It is currently Thu Mar 28, 2024 2:25 am

All times are UTC - 8 hours [ DST ]




Post new topic Reply to topic  [ 189 posts ]  Go to page Previous  1 ... 6, 7, 8, 9, 10
Author Message
Offline
 Post subject: Re: User-friendly Reading of the Japanese 1989 Rules
Post #181 Posted: Mon Jul 12, 2010 10:50 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
Answer will probably be delayed, Robert.

I'm very busy with Igo Hatsuyoron's #120 at the moment. And it will take some time to enlighten the difference between single-instance-force and multi-instance-force.

_________________
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: User-friendly Reading of the Japanese 1989 Rules
Post #182 Posted: Tue Jul 13, 2010 8:12 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
There had been an idea this morning.

With some (minor) modifications you may be able to keep your J2003 multi-step procedure. And you will know what the side-effects are and where they do come from. Please refer to "A comment on ..." in the attached PDF.



Concerning table 2:
- Why is cell of column B31 row B32 a contradiction?

See attachment.

- What shall the table do to assist your claimed proof draft?
See attachment for "no capturable-2 string has any permanent stone in local-1". And for the one and only aspect local-2 might be needed for.

- "force" in the table is not strictly the same "force" of J2003.
I'm aware of this, see attachment the instances of J2003-force.

- You must specify that the opponent moves first, the player second.
Done, see attachment.

- Each "force" would have to be checked independently of the other "force"s before creating the table.
See attachment for J2003-procedure.

Concerning Min-max:
- What is minimized / maximized?

"A" > "B2" > "B31" > "B32"
The opponent chooses the smaller one; the player chooses the bigger one.

- Please explain how the tables represent an algorithm!
Basis is the variation tree. At each branching point there will a concluding decision, following the tables. Direction is "leaves" >>> "root".

- What (beyond basic Min-max) is that algorithm?
See above.

- What is that algorithm used for with respect to your claimed proof draft?
See attachment for "no capturable-2 string has any permanent stone in local-1".


Attachments:
Tables2.pdf [48.37 KiB]
Downloaded 406 times

_________________
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: User-friendly Reading of the Japanese 1989 Rules
Post #183 Posted: Wed Jul 14, 2010 2:15 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
I reply to the second version of your table, dated July 13.

Can you please post the comments on C1 / C2 in J2003 here as plain text again? This makes replying easier.

My early drafts of J2003 had a game tree like description for what now are sets. Therefore surely your idea to use game trees must be possible in principle.

Your ideas to order the statuses in a symbolic relation and base a minmax game tree evaluation and to redefine the statuses by this approach are new and nice alternatives.

You seem assume though that your redefinitions of J2003-alive-2\1 aka Casssandra-J2003-alive, J2003-uncapturable, J2003-capturable-1 and J2003-capturable-2 would equal our previously known (textual) definitions with their iteratively introduced terms. I want to see four proofs for that.

I have not checked every aspect of your decision matrices intuitively yet but anyway, even if you should have made some mistake, a similar construction must exist. With proofs, we would be sure that a given construction is indeed the intended one or whether you would be defining yet a different variation of J2003-style-alive.

"undefined" is maybe the wrong word. Why not simply add a conditional extra layer to the minmax trees, where you can put an "ignore" condition whereever necessary?

Although all you write is somehow related to what you really want to prove (or disprove:)), I do not see relevant contribution to the capturable-2 / -2/1 - conjecture yet. Until I will have seen the four proofs mentioned above, I would not accept something as a final proof for the conjecture anyway.

While you introduce lots of interesting new ideas, the way is still long...

Top
 Profile  
 
Offline
 Post subject: Re: User-friendly Reading of the Japanese 1989 Rules
Post #184 Posted: Thu Jul 15, 2010 3:41 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
You seem assume though that your redefinitions of J2003-alive-2\1 aka Casssandra-J2003-alive, J2003-uncapturable, J2003-capturable-1 and J2003-capturable-2 would equal our previously known (textual) definitions with their iteratively introduced terms. I want to see four proofs for that.
There are NO redefinitions of J2003-terms. The concluding status of a string depends on the statuses found for the leaves (i.e. the ends of the hypothetical sequences) only. And theses leaves' statuses are J2003-alike.

My paper gives a more detailed description only of what is meant by "force". I.e. a clear cut journey from leaves to root of the variation tree, highlighting all the braches and twigs, which could be choosen the other way round to reach a "forced" result of a evaluation sequence.

It is not explicitely written in the paper that it is necessary to have the "special area" defined as local-2 is defined in J2003 - excluding local-1 - to be able to get a concluding status "capturable-2\1".
Including local-1 in this "special" area does not do any harm (but is unnecessary), so you are free to do so if you want to have "capturable-2" instead of "capturable-2\1".

I have not checked every aspect of your decision matrices intuitively yet but anyway, even if you should have made some mistake, a similar construction must exist. With proofs, we would be sure that a given construction is indeed the intended one or whether you would be defining yet a different variation of J2003-style-alive.
As written above, nothing is changed in the J2003-style of defining "alive".

It is shown that the J2003 multi-step procedure gives a well-defined result for each string under evaluation.

And there is the side-effect of having found that it is not necessary to include local-1 in local-2 to acquire the desired results.

"undefined" is maybe the wrong word. Why not simply add a conditional extra layer to the minmax trees, where you can put an "ignore" condition whereever necessary?
"to be ignored" will have the same effect. You will have cells in the decision matrix used for later steps, which are not in effective use any more.

Although all you write is somehow related to what you really want to prove (or disprove:)), I do not see relevant contribution to the capturable-2 / -2/1 - conjecture yet. Until I will have seen the four proofs mentioned above, I would not accept something as a final proof for the conjecture anyway.
See top of posting.

_________________
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: User-friendly Reading of the Japanese 1989 Rules
Post #185 Posted: Thu Jul 15, 2010 4:35 am 
Lives in gote
User avatar

Posts: 325
Location: The shores of sunny Clapham
Liked others: 1
Was liked: 283
GD Posts: 484
Harleqin wrote:
TMark wrote:
Why don't you two take this to email? 171 posts, 162 made by only you two and I don't see any great interest in the subject by anybody else. Discuss it privately, and then produce your agreed version, which everybody else can then ignore.


No, why? I think it is very interesting to see this discussion. I believe that it is a major argument for rules that do not need this kind of discussion.


No, it is two trolls endlessly re-hashing stuff that interests maybe 2 people other than themselves. I have called Robert a troll before and I would justify it by the fact that he never stops posting stuff, the only purpose of which is to show how brilliant he is (in his own view) and which drives away anybody else who may show that he is not too brilliant.

Best wishes.

_________________
No aji, keshi, kifu or kikashi has been harmed in the compiling of this post.
http://www.gogod.co.uk

Top
 Profile  
 
Offline
 Post subject: Re: User-friendly Reading of the Japanese 1989 Rules
Post #186 Posted: Thu Jul 15, 2010 11:07 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
TMark, you confuse your insufficient imagination of motives and purposes with trolling. Research is free and not censored by your disliking.

Top
 Profile  
 
Offline
 Post subject: Re: User-friendly Reading of the Japanese 1989 Rules
Post #187 Posted: Thu Jul 15, 2010 11:15 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
Quote:
There are NO redefinitions of J2003-terms.


Maybe I use definition in a broader sense than you: Something determined by an algorithm can be also considered to be defined by it.

Let me express it a bit differently: Given the definitions in J2003 and the algorithms, I want to see propositions proven of this kind: "The algorithm A determines the strings fitting definition D."

Top
 Profile  
 
Offline
 Post subject: Re: User-friendly Reading of the Japanese 1989 Rules
Post #188 Posted: Fri Jul 16, 2010 10:49 am 
Lives in sente
User avatar

Posts: 1308
Liked others: 14
Was liked: 153
Rank: German 1 Kyu
RobertJasiek wrote:
Quote:
There are NO redefinitions of J2003-terms.


Maybe I use definition in a broader sense than you: Something determined by an algorithm can be also considered to be defined by it.

Let me express it a bit differently: Given the definitions in J2003 and the algorithms, I want to see propositions proven of this kind: "The algorithm A determines the strings fitting definition D."

Why do you want to have "proven" something that is an inherent feature of the minimax-algorithm ?

The minimax-algorithm transports the leaves' statuses via branching point after branching point to the root of the variation tree.

In the end you have the root's status as the concluding status of the string's evaluation.

There must be at least one leaf, which has this concluding status.

Highlight the path between neighbouring branching points (including root and leaves), which habe this concluding status, and you will have marked everything that is "forced".

If the concluding status is A (string remains on the board), this is forced by the player. So it cannot be forced by the opponent. This equals the definition of "uncapturable".

If the concluding status is B2 (string has been captured, successor on local-1), it is forced by the opponent that there is no status A surviving. Thereafter the best status that can be reached from the player's standpoint is B2. This status is reached, so it is forced by the player that there is a successor. The opponent cannot force no successor. This equals the definition of "capturable-1".

If the concluding status is B31 (string has been captured, no successor on local-1, permanent stone on local-2\1), it is forced by the opponent that there are no statuses A or B2 surviving. Thereafter the best status that can be reached from the player's standpoint is B31. This status is reached, so it is forced by the player that there is a permanent stone on local-2\1. The opponoent cannot force no permanent stone on local-2\1. This equals the definition of "capturable-2\1".
Local-2\1 is part of local-2, so this equals the definition of "capturable-2", too.

If the concluding status is B32, it is forced by the opponent that there are no statuses A, B2, or B31 surviving. So it is forced by the opponent that the string is "not alive".

_________________
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: User-friendly Reading of the Japanese 1989 Rules
Post #189 Posted: Fri Jul 16, 2010 11:48 am 
Judan

Posts: 6087
Liked others: 0
Was liked: 786
I do not want to have proven that it is an inherent feature of min-max but that your algorithms determine exactly the strings they claim to determine.

For a proof, the tables representing decisions in your algorithms must be taken into account. Also I want to see a careful application of the definition of force to see why it would conform to your algorithms. The latter should be the easier part, almost trivial?

Top
 Profile  
 
Display posts from previous:  Sort by  
Post new topic Reply to topic  [ 189 posts ]  Go to page Previous  1 ... 6, 7, 8, 9, 10

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