


This play seems obvious. In the marked region Black gets one point of territory while White gets the last move. The net local area score is also one point for Black.

Click Here To Show Diagram Code
[go]$$B Variation
$$ ------------
$$ . O 3 1 X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


In this variation the local territory result is 0, but the local area result is still one point for Black. There are those, myself included, who do not like the idea that Black can get away with :b1:.

Now, to use the language of Berlekamp and Wolfe's Mathematical Go: Chilling gets the Last Point, if there are no kos, we may consider territory scoring to be a chilled form of area scoring. (See http://senseis.xmp.net/?Chilling .) Correct play in the chilled form is also correct in the unchilled form, but not vice versa. (Note that ko-less Japanese go is not a chilled form of ko-less Chinese weiqi, because of the treatment of Japanese seki. Other forms of territory go, such as Lasker-Maas rules and Spight rules, are. :)) The condition of no kos is not as stringent as it might appear. If the ko is resolved early enough, then correct play under the chilled for will still be correct under the unchilled form. The chilled form is usually more stringent than the unchilled form, because there are plays that may be OK under the unchilled form that are mistakes under the chilled form. As this example shows. :) Berlekamp and Wolfe chill territory scoring, for an even more stringent form of go which they call chilled go. They do this by putting a one point tax on a board play.

Click Here To Show Diagram Code
[go]$$B Half point
$$ ------------
$$ . O . . X .
$$ . O X X X .
$$ . O . X . .
$$ . O O X X X[/go]


In chilled go neither player will wish to play, and the local region is scored as 0.5 point for Black.

_________________
"Drooling Banjos"

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #22 Posted: Thu Dec 31, 2015 1:53 pm 
Oza

Posts: 2457
Location: DC
Liked others: 145
Was liked: 435
Universal go server handle: skydyr
Online playing schedule: When my wife is out.
Bill Spight wrote:
But do we want the dame to matter?

Consider this position.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O C C X .
$$ . O X X X .
$$ . O C X . .
$$ . O O X X X[/go]


All stones are alive. The only points of interest, including dame, are the marked points. Black to play.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O 1 . X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


This play seems obvious. In the marked region Black gets one point of territory while White gets the last move. The net local area score is also one point for Black.

Click Here To Show Diagram Code
[go]$$B Variation
$$ ------------
$$ . O 3 1 X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


In this variation the local territory result is 0, but the local area result is still one point for Black. There are those, myself included, who do not like the idea that Black can get away with :b1:.


At the same time, though, black is clearly giving up sente with :b1: in the variation. Not only that, but white can treat :w2: and :b3: as miai and move on right away to another part of the board, if there are other points. Effectively black has taken, using area counting, 2 points in two moves rather than 2 points in one move.

Now, it may not matter if this part of the board is the last one to be played on, of course, but that's not to say it doesn't matter most of the time. To some degree, I think this displays a bias towards the result of territory scoring over area scoring. You could just as well come up with examples where black makes the same number of points via territory scoring for two different moves while one of the moves is clearly better than the other in terms of being able to play more or fewer dame later, or is able to do it in sente one way and gote another.

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #23 Posted: Thu Dec 31, 2015 2:46 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
skydyr wrote:
Bill Spight wrote:
But do we want the dame to matter?

Consider this position.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O C C X .
$$ . O X X X .
$$ . O C X . .
$$ . O O X X X[/go]


All stones are alive. The only points of interest, including dame, are the marked points. Black to play.

Click Here To Show Diagram Code
[go]$$B Black to play
$$ ------------
$$ . O 1 . X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


This play seems obvious. In the marked region Black gets one point of territory while White gets the last move. The net local area score is also one point for Black.

Click Here To Show Diagram Code
[go]$$B Variation
$$ ------------
$$ . O 3 1 X .
$$ . O X X X .
$$ . O 2 X . .
$$ . O O X X X[/go]


In this variation the local territory result is 0, but the local area result is still one point for Black. There are those, myself included, who do not like the idea that Black can get away with :b1:.


At the same time, though, black is clearly giving up sente with :b1: in the variation. Not only that, but white can treat :w2: and :b3: as miai and move on right away to another part of the board, if there are other points. Effectively black has taken, using area counting, 2 points in two moves rather than 2 points in one move.


That's why I said that the only points of interest were the marked points. I meant for the whole board. Guess that was not clear. :(

Quote:
Now, it may not matter if this part of the board is the last one to be played on, of course, but that's not to say it doesn't matter most of the time. To some degree, I think this displays a bias towards the result of territory scoring over area scoring.


Actually, I am a proponent of Button Go (or Double Button Go), both of which integrate area and territory scoring. :D

Quote:
You could just as well come up with examples where black makes the same number of points via territory scoring for two different moves while one of the moves is clearly better than the other in terms of being able to play more or fewer dame later, or is able to do it in sente one way and gote another.


Not sure what you are proposing. Without ko, there are no sequences of play that are incorrect by area scoring but correct by Lasker-Maas or Spight territory rules.

_________________
"Drooling Banjos"

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #24 Posted: Thu Dec 31, 2015 4:52 pm 
Lives with ko

Posts: 176
Liked others: 0
Was liked: 30
Rank: kgs 6 kyu
KGS: Pio2001
Joelnelsonb wrote:
Why on earth is this not the standard way to play Go?


Using stone scoring, the games clearly divide themselves into two parts : the strategic part, where each move can change the outcome, and the filling part, that starts when modern players pass, and where the moves have no meaning anymore.
Playing out the second part is so long, so boring and so meaningless that it was decided that the game would end when the first part is over. Each player is then awarded the number of intersections that he or she conquered. This is area scoring.
Then, territory scoring is a convenient trick allowing not to count the stones.

Thus, for all practical purposes, to count the territory is the most convenient way to play go everyday, as far as players are concerned. This is why it has been the most widespread method.

Why it didn't become the universal method is a long story.

To begin with, the first official complete ruleset published (the japanese rule of 1949) tried to conciliate territory scoring with the traditions in use, such as the filling of teire points (implicit connections after the filling of the dame), and to regroup all the precedents concerning unsolved ko after the two players have passed.
Unfortunately the result was not a success. The rule, dozens of pages long, is much too long to be used for teaching and requires very tedious studying from tournament organizers. Today, it is nearly impossible to program it into a software, and in 1959, it could not prevent an undecided game to happen between Go Seigen and Takagawa Shukaku.
A striking consequence is that today, books introducing the game of go usually don't feature the rule of the game anywhere !

In 1975, the Chinese came with another official rule, using area scoring, this time. Much shorter, it can easily be taught to beginners, it is convenient for tournament organizers and programmers, and it can leave no game undecided. But for everyday use, it is much less convenient : the players must count all the stones.

In 1991, the AGA rule combined the simplicity of the chinese rule with the convenience of territory counting, in exchange for just a little effort from the players : to hand a prisoner each time thay pass.
This progress seem to seduce, and spreads slowly : a similar rule was adopted in 1994 in France, then in 2008 in United Kingdom, and this year it was used in the European Go Congress.

When will we have a unique worldwide rule of go ?

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #25 Posted: Thu Dec 31, 2015 8:21 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
Pio2001 wrote:
To begin with, the first official complete ruleset published (the japanese rule of 1949) tried to conciliate territory scoring with the traditions in use, such as the filling of teire points (implicit connections after the filling of the dame), and to regroup all the precedents concerning unsolved ko after the two players have passed.
Unfortunately the result was not a success. The rule, dozens of pages long, is much too long to be used for teaching and requires very tedious studying from tournament organizers. Today, it is nearly impossible to program it into a software, and in 1959, it could not prevent an undecided game to happen between Go Seigen and Takagawa Shukaku.


The Nihon Kiin 1989 rules are too ambiguous to be programmed, IMO.

As for the rules dispute between Go Seigen and Takagawa, at the time Go was not a member of the Nihon Kiin and had not agreed to use those rules, as Takagawa had assumed. The 1949 rules were clear that Go had to fill the final ko in the game continuation that was not played out, as both players had read it out. It was the failure to agree on the rules before the game that created the problem.

Edit: Some reports of that dispute are misleading. They say that the rules would have forced Go Seigen to protect against a ko that he would have won. Yes, he would have won it, but the point was that he would have had to fill it. The extra move to fill it would have come to the same thing as making a move to prevent it. Go Seigen argued that he would not have had to fill it, because there would have been no dame left at that point.

_________________
"Drooling Banjos"


Last edited by Bill Spight on Fri Jan 01, 2016 11:42 pm, edited 1 time in total.
Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #26 Posted: Fri Jan 01, 2016 11:39 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
Joelnelsonb wrote:
So this may have been thought of and talked about already but I have never heard of it. I suggest that a player ought to lose 2 points for each independent group he/she has on the board.


Indeed, it is an interesting idea, and congratulations for having come up with it on your own. :) As you have discovered, it is an old idea, but that does not diminish your accomplishment.

I want to thank you because I had always associated the group tax with stone scoring without giving it much thought. I knew about its association with no pass go and territory scoring long before I learned that ancient territory scoring had a group tax. Your question made me examine why stone scoring had a group tax. It is a mystery. ;)

_________________
"Drooling Banjos"

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #27 Posted: Sat Jan 02, 2016 3:09 pm 
Lives in gote
User avatar

Posts: 385
Liked others: 13
Was liked: 23
OGS: Saint Ravitt
Thank you, Bill :salute:

As for my question about why we don't commonly see this in modern Go: I'm not so much asking why it is that people prefer to play a certain way but rather I was wondering if there is another theoretical explanation for the core objective of the game that makes group tax unnecessary or even not make sense.

_________________
Thinking like a go player during a game of chess is like bringing a knife to a gun-fight. Thinking like a chess player during a game of go feels like getting knifed while you're holding a gun...

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #28 Posted: Sat Jan 02, 2016 3:40 pm 
Oza
User avatar

Posts: 2694
Location: Seattle, WA
Liked others: 247
Was liked: 525
KGS: oren
Tygem: oren740, orenl
IGS: oren
Wbaduk: oren
Joelnelsonb wrote:
As for my question about why we don't commonly see this in modern Go: I'm not so much asking why it is that people prefer to play a certain way but rather I was wondering if there is another theoretical explanation for the core objective of the game that makes group tax unnecessary or even not make sense.


It makes scoring more complicated. It's much easier to count territory after filling prisoners and leave it at that.

We get enough tax in the rest of the world. We don't need group taxes :)

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #29 Posted: Sun Jan 03, 2016 9:56 am 
Lives with ko

Posts: 176
Liked others: 0
Was liked: 30
Rank: kgs 6 kyu
KGS: Pio2001
Joelnelsonb wrote:
I was wondering if there is another theoretical explanation for the core objective of the game that makes group tax unnecessary or even not make sense.


Rules are arbitrary. They don't need justification. Which ones are better is all a matter of taste, convenience, beauty etc.

Go is a game about shapes. Having area as a goal is also a matter of shape.
Having stones as a goal would also be a matter of numbers. I prefer shapes to numbers.

I don't like territory counting because of the "prisoner tax". I always think about my score in area counting, even when I play under japanese rules. My points appear as group of intersections laying on the goban. It's purely visual.

I must admit that it causes me some trouble when I practice exercises about yose : the "area" swing value of a given move can be very different from its "territory" swing value, because the number of black and white stones played locally is not always the same.

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #30 Posted: Sun Jan 03, 2016 11:14 am 
Dies with sente

Posts: 94
Liked others: 6
Was liked: 4
I think that one of the main problems with group tax is defining what a group is. "A group" may seem like a simple concept, but it is actually quite hard to define it precisely. You might want to post a good definition of a group before going further with your group tax idea.

_________________
“Discussion is an exchange of knowledge; argument an exchange of ignorance.” ― Robert Quillen

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #31 Posted: Sun Jan 03, 2016 11:19 am 
Lives in gote
User avatar

Posts: 385
Liked others: 13
Was liked: 23
OGS: Saint Ravitt
An "independent group" is a group of connected stones, all sharing liberties which require two separate eyes in order to permanently maintain their place on the board. Compare this to a group left in seki which is not independent but depends and shares liberties with an enemy group. In Japan they call it seki, in America we call it a Mexican Standoff :D

_________________
Thinking like a go player during a game of chess is like bringing a knife to a gun-fight. Thinking like a chess player during a game of go feels like getting knifed while you're holding a gun...

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #32 Posted: Sun Jan 03, 2016 11:46 am 
Tengen

Posts: 4357
Liked others: 0
Was liked: 574
Pio2001 wrote:
for all practical purposes, to count the territory is the most convenient way to play go everyday


Then why does nobody "count the territory" but are there rulesets that count "territory plus opposing prisoners"? ;)

Convenient - no.

Quote:
Rules are arbitrary. They don't need justification. Which ones are better is all a matter of taste, convenience, beauty etc.


Rules in general are not arbitrary, but there are explanations and justifications for them. For particular rules, there is a choice among equally valid alternatives; for every such choice, it is a matter of external requirements or, if none are presumed, personal preferance.

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #33 Posted: Sun Jan 03, 2016 1:29 pm 
Dies with sente

Posts: 94
Liked others: 6
Was liked: 4
Joelnelsonb wrote:
An "independent group" is a group of connected stones, all sharing liberties which require two separate eyes in order to permanently maintain their place on the board. Compare this to a group left in seki which is not independent but depends and shares liberties with an enemy group. In Japan they call it seki, in America we call it a Mexican Standoff :D

May I ask you to identify individual groups and the final score in the following (made up) game:



It would seem to me that under your definition the single stone in the top left corner is a group on it's own, since it is not connected to any other stones. Yet, I would usually consider it a part of the independently alive black group.

_________________
“Discussion is an exchange of knowledge; argument an exchange of ignorance.” ― Robert Quillen

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #34 Posted: Sun Jan 03, 2016 1:34 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
tiger314 wrote:
I think that one of the main problems with group tax is defining what a group is. "A group" may seem like a simple concept, but it is actually quite hard to define it precisely. You might want to post a good definition of a group before going further with your group tax idea.


Well, a "group tax" is not a tax, maybe it does not apply to groups, either. ;) Language is funny.

As for a precise definition, there is an operational definition by play. Play reaches a point where there is one or more one point eyes which, if and only if filling one of them allows the opponent to take the stones forming those eyes on the next move. The number of such eyes is the group tax for those stones.

You can also use hypothetical play, which is usually how the group tax is determined. :)

Edit: Oops! There are eyes that, if filled, do not allow the opponent to take on the next move, but the stones that formed that eye are still dead. That is the test, isn't it? Filling an eye kills your own stones.

_________________
"Drooling Banjos"


Last edited by Bill Spight on Sun Jan 03, 2016 4:17 pm, edited 2 times in total.
Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #35 Posted: Sun Jan 03, 2016 1:50 pm 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
tiger314 wrote:
May I ask you to identify individual groups and the final score in the following (made up) game:



It would seem to me that under your definition the single stone in the top left corner is a group on it's own, since it is not connected to any other stones. Yet, I would usually consider it a part of the independently alive black group.


Under any number of variants of No Pass Go the final net score is 1 point for White. Given the net score and the number of points of territory, you can figure out the group tax. :)

_________________
"Drooling Banjos"

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #36 Posted: Sun Jan 03, 2016 11:38 pm 
Lives in gote
User avatar

Posts: 385
Liked others: 13
Was liked: 23
OGS: Saint Ravitt
tiger314 wrote:
May I ask you to identify individual groups and the final score in the following (made up) game:



It would seem to me that under your definition the single stone in the top left corner is a group on it's own, since it is not connected to any other stones. Yet, I would usually consider it a part of the independently alive black group.


The score is white wins by 1 (given that both have 35 as is yet white can still fill in one more eye to get a 36th stone). That would of course be under strict stone scoring without any additional compensation points. It's a weird example, I admit. However, I would consider that stone in the top left to be sharing liberties with it's surrounding group given that its laying on the neutral fringe of the board (which in this case is treated like a wall of black stones). BUT, if you don't like that answer then that just means that we don't get to have a nice, pretty little concise statement to confine the group tax. You must simply say: A player must deduct one point for every necessary eye he has on the board (oh wait, that's also rather concise, isn't it...). Notice as in the example that white still has an unnecessary eye which is why he wins by my count. It's important to remember though that the idea of this thing were calling "group tax" is in no way a rule or even a subrule of the game. It's simply an aspect of the learning to count up the score without having to completely fill in the board as I did in the example I gave before. Compare this to life and death. No where in the rule book of Go should the terms "life and death" even show up. These are simply jargon that we use between players to describe a certain type of position. Instead of playing it out to completion, two skilled players can look and say "if we DID play this out, those stones would be captured without any chance of survival. Let's not waste our time. And for future reference, when we encounter stones like these, lets just call them "dead" so we both know what we mean. agreed?" The same is true for miai, seki, atari, along with kosumi, kakari, nobi and so forth. The point being that striving to write up a perfect definition that can be universally among players is a little pointless anyways. Just as long as were all on the same page and when the games over, we both know who won.

_________________
Thinking like a go player during a game of chess is like bringing a knife to a gun-fight. Thinking like a chess player during a game of go feels like getting knifed while you're holding a gun...


This post by Joelnelsonb was liked by: Bill Spight
Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #37 Posted: Mon Jan 04, 2016 10:52 am 
Lives with ko

Posts: 176
Liked others: 0
Was liked: 30
Rank: kgs 6 kyu
KGS: Pio2001
Joelnelsonb wrote:
Compare this to life and death. No where in the rule book of Go should the terms "life and death" even show up.


This is true for stone or area scoring.
But territory scoring, as used in japanese rules, must rely on a strict definition of life and death. Under japanese rules, removing dead stones without paying the price of their capture is a right given by the rule. In area or stone scoring, it is a conventional practice not depending on the rule.

It is funny to think that a strict territory-minus-prisoner scoring would have nothing to do with the game of go as we know it, because every so-called "dead" stone would have to be actually captured for the eyespace to become one player's territory.
In fact, when we say territory-minus-prisoners, we actually mean territory-minus-prisoners-minus-dead-stones, where "dead" is supposed to be defined somewhere in the complete version of the rule.
Robert has found a very good definition of "life and death" : http://home.snafu.de/jasiek/sj.html

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #38 Posted: Mon Jan 04, 2016 11:03 am 
Tengen

Posts: 4357
Liked others: 0
Was liked: 574
Robert Pauli suggested a predecessor of the definition.

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #39 Posted: Mon Jan 04, 2016 11:06 am 
Judan

Posts: 6025
Liked others: 1403
Was liked: 2315
Joelnelsonb wrote:
It's important to remember though that the idea of this thing were calling "group tax" is in no way a rule or even a subrule of the game. It's simply an aspect of the learning to count up the score without having to completely fill in the board as I did in the example I gave before.


Very important point. It underscores the fact that the so-called "group tax" logically depends upon the score, not the other way around. :) Yes, it is a shortcut to count the score, but unnecessary.

_________________
"Drooling Banjos"

Top
 Profile  
 
Offline
 Post subject: Re: An interesting new rule idea.
Post #40 Posted: Mon Jan 04, 2016 11:50 am 
Oza

Posts: 3957
Location: North Carolina
Liked others: 390
Was liked: 649
Rank: AGA 3k
GD Posts: 65
OGS: Hyperpape 4k
Pio2001 wrote:
Rules are arbitrary. They don't need justification. Which ones are better is all a matter of [several forms of justification].

_________________
Occupy Babel!

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 42 posts ]  Go to page Previous  1, 2, 3  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 2 guests


You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot post attachments in this forum

Search for:
Jump to:  
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group