Fixed Ko Rule: Continued Discussion

[RobertJasiek]

can you find another one that does not increase the complexity of the initial rule?

The textual complexity? No (not immediately). Everything related with short text length is already discovered. Also see http://home.snafu.de/jasiek/koclass.html

The question is that choosing a fixed ko rule that would in most cases choose which side wins, can either be fairly chosen? I don't think so. This is why, while simple, I think there must be a better solution than the fixed ko rule.

The fixed ko rule has never had the purpose to be used as the only ko rule in a real world ruleset. Therefore I do not care about perception of fairness. - I invented the fixed ko rule as a theoretical study tool when substituting other ko rulesets. Once invented, I noticed more clearly than before a) the disturbing ko nature of the rule's effect and b) that the fixed ko rule can be combined with other ko rules to form new ko rulesets useful for yet further purposes: 1) Basic-Fixed-Ko Rules are a very well possible real world ko rulesets. 2) The fixed ko rule as part of the Default Restriction Rules contributes to defining "ko" in general: http://home.snafu.de/jasiek/ko.pdf

The fixed ko rule is the best solution for the purpose of creating (a) and for the purpose (2). Whether (1) is better or worse than other ko rulesets for real world play is subject to subjective preference or one's set axioms of aims.

So it does not make much sense to call other ko rulesets "better"; sense we get only in the context of also choosing purposes of application. Such a context you are missing so far, I think.

[emeraldemon]

I didn't really follow the original discussion, but to me plain positional superko seems much superior to more complicated options. It is
(1) Easy to state
(2) Easy to understand and predict
(3) Doesn't void games

I've heard most pro tournaments don't use this rule, but it seems to have more to do with inertia than anything to me. FWIW I also find area scoring (or AGA style pass stones) preferable to territory scoring, for a similar reason.

[RobertJasiek]

emeraldemon, you seem to discuss ko rule(s) for the purpose of human play and to compare superko with Japanese style ko rules?

For that purpose, lots of ko rulesets are better than Japanese style ko rules. Examples of possible ko rulesets for that purpose:

1) superko

2) Basic-Fixed-Ko Rules

3) Short Cycle Rules

4) Long Cycle Rules

(2), (3) and (4) are not as concise and easily understood as rule text as (1) but have another - what might be perceived as - advantage: long cycle ko strategy is understood easily.

[willemien]

emeraldemon, you seem to discuss ko rule(s) for the purpose of human play and to compare superko with Japanese style ko rules?

For that purpose, lots of ko rulesets are better than Japanese style ko rules. Examples of possible ko rulesets for that purpose:

1) superko

2) Basic-Fixed-Ko Rules

3) Short Cycle Rules

4) Long Cycle Rules

(2), (3) and (4) are not as concise and easily understood as rule text as (1) but have another - what might be perceived as - advantage: long cycle ko strategy is understood easily.

What do you mean by:

3) Short Cycle Rules
and
4) Long Cycle Rules?

[emeraldemon]

How is long cycle ko strategy different from short (normal) ko strategy? In both cases, you have to find a ko threat to break the cycle, yes?

[RobertJasiek]

Short Cycle Rules in one possible wording:

"Creating a cycle of 2 or 3 plays is prohibited.
Creating a cycle of 4+ plays ends the game as a tie."


Long Cycle Rules see here under (2.) to (4.):

http://home.snafu.de/jasiek/int.html#AlternativeBasicKo


For long cycles, one needs to distinguish whether a ko or a set of kos behaves like a fighting ko, disturbing ko or semi-stable ko. Usually disturbing kos in their stable state do not need ko threats.

Changing the ko type might be interesting strategy.

Instead of only one cycle, a set of cycles might have to be considered.

Instead of ko-stone-sequences of constant length 1 until the next local ko ban, greater lengths have to be considered and variation might occur.

Ko fight details (like passes, nature of local threats, count evaluation or its characterists) might differ, too.

Ko threats for one ko might be ko plays in a different ko.

[willemien]

Short Cycle Rules in one possible wording:

"Creating a cycle of 2 or 3 plays is prohibited.
Creating a cycle of 4+ plays ends the game as a tie."

There is no "Short Cycle Rule" page on Sensei's page on this (So I don't understand it can we make a page for it?)

I also think that a cycle of 3 is positional (the other player is on the move)
Is this on correct or does is this an oversight?
(also some examples would be nice)

Long Cycle Rules see here under (2.) to (4.):

http://home.snafu.de/jasiek/int.html#AlternativeBasicKo


For long cycles, one needs to distinguish whether a ko or a set of kos behaves like a fighting ko, disturbing ko or semi-stable ko. Usually disturbing kos in their stable state do not need ko threats.

Changing the ko type might be interesting strategy.

Instead of only one cycle, a set of cycles might have to be considered.

Instead of ko-stone-sequences of constant length 1 until the next local ko ban, greater lengths have to be considered and variation might occur.

Ko fight details (like passes, nature of local threats, count evaluation or its characterists) might differ, too.

Ko threats for one ko might be ko plays in a different ko.

I just understood the long Cycle rule as described on Sensei's LongCycleRule but that has none of the disturbing ko definitions (that i am told only 2 people understand , you being one of them)
are they on the same principle or are you talking about something more complex?

[RobertJasiek]

I also think that a cycle of 3 is positional

Yes.

that has none of the disturbing ko definitions

Of course not. The text in my previous message starting with "For long cycles, one needs" is a reply to emeraldemon and does not explain the Long Cycle Rules, as I have meant to imply from consistent usage of 2 blank lines. (I format with meaning!)

(that i am told only 2 people understand

That joke is outdated since
http://home.snafu.de/jasiek/ko_types.pdf

[willemien]

How is long cycle ko strategy different from short (normal) ko strategy? In both cases, you have to find a ko threat to break the cycle, yes?

NO

This tread is not so much about strategy (it is about Go rules)

Under "normal" superko rules you are Strategicly right.

Under cycle rules it maybe possible to make no ko threat and still win or draw.

so it is a bit more nuanced than a simple yes/no

The problem is that a rule set needs to
1) describe what a repeat is.
2) what the consecuence of the repaet is

Wat a repeats is:

- The recreation of the position is a repeat? or only if the same player has the move? (discussion situational/ postional superlo)
- more limited rule (fixed ko rule) only making the same move in the same position is a repeat.
- what is the influence of passes on it (may you retake a ko, and so recreate a earlier position, after your opponent passes?) (this then leads on to questions like how many passes ends the game and so) and has also to do with the natural situational superko rule

What is the consequence of the repeat?

- The player who makes the repeat loses (that seems to me the essence of a superko rule)
- It depends on how many moves have been made between the repeats (that seems to me now the essence of the short cycle rule, but even i am not sure of it)
- It depend on how many captures have been made between the repeats (that seems to me now the essence of the long cycle rule)


I Hope this overview is clearand complete . (it shows all variations possible)
And maybe even correct. But i guess not everybody will agree with this

[RobertJasiek]

There is no "Short Cycle Rule" page on Sensei's page on this

Don't complain but make one.

(also some examples would be nice)

4 years ago I wrote examples, now again available:

http://home.snafu.de/jasiek/3_plays.pdf
http://home.snafu.de/jasiek/long_cycle.pdf

3-plays Ko Rules = Short Cycle Ko Rules

BTW, since I still have not had time to write a superko strategy and examples paper, maybe you could write some?

[willemien]

There is no "Short Cycle Rule" page on Sensei's page on this

Don't complain but make one.

I would if i understood it, for the moment to me it is no more than:

- Under the the short cycle rule it depends on how many moves have been made between the repeats.
if there are 4 or more plays between the repeats it is a draw, if 3 or less it is a loss for the player who makes the repeat.
And a repeat here is of the positional type (the same position with no regard for who's turn it is)

But i guess you will say i simplify it to much. (and I hope you can explain where i simply it to much)

4 years ago I wrote examples, now again available:

http://home.snafu.de/jasiek/3_plays.pdf
http://home.snafu.de/jasiek/long_cycle.pdf

3-plays Ko Rules = Short Cycle Ko Rules

Thanks

BTW, since I still have not had time to write a superko strategy and examples paper, maybe you could write some?

I am not strong enough for that, what i wrote earlier is about all i can come up with.

[RobertJasiek]

Rules application is possible by anybody who understands the rules; therefore also you can write a superko commentary:)

[emeraldemon]

In my first post I said I thought positional superko (PSK) was simple and easy to understand, and Robert replied

(2), (3) and (4) are not as concise and easily understood as rule text as (1) but have another - what might be perceived as - advantage: long cycle ko strategy is understood easily.

So I would like to argue that PSK has easily understood strategy. Here's the first position in the 3_plays.pdf where a long cycle would be optimal play if it were allowed:

$$Bcm1 2xko seki plus other ko
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O . |
$$ | X . X O X O X O |
$$ | X X O O X X X X |
$$ | X O . O X . X . |
$$ ------------------

Click Here To Show Diagram Code

[go]$$Bcm1 2xko seki plus other ko
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O . |
$$ | X . X O X O X O |
$$ | X X O O X X X X |
$$ | X O . O X . X . |
$$ ------------------[/go]

Black tries to capture the white stones, and white uses the double-ko seki as a threat:

$$Bcm1
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O 1 |
$$ | X 2 X O X O X 4 |
$$ | X X O O X X X X |
$$ | X O 3 O X . X . |
$$ ------------------

Click Here To Show Diagram Code

[go]$$Bcm1
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O 1 |
$$ | X 2 X O X O X 4 |
$$ | X X O O X X X X |
$$ | X O 3 O X . X . |
$$ ------------------[/go]

Now the board looks like this:

$$Bcm1
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O . |
$$ | X O . O X O X 4 |
$$ | X X O O X X X X |
$$ | X . X O X . X . |
$$ ------------------

Click Here To Show Diagram Code

[go]$$Bcm1
$$ ------------------
$$ | . O O O X O . O |
$$ | X X O O X O O . |
$$ | X O . O X O X 4 |
$$ | X X O O X X X X |
$$ | X . X O X . X . |
$$ ------------------[/go]

Now with PSK, black can capture, and white cannot retake, because it will repeat the position above:

$$Bcm1 White 6 must pass
$$ ------------------
$$ | 7 O O O X O . O |
$$ | X X O O X O O . |
$$ | X . 5 O X O X O |
$$ | X X O O X X X X |
$$ | X . X O X . X . |
$$ ------------------

Click Here To Show Diagram Code

[go]$$Bcm1 White 6 must pass
$$ ------------------
$$ | 7 O O O X O . O |
$$ | X X O O X O O . |
$$ | X . 5 O X O X O |
$$ | X X O O X X X X |
$$ | X . X O X . X . |
$$ ------------------[/go]

So white is doomed. The strategy is simple and easy to understand: in general the double-ko seki cannot be used as a source of ko threats. I feel the other rulesets mentioned only complicate the situation, making the possibilty of ties. If this were part of a larger board where black was already winning, he might be forced to let the white stones live rather than risk having the game voided due to the ko rule.

[HermanHiddema]

You said that superko is easy to understand, then Robert replied that superko is indeed easy to understand, therefore you felt the need to elaborate on why superko is easy to understand?

[emeraldemon]

You said that superko is easy to understand, then Robert replied that superko is indeed easy to understand, therefore you felt the need to elaborate on why superko is easy to understand?

Yes

Actually as I understand it we're talking about a subtle difference: We both agree that the rules of superko are easy, but Robert said that in his variants the strategy is easy to understand, implying that the strategy of superko is difficult or convoluted. I think the strategy of superko is not significantly more difficult than the strategy of regular ko.

Of course, regular kos are certainly enough to make my head spin