It is currently Mon Mar 18, 2024 8:34 pm

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 146 posts ]  Go to page Previous  1 ... 4, 5, 6, 7, 8  Next
Author Message
Offline
 Post subject: Re: GT territory rule
Post #121 Posted: Mon Sep 06, 2021 3:46 pm 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Cassandra wrote:
Gérard TAILLE wrote:
Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . O . X W . O X . . .
$$ | O X X X W X O X . . .
$$ | . X W W W X O X . . .
$$ | X X W O . X O X . . .
$$ | W W W X O O O X . . .
$$ | X X X X X X X X . . .
$$ | . . . . . . . . . . .
$$ | . . . . . . . . . . .
$$ | . . . . . . . . . . .
$$ | . . . . . . . . . . .
$$ | . . . . . . . . . . .
$$ +----------------------[/go]
Strict J89 application says that the two black groups are dead. Isn't it an unexpected result which goes against the go logic?

Quite apparently, traditional Japanese understanding includes the general principle that there are NO ko-threats in a (temporary) seki during status confirmation.
Which also includes double-ko formations as a special case dealt with in Examples 16-18.

SUPPLEMENT:

Probably...

Commentary on Article 7, Life and death, Clause 2
...
Even if a player has unlimited ko threats (in a double-ko seki for example), he cannot use them to recapture a ko.


... has been underrated so far. Or it wasn't taken seriously enough all the decades.
Please note that Japanese language does not know explicite plural!

If we leave the mentioned example aside, that comment reads:

"Even if a player has a ko threat, he cannot use it to re-capture a ko."

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #122 Posted: Tue Sep 07, 2021 7:08 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
The problem with the examples chosen is that all ko fights illustrated are built with a local ko and ko threats which could be more or less independant of the local area => we cannot easily guess how to handle a local move following a ko capture.
A lot of positions may be built OC. Here is another one:

Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | X O . O O O X . X O . . . |
$$ | . X . X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]


In normal play, if is white to play it could follow
Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | X O . O O O X . X O . . . |
$$ | 1 X 2 X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]
and the game is finished with all the top of the board being black territory.
As a consequence, instead of playing :w1: in normal play, white simply passes.

Now it is black to play and we are always in normal play. Do you feel logical to add a move before stopping the game?

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #123 Posted: Tue Sep 07, 2021 7:44 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Gérard TAILLE wrote:
The problem with the examples chosen is that all ko fights illustrated are built with a local ko and ko threats which could be more or less independant of the local area => we cannot easily guess how to handle a local move following a ko capture.
A lot of positions may be built OC. Here is another one:

Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | X O . O O O X . X O . . . |
$$ | . X . X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]


In normal play, if is white to play it could follow
Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | X O . O O O X . X O . . . |
$$ | 1 X 2 X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]
and the game is finished with all the top of the board being black territory.
As a consequence, instead of playing :w1: in normal play, white simply passes.

Now it is black to play and we are always in normal play. Do you feel logical to add a move before stopping the game?


Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | . O . O O O X . X O . . . |
$$ | O X X X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]

Under J89, there is NO teire needed in your final position.

However, this is another example of the kind of positions that you seem to like so very much:
A compound with two ko-shapes that allows an unlimited number of ko threats.

In a "normal" game, Black would have captured White's five stones at a suitable moment to get rid of this source of White ko-threats.

Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | X O . O O O X . X O . . . |
$$ | . X X X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]

A teire would be necessary in your starting position, but which follows Japanese tradition with this topic. A lot of rule discussions / disputes before J89 were about such teire.

Click Here To Show Diagram Code
[go]$$W
$$ +---------------------------+
$$ | M O . O O O X . X O . . . |
$$ | 1 X 2 X X O O X X O . . . |
$$ | X X X O X X X X O O . . . |
$$ | O O O O O O O O O O . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |
$$ | . . . . . . . . . . . . . |[/go]

However, Black could have circumnavigated the status confirmation cliff, if he had answered White's atari by connecting at :b2:, instead of capturing with :bx: .
But then, the unlimited-number-of-ko-threats issue would come out from around the corner.

A teire would be practically necessary in both cases, so there is no cause for grief.

However, the alternative you choose (Black captured in the corner) would let the pressure out of the boiler during the "hot" phase of the game. Thereafter, the required teire can be added in peace and quiet during the occupation of the DAME.


++++++++++++++++++++++++++++++++

By the way, in principle all your examples of this family are similar to the following position:

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

Would you assess it suitable for Black to add a move before he declares the game to have ended?

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #124 Posted: Tue Sep 07, 2021 9:38 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Meaningless reinforcement moves are incorrect compared to normal go, but from the Japanese view there is some logic behind them.

It should be emphasized and appreciated that both the Korean and Japanese approach has a coherent theory behind the handling of ALL positions. The same for the third potential approach of "only moonshine kos disallowed, otherwise normal play decides".

Your combination, on the other hand, makes moonshine dead both with nearby and remote double ko, while bent4 life differs in nearby and in remote cases. It seems hard to find convincing theory/explanation for that.

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #125 Posted: Tue Sep 07, 2021 1:54 pm 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Meaningless reinforcement moves are incorrect compared to normal go, but from the Japanese view there is some logic behind them.

It should be emphasized and appreciated that both the Korean and Japanese approach has a coherent theory behind the handling of ALL positions. The same for the third potential approach of "only moonshine kos disallowed, otherwise normal play decides".

Your combination, on the other hand, makes moonshine dead both with nearby and remote double ko, while bent4 life differs in nearby and in remote cases. It seems hard to find convincing theory/explanation for that.


Let's look at a position with one local and a double ko nearby:
Click Here To Show Diagram Code
[go]$$W
$$ +-------------------+
$$ | X O . O O O O O X |
$$ | . X . X X X X X . |
$$ | X X X O X X O X X |
$$ | O O O O O O . O O |
$$ | . . . . . O O O . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ +-------------------+[/go]

Click Here To Show Diagram Code
[go]$$W
$$ +-------------------+
$$ | X O . O O O O O X |
$$ | 1 X 2 X X X X X . |
$$ | X X X O X X O X X |
$$ | O O O O O O . O O |
$$ | . . . . . O O O . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ +-------------------+[/go]

If in normal play white plays first then after the exchange :w1: :b2: all white stones at the top are dead either in normal play or in confirmation phase.
If in normal play black plays first then obviously all white stones at the top are dead.
In such situation we could expect that all white stones are dead if the game stops and the confirmation phase takes place.

But it follows in confirmation phase:
Click Here To Show Diagram Code
[go]$$W
$$ +-------------------+
$$ | X O 5 O O O O O X |
$$ | 1 X 2 X X X X X 3 |
$$ | X X X O X X O X X |
$$ | O O O O O O 4 O O |
$$ | . . . . . O O O . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ | . . . . . . . . . |
$$ +-------------------+[/go]
and all BLACK stones are dead => anti seki.
Instead of :b2: black can at :w5: but the result is not changed, black cannot kill all white stones.

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #126 Posted: Tue Sep 07, 2021 2:11 pm 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Was this a response to what I wrote? I talked about your rules (combination), and bent4 handling difference to moonshine ko with double ko (seki). This position is a variant of the previous one with the same reinforcing problem for Japanese (also no double ko and no moonshine ko here).

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #127 Posted: Wed Sep 08, 2021 1:33 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Gérard, rather than continuing to travel to secondary theaters of war where you cannot succeed, it would be far more convenient for you to return to GT City.
And back to its Roaring Twenties, that is, to the era before the earthquake you caused, when people were still raving about the beneficial effects of loops and cycles.

Let me explain what you overlooked in your desperate -- but uncessessful -- attempt to "simplify" the application of your rules...

--------------------------

Prohiting an attacker's (Black below) ko-capture before a defender's (White below) pass is suitable for stopping the endless repetition of a cycle

:black: :white: :black: :white: :b1: :wx: _._ :bc: :wc: :bc: :wc: :b1: :wx: _._ :bc: ...

:black: :white: = GENUINE moves
:b1: = ko-capture
:wx: = pass
_._ = cycle separator (for easier understanding)
:bc: :wc: = repeated GENUINE moves
... = ad infinitum


... that would prevent the defender from approaching attacker's stones to be captured.

--------------------------

Is is also suitable for stopping the endless repetition of the following cycle with THREE ko-shapes (one GENUINE DOUBLE-ko included; #2, #3)

:w1: :b2: :w3: :b1: :wx: :bx: :w1: :b3: :w2: :b1: :wx: :bx: _._ :w1: :b2: :w3: :b1: :wx: :bx: :w1: :b3: :w2: :b1: :wx: :bx: _._ :w1: ...

:w1: :b1: , :w2: :b2: , :w3: :b3: = ko-captures in the respective ko-shape

... that would also prevent the defender from approaching attacker's stones to be captured.


At first sight, it seems as if that cycle above was initiated by the defender, but a deeper look will reveal that it really was the ATTACKER, who started it.
This is because :b2: (the first ko-capture into the DOUBLE-ko) is NOT enforced by :w1: (a ko-capture into the other ko-shape). Thus, we can re-write the cycle a bit to make this more evident:

( :w1: _._ ) :b2: :w3: :b1: :wx: :bx: :w1: :b3: :w2: :b1: :wx: :bx: :w1: _._ :b2: :w3: :b1: :wx: :bx: :w1: :b3: :w2: :b1: :wx: :bx: :w1: _._ :b2: ...

You will recognise that there are TWO :b1: :wx: in the cycle, one in the middle, and the other one near the cycle's end.
Your abandonment of loops has resulted in the possibibily of banning :b1: as soon as after its VERY FIRST appearance (and not as late as after the completition of the cycle). What does not hurt here, as #1 is a SINGLE-ko-shape.

--------------------------

However, you overlooked the DEFENDER's option of utilising a DOUBLE-ko for prolonging the lifespan of an attached group of her that is caught in a semeai:

:black: :w1: :b2: :wx: :black: :w2: :b1: :wx: :black: ( :white: :black: ) :w1: :b2: :wx: :black: :w2: ...

Please note that there are NO :bc: :wc: included, as a GENUINE semeai, which ends with the death of one colour's stones, is NEVER a cycle.
Please also note that it is the DEFENDER here, who starts EVERY pair of DOUBLE-ko captures.


Prohibiting :b2: will result in

:black: :w1: :b2: :wx: :black: :w2: :b1: :wx: :black: ( :white: :black: ) :w1: :black: :white:

After the second :w1: (which is an atari) the attacker can only tenuki (best for playing another move in the attached semeai), whereafter the defender's final :white: captures the attacker's DOUBLE-ko group. Thus, the defender has WON a semeai that she should have LOST FOR SURE.
This means that any attached group that has MORE THAN FOUR liberties (MORE THAN THREE in the case of a nested DOUBLE-ko shape for the attacker, which would result in the survival of the defender's DOUBLE-ko group at least) can NEVER be captured (with simultaneously frustrating a defender's "enabled" stone in the "nested" case).

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #128 Posted: Wed Sep 08, 2021 7:41 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Was this a response to what I wrote? I talked about your rules (combination), and bent4 handling difference to moonshine ko with double ko (seki). This position is a variant of the previous one with the same reinforcing problem for Japanese (also no double ko and no moonshine ko here).


Not really Yann. I am wondering what do you mean by "combination". I am a little lost here. Maybe a diagram will help me to understand.

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #129 Posted: Wed Sep 08, 2021 7:49 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Gérard TAILLE wrote:
I am wondering what do you mean by "combination". I am a little lost here.

I referred to your proposed rules, which I think are mostly a COMBINATION of Korean localization and your moonshine ko rule.

And what I tried to say was this: while we can agree the unnecessary reinforcement moves of J89 are incorrect (compared to normal go play - maybe even ridiculous), they are natural consequence of the ko-fight-less L/D concept. So you can criticize the theory but not the consequence (which is correct for the theory).

On the other hand your rules produce the mentioned slight inconsistency, for which it seems hard to find a convincing theory/explanation (and which is also incorrect compared to normal go play btw).

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #130 Posted: Wed Sep 08, 2021 8:43 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Gérard TAILLE wrote:
I am wondering what do you mean by "combination". I am a little lost here.

I referred to your proposed rules, which I think is mostly a COMBINATION of Korean localization and your moonshine ko rule.

And what I tried to say was this: while we can agree the unnecessary reinforcement moves of J89 are incorrect (compared to normal go play), they are natural consequence of the ko-fight-less L/D concept. So you can criticize the theory but not the consequence.

On the other hand your rules produce the mentioned slight inconsistency, for which it seems hard to find a convincing theory/explanation (and also incorrect compared to normal go play).


Let's try to explain. Many problems have to take into account when creating a rule (J89, J2003, GT territoy rule ...).
The three main problems are the following:
1) Avoid ko fight through all the board and especially a local ko and ko threats in a remote area (typically bent four in the corner)
2) Be aware of local consequences when a player kill a local group at the cost of allowing another local group to live (typically L&D example 4)
3) loop breaking, especially a loop advantageous to one player which could not be solved in normal play (typically mooshine life).

In J89 the "pass-for-ko" tool was created and the intention was surely to deal with points 1) and 3) above.
Though some side effects has been identified, it appears that this tool works very efficiently in the great majority of cases. Note also that rare loops are not breaked (see https://lifein19x19.com/viewtopic.php?p=266770#p266770) but because the tool is really efficient I can say it is acceptable.
It remains the difficult point 2). In J89 this point is handled by the famous "enable" region (in J2003 it is the local2 region). Yes "enable" is ambiguous because it looks an intuitive notion but the point is quite understandable and even "enable" is for me acceptable.

Now is the point: it appears that "enable" region might be also used for handling a number of positions corresponding to point 1), though it was not the primary intention OC.
What does that mean?
In J89, when considering the "pass-for-ko" tool in one hand and the "enable" region tool in the other hand the two tools can handled a great part of point 1), the first tool handled point" 3) and the second tool handled point 2)

I do not know Korean rule, but the intention of GT territory rule is to try other tools hoping to solve the majority of side effects known in J89:
The "controlled area" is used only to handled points 1) and 2)
The "permanently prohibited ko" is used only to handled point 3)
With this clear separation I hope it will be easier to solved side effects.

I do not claim that I fully understood what is behind the point 1) 2) 3) which represent in my mind the main logic of japanese go. I agree these three points may not be logical (at least to area go player) but that is something else.
What I claim is only the following : providing points 1) 2) 3) are clear or (more generally) providing the logic of japanese go has be been clearly defined, then other approach (different from J89) may be worth to study and might even reach better results (less side effects).

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #131 Posted: Wed Sep 08, 2021 9:50 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Quote:
The three main problems are the following:
1) Avoid ko fight through all the board and especially a local ko and ko threats in a remote area (typically bent four in the corner)

No. Normal play is king, and it doesn't work like this. "Enable" is good enough (cannot ignore threats anywhere).

OC this is not the Japanese approach, but you seem to state the above as a general requirement - it is not, only a traditional rule hack.

Quote:
2) Be aware of local consequences when a player kill a local group at the cost of allowing another local group to live (typically L&D example 4)

You should remove "local" from the above (and enable is enough again).

Quote:
I do not know Korean rule

I linked jaeup's post with insightful comments earlier.

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #132 Posted: Wed Sep 08, 2021 10:53 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Quote:
The three main problems are the following:
1) Avoid ko fight through all the board and especially a local ko and ko threats in a remote area (typically bent four in the corner)

No. Normal play is king, and it doesn't work like this. "Enable" is good enough (cannot ignore threats anywhere).

OC this is not the Japanese approach, but you seem to state the above as a general requirement - it is not, only a traditional rule hack.


It seems I have difficulty to understand this point. So let's take a diagram
Click Here To Show Diagram Code
[go]$$W
$$ -------------------------
$$ | O O O . X O . . . . . |
$$ | . X X X X O . . . . . |
$$ | X X O O O O . . . . . |
$$ | O O O . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | X X X X X X X . . . . |
$$ | O O O O O O X . . . . |
$$ | . X X X . O X . . . . |
$$ -------------------------[/go]


To avoid being killed in the bent four black use the ko threat at the bottom.
Now I do not take a specific rule (neithe J89, nor korean, nor J2003, nor ...)
What is the general japanese principle (logic? or whatever word you use) you apply to say that the bent four is dead while the seki remains seki?

1) No ko threat are allowed approach:
Click Here To Show Diagram Code
[go]$$W
$$ -------------------------
$$ | X O . X X O . . . . . |
$$ | 1 X X X X O . . . . . |
$$ | X X O O O O . . . . . |
$$ | O O O . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | X X X X X X X . . . . |
$$ | O O O O O O X . . . . |
$$ | 2 X X X 3 O X . . . . |
$$ -------------------------[/go]

even after the sequence :w1: :b2: :w3: black is not allowed to retake the ko => bent four is dead

2) "enable" region approach:
Click Here To Show Diagram Code
[go]$$W
$$ -------------------------
$$ | X O 3 X X O . . . . . |
$$ | 1 X X X X O . . . . . |
$$ | X X O O O O . . . . . |
$$ | O O O . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | X X X X X X X . . . . |
$$ | O O O O O O X . . . . |
$$ | 2 X X X 4 O X . . . . |
$$ -------------------------[/go]

black stones in the bent four are dead, black was allowed to take new region at the bottom. This new region is not in the "enable" region => bent four is dead

The result is the same but the motivation to reach the result is different.
What is the good approach in order to build correctly the rule?

Click Here To Show Diagram Code
[go]$$W
$$ -------------------------
$$ | . O a X X O . . . . . |
$$ | O X X X X O . . . . . |
$$ | X X O O O O . . . . . |
$$ | O O O . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | X X X X X X X . . . . |
$$ | O O O O O O X . . . . |
$$ | X X X X b O X . . . . |
$$ -------------------------[/go]

IOW, what is the more logic :w3: move? a move at "a" or a move at "b"?

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #133 Posted: Wed Sep 08, 2021 11:17 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Gérard TAILLE wrote:
Let's try to explain. Many problems have to take into account when creating a rule (J89, J2003, GT territoy rule ...).
The three main problems are the following:
1) Avoid ko fight through all the board and especially a local ko and ko threats in a remote area (typically bent four in the corner)
2) Be aware of local consequences when a player kill a local group at the cost of allowing another local group to live (typically L&D example 4)
3) loop breaking, especially a loop advantageous to one player which could not be solved in normal play (typically mooshine life).

My suggestion would be an addition of at least

4) You will NEVER reach identical results with different types of rulesets.

However, this should not be a grave barrier that prevented you from creating a consistent one.


+++++++++++++++++++++++++

The main task in "uncapturable"-type rulesets (J89, J2003, ...) is
"Kill me, if you can!"

The main task in "two-eye-formation"-type rulesets (GT territory rules, ...) is
"Prevent me from creating a two-eye-formation, if you can!"

It should be evident that both tasks are NOT the opposite / complement of each other, as FAILURE in the "uncapturable" world can mean "unconditionally alive" or "alive in seki", while SUCCESS in the "two-eye-formation" world can mean "alive in seki" or "dead".

This does not only result in the need for additional teire is special cases (costing one point of territory, equivalent to the one-move difference during status confirmation between both worlds) in ONLY ONE of both worlds, but also in a SEKI in one world, where is TERRITORY in the other one.


+++++++++++++++++++++++++

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | X X . X . X O |
$$ | X O X X X X O |
$$ | O O O O O O O |
$$ | . O X X X X X |
$$ | X O X . . X . |
$$ | . O X . . X X |
$$ | O O X . . X . |
$$ -----------------[/go]

You will recognise this position for sure.

--------------------

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | Z Z . B . B W |
$$ | Z W B B B B W |
$$ | W W W W W W W |
$$ | . W X X X X X |
$$ | B W X . . X . |
$$ | . W X . . X X |
$$ | W W X . . X . |
$$ -----------------[/go]

J89 has this result of the first run of status confirmation.

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | # # . # . # @ |
$$ | # @ # # # # @ |
$$ | @ @ @ @ @ @ @ |
$$ | . @ X X X X X |
$$ | # @ X . . X . |
$$ | . @ X . . X X |
$$ | @ @ X . . X . |
$$ -----------------[/go]

J89's final result is "seki".

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | X X 1 X . X O |
$$ | X O X X X X O |
$$ | O O O O O O O |
$$ | . O X X X X X |
$$ | X O X . . X . |
$$ | . O X . . X X |
$$ | O O X . . X . |
$$ -----------------[/go]

However, this position will NOT reach status confirmation, as :w1: will capture Black's three stones in the corner.

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | a 2 O X . X O |
$$ | 3 O X X X X O |
$$ | O O O O O O O |
$$ | 4 O X X X X X |
$$ | X O X . . X . |
$$ | 5 O X . . X X |
$$ | O O X . . X . |
$$ -----------------[/go]

:b2: captures one White stone.
:w3: turns this position into a seki -- in principle.
:b4: is enforced. Otherwise, White would be able to give atari at A, resulting in the capture of Black's entire moonshine group.
:w5: captures another two Black stones.

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | a X . X . X O |
$$ | O O X X X X O |
$$ | O O O O O O O |
$$ | 6 O X X X X X |
$$ | 7 O X . . X . |
$$ | O O X . . X X |
$$ | O O X . . X . |
$$ -----------------[/go]

:b6: has to stop White from giving atari at A again.

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | . X . X . X O |
$$ | O O X X X X O |
$$ | O O O O O O O |
$$ | . O X X X X X |
$$ | O O X . . X . |
$$ | O O X . . X X |
$$ | O O X . . X . |
$$ -----------------[/go]

:w7: captures another Black stone. The final position is "seki".

In total, we can count 6 captives for White vs. 1 captive for Black => 5 points for White during play.

--------------------

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | Z Z . # . # @ |
$$ | Z @ # # # # @ |
$$ | @ @ @ @ @ @ @ |
$$ | . @ X X X X X |
$$ | # @ X . . X . |
$$ | . @ X . . X X |
$$ | @ @ X . . X . |
$$ -----------------[/go]

This diagram shows the result of the first run of status confirmation with a "two-eye-formation" ruleset without any special regulations.

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | # # . # . # @ |
$$ | # @ # # # # @ |
$$ | @ @ @ @ @ @ @ |
$$ | . @ X X X X X |
$$ | # @ X . . X . |
$$ | . @ X . . X X |
$$ | @ @ X . . X . |
$$ -----------------[/go]

The final conclusion is "seki" again.

Same procedure as above will apply during play.

--------------------

Click Here To Show Diagram Code
[go]$$W
$$ -----------------
$$ | Z Z . Z . Z W |
$$ | Z W Z Z Z Z W |
$$ | W W W W W W W |
$$ | . W X X X X X |
$$ | Z W X . . X . |
$$ | . W X . . X X |
$$ | W W X . . X . |
$$ -----------------[/go]

This is the result of status confirmation under GT territory rules.
Please note that White -- contrary to the "uncapturable" world -- will not want to continue in the upper left corner.

You will have to come up with a very good reason why White should have the upper hand in a position that combines a Black "moonshine-life" with a White one.


+++++++++++++++++++++++++

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O . O X X O , .
$$ | O O O . O X . X O . .
$$ | X X O O X . X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

J89's L&D example 18 again.

--------------------

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . W . Z W Z O . . . .
$$ | W Z Z Z W Z O O . . .
$$ | . Z W W W Z Z O O . .
$$ | Z Z W W . W Z Z O , .
$$ | W W W . W Z . Z O . .
$$ | X X W W Z . Z Z O . .
$$ | . X X W W Z Z O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

This is the result of status confirmation in the "uncapturable" world (J2003, J89's adjusted edition), the entire corner is White's territory.

--------------------

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . W . Z W # O . . . .
$$ | W Z Z Z W # O O . . .
$$ | . Z W W W # # O O . .
$$ | Z Z W W . P # # O , .
$$ | W W W . W # . # O . .
$$ | X X W W Z . # # O . .
$$ | . X X W W # # O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

This diagram shows the result of the first run of status confirmation with a "two-eye-formation" ruleset without any special regulations.

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . W . Z W # O . . . .
$$ | W Z Z Z W # O O . . .
$$ | . Z W W W # # O O . .
$$ | Z Z W W . @ # # O , .
$$ | W W W a W # . # O . .
$$ | X X W W # . # # O . .
$$ | . X X W W # # O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

The concluding diagram might look a bit strange, but it is consistent.

Black's bent-four-in-the-corner is "dead", while Black's double-ko group at the right is "alive in seki".
White's stones in the corner are "alive", as are her stones in the double-ko (but one, which is "alive in seki").
DAME do not play a role, so A is one point of territory for White.

--------------------

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . @ . # @ # O . . . .
$$ | @ # # # @ # O O . . .
$$ | . # @ @ @ # # O O . .
$$ | # # @ @ . @ # # O , .
$$ | @ @ @ . @ # . # O . .
$$ | X X @ @ # . # # O . .
$$ | . X X @ @ # # O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

This is the concluding result under GT territory rules (version before the earthquake; otherwise the entire corner would be Black TERRITORY).

You will have to come up with a very good reason for "SEKI" or to find an additional machanism for your ruleset that prevents this undesired result.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #134 Posted: Wed Sep 08, 2021 11:37 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Gérard TAILLE wrote:
To avoid being killed in the bent four black use the ko threat at the bottom.
Now I do not take a specific rule (neithe J89, nor korean, nor J2003, nor ...)
What is the general japanese principle (logic? or whatever word you use) you apply to say that the bent four is dead while the seki remains seki?

As I wrote this is not the Japanese approach. But a straightforward (territory scoring) logic says stones are only dead if can be cleanly captured without giving up (or enabling) anything. So bent4+seki is alive - as is often the case in actual Chinese games!

L/D confirmation is static, it cannot worry if an exchange is favorable or not, and cannot be as accurate as playout with pass stones. But erring on the safe (alive) side seems better (only dead if 100% dead). A doubtful alive may mean 1-2 reinforcing moves for small score drift, while doubtful dead can be large gift and error. OC, this approach can also be unfair for some bent4s.

But other logic and approaches are also possible, and can be reasonable (tolerable) including the J89 one. The key is consistency - you NEED a convincing explanation / theory. Consider four examples:

  1. Bent4 with remote unremovable threat (or double ko seki)
  2. Bent4 with nearby unremovable threat (or double ko seki)
  3. Moonshine life with remote double ko seki
  4. Moonshine life with nearby double ko seki

Normal play says: All 4 alive.
REASON: cannot capture (without giving up something), so those are clearly not dead stones in territory.

Korean rules say: #1 and #3 is dead, others alive.
REASON: we should look at L/D in regions separately.

Japanese rules say: All 4 dead.
REASON: we should look at L/D without relying on any ko fight.

Normal play with a moonshine ko rule: #3 and #4 is dead, others alive.
REASON: moonshine loops are forbidden (there is a rich theory behind this, see elsewhere).

Your rules: only #2 is alive.
REASON: ... ?

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #135 Posted: Wed Sep 08, 2021 1:58 pm 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Gérard TAILLE wrote:
To avoid being killed in the bent four black use the ko threat at the bottom.
Now I do not take a specific rule (neithe J89, nor korean, nor J2003, nor ...)
What is the general japanese principle (logic? or whatever word you use) you apply to say that the bent four is dead while the seki remains seki?

As I wrote this is not the Japanese approach. But a straightforward (territory scoring) logic says stones are only dead if can be cleanly captured without giving up (or enabling) anything. So bent4+seki is alive - as is often the case in actual Chinese games!

Confirmation is static, it cannot worry about whether an exchange is favorable or not, and cannot be as accurate as playout with pass stones. But erring on the safe (alive) side seems better (only dead if 100% dead). A doubtful alive may mean 1-2 reinforcing moves for small score drift, while doubtful dead can be large gift and error. OC, this approach can also be unfair for some bent4s.

But other logic and approaches are also possible, and can be reasonable (tolerable) including the J89 one. The key is consistency - you NEED a convincing explanation / theory. Consider four examples:

  1. Bent4 with remote unremovable threat (or double ko seki)
  2. Bent4 with nearby unremovable threat (or double ko seki)
  3. Moonshine life with remote double ko seki
  4. Moonshine life with nearby double ko seki

Normal play says: All 4 alive.
REASON: cannot capture (without giving up something), so those are clearly not dead stones in territory.

Korean rules say: #1 and #3 is dead, others alive.
REASON: we should look at L/D in regions separately.

Japanese rules say: All 4 dead.
REASON: we should look at L/D without relying on any ko fight.

Normal play with a moonshine ko rule: #3 and #4 is dead, others alive.
REASON: moonshine loops are forbidden (there is a rich theory behind this, see elsewhere).

Your rules: only #2 is alive.
REASON: ... ?


What is GT territory rule logic : to respect as far as possible the traditionnal japanese rule (I mean J89), except if the result given by J89 contradicts normal play.
"normal play" is easy to understand but another important point appears in J89 : when a player kills a group of stones then, before concluding the group of stones is dead, we have to look at all consequences in the "enable" region => this group of stone might be finally declared alive.
Having this in mind the GT territory rule logic became:
To respect as far as possible the traditionnal japanese rule (I mean J89), except if the result given by J89 contradicts normal play inside "enable" region.

When this contradiction arises?
The majority of the cases corresponds to a pass-for-ko which disturbs normal play in the "enable" region.
I gave several exemples on this point and the "Bent4 with nearby unremovable threat" is nothing else that one of these cases: under normal play within the "enable" region, if you kill the bent four then the unremovable ko threat in the seki is efficient. In these conditions I chose to change J89 result in oder to align the result to normal play inside the "enable" region.

I hope that all differencies between J89 and GT territory rule are the consequence of this strategy. Maybe you can verify youself with all examples I showed here.

Remember your post https://lifein19x19.com/viewtopic.php?p=266868#p266868. It was a serious blow for me because I felt I missed something by building my strategy. The idea was not to change J89 for such unexpected situation and I had to work hard to clarify my proposal accordingly.

I know I could have chosen another strategy but here again the goal was not to change J89 but to put more consistency by aligning the result to normal play in the "enable" region.

I believe I am realist Jann, you have not to consider GT territory rule as a new rule but rather as a tool to have a better understanding of J89 and its weakness if not its inconsistencies.

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #136 Posted: Wed Sep 08, 2021 10:57 pm 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Gérard TAILLE wrote:
I believe I am realist Jann, you have not to consider GT territory rule as a new rule but rather as a tool to have a better understanding of J89 and its weakness if not its inconsistencies.

There are NO inconsistencies in J89(<= 2007 revised edition)!

Quite apparently, no one, neither in JAPAN nor in KOREA, considered it worthwhile to use "normal" play in the phase of judging life and death UNDER TERRITORY RULES.
So why do you, as a European, who will never understand the underlying culture in every detail, attach so much importance to it?

Probably because you do NOT understand your OWN approach?

Click Here To Show Diagram Code
[go]$$B
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O . O X X O , .
$$ | O O O . O X . X O . .
$$ | X X O O X . X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

J89's L&D example 18 again.
Question: Is the corner White territory under GT territory rules(before the earthquake; advantageous loop for the defender only)?

Click Here To Show Diagram Code
[go]$$B :b3: pass
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O 1 O X X O , .
$$ | O O O . O X . X O . .
$$ | X X O O X 2 X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

Click Here To Show Diagram Code
[go]$$B :w6: pass; :b7: pass
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O X 4 X X O , .
$$ | O O O . O X . X O . .
$$ | X X O O 5 O X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

Click Here To Show Diagram Code
[go]$$B :w10: pass; :b11: pass
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O 9 O X X O , .
$$ | O O O . O X . X O . .
$$ | X X O O X 8 X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

The defender has an "advantageous loop", and therefore claims "permanently prohibited ko" for both ko-shapes.

Click Here To Show Diagram Code
[go]$$Wm12 :b13: pass
$$ +----------------------
$$ | . O . X O X O . . . .
$$ | O X X X O X O O . . .
$$ | . X O O O X X O O . .
$$ | X X O O X 1 X X O , .
$$ | O O O . O X 3 X O . .
$$ | X X O O M O X X O . .
$$ | . X X O O X X O O . .
$$ | . . X X X O O O . . .
$$ | . . . . X X . . . . .
$$ | . . . , . . . . . , .
$$ | . . . . . . . . . . .[/go]

Quod erat demonstrandum.

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #137 Posted: Thu Sep 09, 2021 7:34 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Gérard TAILLE wrote:
jann wrote:
  1. Bent4 with remote unremovable threat (or double ko seki)
  2. Bent4 with nearby unremovable threat (or double ko seki)
  3. Moonshine life with remote double ko seki
  4. Moonshine life with nearby double ko seki

Normal play says: All 4 alive.
REASON: cannot capture (without giving up something), so those are clearly not dead stones in territory.

Korean rules say: #1 and #3 is dead, others alive.
REASON: we should look at L/D in regions separately.

Japanese rules say: All 4 dead.
REASON: we should look at L/D without relying on any ko fight.

Normal play with a moonshine ko rule: #3 and #4 is dead, others alive.
REASON: moonshine loops are forbidden (there is a rich theory behind this, see elsewhere).

Your rules: only #2 is alive.
REASON: ... ?

... the GT territory rule logic became:
To respect as far as possible the traditionnal japanese rule (I mean J89), except if the result given by J89 contradicts normal play inside "enable" region.

This isn't met in #4 above? J89 contradicts normal play there (even locally), yet you agree. What you say sounds like simple Korean rules (with consistent #2 and #4 handling).

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #138 Posted: Thu Sep 09, 2021 7:59 am 
Lives in gote

Posts: 445
Liked others: 0
Was liked: 37
Cassandra wrote:
no one, neither in JAPAN nor in KOREA, considered it worthwhile to use "normal" play in the phase of judging life and death UNDER TERRITORY RULES.
So why do you, as a European, who will never understand the underlying culture in every detail, attach so much importance to it?

I'm all in favor of modesty and careful approach to rules, but this is wrong.

The very difference between Japanese and Korean rules is that the latter doesn't use pass for ko and rely on normal play (in each region). If "normal play" means normal global play, then: it was mentioned that normal play can solve almost everything except a single problem (moonshine life).

If there were other problems then the game itself (using normal play) wouldn't work to begin with. And resumption is always an option, particularly playout in Chinese or with pass stones - this dictates REALITY, and your only hope is to get reasonably close to that. So obviously some kind of extra rule is necessary - preferably affecting that particular problem only.

Quote:
Is the corner White territory under GT territory rules(before the earthquake; advantageous loop for the defender only)?
...
The defender has an "advantageous loop", and therefore claims "permanently prohibited ko" for both ko-shapes.

By this logic all double ko sekis die (I think you applied his ideas incorrectly here). Anyway, rules need much more careful (and principled) approach.


Last edited by jann on Thu Sep 09, 2021 8:43 am, edited 2 times in total.
Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #139 Posted: Thu Sep 09, 2021 8:38 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
jann wrote:
By this logic all double ko sekis die (I think you applied his ideas incorrectly here). Anyway, rules need much more careful (and principled) approach.

Indeed, this might be an unwanted side-effect.

By the way, "all" seems to be kind of exaggeration.
By the way #2, that unwanted side-effect mentioned above would not matter in a TWO-run process for status confirmation. Whether both DOUBLE-ko groups were "alive" or "dead" would result in the SAME conclusion of "seki".

Probably Gérard will find a modification without, who knows?
Everything needed has already been published...

_________________
The really most difficult Go problem ever: https://igohatsuyoron120.de/index.htm
Igo Hatsuyōron #120 (really solved by KataGo)

Top
 Profile  
 
Offline
 Post subject: Re: GT territory rule
Post #140 Posted: Thu Sep 09, 2021 9:59 am 
Lives in sente

Posts: 1226
Liked others: 21
Was liked: 57
Rank: 1d
jann wrote:
Gérard TAILLE wrote:
jann wrote:
  1. Bent4 with remote unremovable threat (or double ko seki)
  2. Bent4 with nearby unremovable threat (or double ko seki)
  3. Moonshine life with remote double ko seki
  4. Moonshine life with nearby double ko seki

Normal play says: All 4 alive.
REASON: cannot capture (without giving up something), so those are clearly not dead stones in territory.

Korean rules say: #1 and #3 is dead, others alive.
REASON: we should look at L/D in regions separately.

Japanese rules say: All 4 dead.
REASON: we should look at L/D without relying on any ko fight.

Normal play with a moonshine ko rule: #3 and #4 is dead, others alive.
REASON: moonshine loops are forbidden (there is a rich theory behind this, see elsewhere).

Your rules: only #2 is alive.
REASON: ... ?

... the GT territory rule logic became:
To respect as far as possible the traditionnal japanese rule (I mean J89), except if the result given by J89 contradicts normal play inside "enable" region.

This isn't met in #4 above? J89 contradicts normal play there (even locally), yet you agree. What you say sounds like simple Korean rules (with consistent #2 and #4 handling).


Click Here To Show Diagram Code
[go]$$W
$$ -------------------------
$$ | X X O X . X . X O . . |
$$ | . X O O X X X X O . . |
$$ | X X O . O X O O O . . |
$$ | . X O O X X O . . . . |
$$ | X O O . O X O . . . . |
$$ | O O O O O O X . . . . |
$$ | X X X X X X X . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ | . . . . . . . . . . . |
$$ -------------------------[/go]

In one hand it is true that in normal play black stones in the corner cannot be killed but in the other hand it looks to me that the author of the J89 rule wanted really to see black group dead. Obviously this is an advantageous loop. You can note in particular that in normal play white can force a NO RESULT game and, then white can also force the stop of the game. Providing you do not accept an infinite number of resumption then black cannot force a NO RESULT game. In this context, because my idea was really not to change the basis of J89, I chose to give the same result as J89 => black stones are dead.
It looks to me a very specific situation comparing to situations where the result in normal play is reached without a loop or is reached by a loop that cannot be interrupted (solid triple ko, chosei ...)

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 146 posts ]  Go to page Previous  1 ... 4, 5, 6, 7, 8  Next

All times are UTC - 8 hours [ DST ]


Who is online

Users browsing this forum: No registered users and 1 guest


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:  
cron
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group