Could you explain your proposal on the mooshine life position?kvasir wrote: This position demonstrates how dangerous it is to try to pick a winning side in ko fights for the purpose of status confirmation.
One question I have that pertains to this position is how many pass-ko bans do we really need for each player? In this position black is dead because white can create two pass-ko bans in time to fill blacks last two liberties. Is it really needed to allow multiple pass-ko bans on the same player to achieve the desired effect? For example if white had to pick which ko had a pass-ko ban (that is one or the other ko stone would not be subject to a pass-ko), that would solve this position. That is if a player is allowed only one pass-ko ban at a time, could that work? If not, how many pass-ko bans do we really need?
The "no ban in double ko" rule to solve pass-for-ko issue ?
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
It is exactly the same as example 11 because the normal ko rule prevents white from taking again in the double-ko.
No one really knows why example 11 has white pass for
instead of 
and not create the pass-ko cycle in the double-ko. You will need another explanation of this (as you are actually working on) but maybe some of the tricky cases could be handled by limiting the number of pass-ko bans? I did not mean to put this up against some method to treat the double-ko cycle (that some believe doesn't exists for some reason) but to avoid using ko shapes to play multiple approach moves during status confirmation.
Example 7-2 (a bent-4 vs. 10'000 year ko) also works fine because again there is a point when two ko bans are needed buy one is provided with the normal ko rule.
I do not quite see why one needs to allow multiple pass-ko bans to get "Japanese" results. Obviously, there are different results in positions when one side could otherwise use pass-ko to play multiple moves in a row but that is not really bread and butter "Japanese" approach is it? Maybe there is some position someone came up with? But (by potentially fault intuition) I'd suggest that positions that require multiple pass-ko bans are quite in a flux, having more than one possible ko fight at once.
No one really knows why example 11 has white pass for
instead of and not create the pass-ko cycle in the double-ko. You will need another explanation of this (as you are actually working on) but maybe some of the tricky cases could be handled by limiting the number of pass-ko bans? I did not mean to put this up against some method to treat the double-ko cycle (that some believe doesn't exists for some reason) but to avoid using ko shapes to play multiple approach moves during status confirmation. Example 7-2 (a bent-4 vs. 10'000 year ko) also works fine because again there is a point when two ko bans are needed buy one is provided with the normal ko rule.
I do not quite see why one needs to allow multiple pass-ko bans to get "Japanese" results. Obviously, there are different results in positions when one side could otherwise use pass-ko to play multiple moves in a row but that is not really bread and butter "Japanese" approach is it? Maybe there is some position someone came up with? But (by potentially fault intuition) I'd suggest that positions that require multiple pass-ko bans are quite in a flux, having more than one possible ko fight at once.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
The position reached is the following:kvasir wrote:It is exactly the same as example 11 because the normal ko rule prevents white from taking again in the double-ko.
No one really knows why example 11 has white pass for
Example 7-2 (a bent-4 vs. 10'000 year ko) also works fine because again there is a point when two ko bans are needed buy one is provided with the normal ko rule.
I do not quite see why one needs to allow multiple pass-ko bans to get "Japanese" results. Obviously, there are different results in positions when one side could otherwise use pass-ko to play multiple moves in a row but that is not really bread and butter "Japanese" approach is it? Maybe there is some position someone came up with? But (by potentially fault intuition) I'd suggest that positions that require multiple pass-ko bans are quite in a flux, having more than one possible ko fight at once.
In this position black has created two pass-for-ko ban => is white allowed to retake the ko at the bottom?
Compare with the following position:
and white has created two pass-for-ko ban => is black allowed to retake the ko at the top?
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
White is not allowed to take at the bottom because black chooses to protect there with the pass-ko, likewise white is not allowed to take in the double-ko because of the regular ko rule. There is no difference because black only needs to protect one ko using pass-ko (hypothetically because there is the pass-ko double-ko cycle anyway).
It would be up to white which ko to protect with pass-ko, but he could not protect both if the number of active pass-ko bans was limited to one per player at time. The position is symmetric so white just picks either one and black then takes the other ko.
Please come up with something to challenge this if I was able to explain what I meant
What I am trying to say is that if each player is allowed to protect one ko at a time using pass-ko it seems to work with all the J89 examples and rule out some forms of pass-ko abuse (for lack of a neutral phrase). It is also similar to pass-once but doesn't preclude using the pass-ko rule repeatedly in the same shape.
It would be up to white which ko to protect with pass-ko, but he could not protect both if the number of active pass-ko bans was limited to one per player at time. The position is symmetric so white just picks either one and black then takes the other ko.
Please come up with something to challenge this if I was able to explain what I meant
What I am trying to say is that if each player is allowed to protect one ko at a time using pass-ko it seems to work with all the J89 examples and rule out some forms of pass-ko abuse (for lack of a neutral phrase). It is also similar to pass-once but doesn't preclude using the pass-ko rule repeatedly in the same shape.-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
I do not understand. As in the previous example white can choose simply to protect the ko at the top and the other ko is protected by regular ko. What is the difference?kvasir wrote: White is not allowed to take at the bottom because black chooses to protect there with the pass-ko, likewise white is not allowed to take in the double-ko because of the regular ko rule. There is no difference because black only needs to protect one ko using pass-ko (hypothetically because there is the pass-ko double-ko cycle anyway).
It would be up to white which ko to protect with pass-ko, but he could not protect both if the number of active pass-ko bans was limited to one per player at time. The position is symmetric so white just picks either one and black then takes the other ko.
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
There is no regular ko ban in this position, only the pass-ko bans. If white can only have a single pass-ko ban at a time then black could take back in one of the kos without first passing. That is the difference.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
Oh I see where is the misunderstanding. The position I proposed is not that one above but the following:kvasir wrote: There is no regular ko ban in this position, only the pass-ko bans. If white can only have a single pass-ko ban at a time then black could take back in one of the kos without first passing. That is the difference.
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
That is a different position and white would be able to use the pass-ko and the regular ko ban to demonstrate the black group is dead. The result would then be a seki. This is exactly same as with regular pass-ko rule because you still have two ko bans.Gérard TAILLE wrote:
You could create a position like this to get different results.
Is it desirable or undesirable to let white claim there are three ko bans now? I'd say it is undesirable, and I find it hard to imagine a position when we want one side to defend SIMULTANIOUSLY in more than two kos. Such things may exist but the examples in j89 don't require it.
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
I think now your are right Jann. I am not able to find an example in which one player need more than one pass-ko ban. Good point indeed.kvasir wrote: This position demonstrates how dangerous it is to try to pick a winning side in ko fights for the purpose of status confirmation.
One question I have that pertains to this position is how many pass-ko bans do we really need for each player? In this position black is dead because white can create two pass-ko bans in time to fill blacks last two liberties. Is it really needed to allow multiple pass-ko bans on the same player to achieve the desired effect? For example if white had to pick which ko had a pass-ko ban (that is one or the other ko stone would not be subject to a pass-ko), that would solve this position. That is if a player is allowed only one pass-ko ban at a time, could that work? If not, how many pass-ko bans do we really need?
Coming back to my proposal with the "no ban in double ko" rule the point is that in some double ko situations even a single pass-ko ban is an issue:
and white can neither retake the ko at
(pass-ko ban) nor the ko at (regular ko).Re: The "no ban in double ko" rule to solve pass-for-ko issu
That was not my quote, I still think the left-or-right examples from the other topic show why one pass-ko ban is not enough - IF Japanese-style results (no ko interaction) are desired.Gérard TAILLE wrote:I think now your are right Jann. I am not able to find an example in which one player need more than one pass-ko ban.kvasir wrote:That is if a player is allowed only one pass-ko ban at a time, could that work? If not, how many pass-ko bans do we really need?
-
Gérard TAILLE
- Gosei
- Posts: 1346
- Joined: Sun Aug 23, 2020 2:47 am
- Rank: 1d
- GD Posts: 0
- Has thanked: 21 times
- Been thanked: 57 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
Yes I agree Jann. In this thread I only consider positions with double ko and in this context and in Japanese-style rule your comment was to me quite interesting.jann wrote:That was not my quote, I still think the left-or-right examples from the other topic show why one pass-ko ban is not enough - IF Japanese-style results (no ko interaction) are desired.Gérard TAILLE wrote:I think now your are right Jann. I am not able to find an example in which one player need more than one pass-ko ban.kvasir wrote:That is if a player is allowed only one pass-ko ban at a time, could that work? If not, how many pass-ko bans do we really need?
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
I don't get what interactions in you mean in that specific position, Jann. I think I treated the position incorrectly before, it is a bit tricky to have a pass-ko and a regular ko ban, now I don't think there is any difference from regular pass-ko in this case because of how black takes the kos first. Because black takes the kos first he will have a regular ko ban in one and a pass-ko ban in the other, allowing black to capture everything.jann wrote:That was not my quote, I still think the left-or-right examples from the other topic show why one pass-ko ban is not enough - IF Japanese-style results (no ko interaction) are desired.Gérard TAILLE wrote:I think now your are right Jann. I am not able to find an example in which one player need more than one pass-ko ban.kvasir wrote:That is if a player is allowed only one pass-ko ban at a time, could that work? If not, how many pass-ko bans do we really need?
It is basically the same as your pass-once except that possibility of reusing the pass-ko ban is not discarded AND it solves the issue of allowing use of pass-ko to force multiple moves in a row in at least some positions. Like with any change or interpretation of anything in J89 there will be different results in some positions (that is the intention) but it does appear to work with all of the examples.
Anyway, I don't expect it to work 100%. For one thing J89 still has unclear "enable", a double-ko cycles, and a lofty idea of a "Japanese" understanding of what is the correct result. It is too subjective to really match everyone's idea of what 100% is. I am more interested in understanding the benefits and limitations.
Re: The "no ban in double ko" rule to solve pass-for-ko issu
I thought that: B starts the bent4 ko and takes it first, W throws in at right, B captures there as well. At this point (if there is only one pass-ko ban) W passes for whichever of the two kos B chooses to have protected. Next B resolves one of the kos and W captures in the other, thus saving central stones and even bent4 will be alive (directly or on enabling). Right?kvasir wrote:I don't get what interactions in you mean in that specific position, Jann. I think I treated the position incorrectly before, it is a bit tricky to have a pass-ko and a regular ko ban, now I don't think there is any difference from regular pass-ko in this case because of how black takes the kos first. Because black takes the kos first he will have a regular ko ban in one and a pass-ko ban in the other, allowing black to capture everything.
Hmm, I think J89 is actually not bad (if you accept its approach to ko), better than I thought in the past.For one thing J89 still has unclear "enable", a double-ko cycles, and a lofty idea of a "Japanese" understanding of what is the correct result.
-
kvasir
- Lives in sente
- Posts: 1040
- Joined: Sat Jul 28, 2012 12:29 am
- Rank: panda 5 dan
- GD Posts: 0
- IGS: kvasir
- Has thanked: 25 times
- Been thanked: 187 times
Re: The "no ban in double ko" rule to solve pass-for-ko issu
This is my current thought of how the position goes when black tries to capture in confirmation.
Black protects with pass-ko, the other ko will be protected by the regular ko rule.
White passes for
because he can take in neither ko shape, one protected by pass-ko and the other by regular ko rule.
finishes the ko that was protected by the regular ko rule. (There may be an ambiguity about if you are allowed to switch the pass-ko ban around but we don't need it here).
White still can't take in the bent-4 because of the pass-ko ban, black takes there also with
.
So it appears that black only needs one pass-ko ban to capture everything in this position.
I'd also add that the faults in some other rule texts are lot more comical than J89, for one thing we tend to just ignore anything not doing with the position on the board. That is really a faulty approach but understandable because it is the position in the game that interests us.
Black protects
with pass-ko, the other ko will be protected by the regular ko rule.White passes for
because he can take in neither ko shape, one protected by pass-ko and the other by regular ko rule. finishes the ko that was protected by the regular ko rule. (There may be an ambiguity about if you are allowed to switch the pass-ko ban around but we don't need it here).White still can't take in the bent-4 because of the pass-ko ban, black takes there also with
.So it appears that black only needs one pass-ko ban to capture everything in this position.
I agree that J89 is not so bad. It is just that I think that we can't talk about perfection or treat its concepts formally all the time. The semantics of J89 are not always clear enough, but I do like how it presents a simple and elegant approach. I still doubt it really works except in practice (when we are really just happy to finish the game).jann wrote:Hmm, I think J89 is actually not bad (if you accept its approach to ko), better than I thought in the past.For one thing J89 still has unclear "enable", a double-ko cycles, and a lofty idea of a "Japanese" understanding of what is the correct result.
I'd also add that the faults in some other rule texts are lot more comical than J89, for one thing we tend to just ignore anything not doing with the position on the board. That is really a faulty approach but understandable because it is the position in the game that interests us.
Re: The "no ban in double ko" rule to solve pass-for-ko issu
I don't understand this. As I wrote I think W passes for the protected ko (B 1). What does it mean to pass for a ko that is not protected anyway?kvasir wrote:Black protects
White passes for
Hm, maybe you mean passing for the ko itself is forbidden by the protection? But then how can such protection be eliminated? You need a way for that to prevent anomalies (or a different kind of double ko abuse).