|
| Page: [Previous] [Next] 1 2 3 |
| (41) Posted by Hauke Reddmann [Monday, Aug 21, 2023 22:10] |
Of course you know that for a mathematician "main line" doesn't exist -
Black loses in each line, what is the point of choosing a particular? :-)
I.e. "main line" can only be defined from an esthetic standpoint.
|
|
| (42) Posted by Mu-Tsun Tsai [Tuesday, Aug 22, 2023 02:00] |
@(40)
He also mentioned that such an idea is controversial.
@(41)
> "main line" can only be defined from an esthetic standpoint
In other words, it's subjective, and I can't possibly please everyone.
What if I allow the user to manually suggest a mainline? Would that help?
|
|
| (43) Posted by Kevin Begley [Tuesday, Aug 22, 2023 02:26] |
Thanks for providing this translation, Joost.
|
|
| (44) Posted by Hauke Reddmann [Tuesday, Aug 22, 2023 08:29] |
@Mu-Tsun Tsai: Of course I trolled a bit :-)
Analyzing a position is not unlike doing a mathematical proof.
You have some axioms: these are mate, stalemate and (forced) repetition
(50 moves *could* be one but was willingly excluded).
You have lemmas: Q wins against R, except <case list>
To prove these, there are for example "theoretical studies"
(it's long, there are finesses, it's not unique, ...).
Or maybe the win is utterly trivial (say Q v B) and
"left for the reader".
You have your position and make a case analysis:
see, Black has no more move to hinder White queening
and going into Q v R. Subline stop, qed.
We can check for the length of the subline and how much
tactical flashbang goes off. I think in 90% we get
all sane study fans to agree on the subline. If not,
this *will* be an issue in the comments (assuming the
study was in an informal and/or a prize report).
|
|
| (45) Posted by Viktoras Paliulionis [Tuesday, Aug 22, 2023 13:38] |
The tool could have several analysis modes depending on the possible use cases.
1. Simple search:
The user inputs a position and wants to know what the result is. He can choose any branch and go to a deeper level. This mode is analogous to other known online EGTB search tools.
2. Study solving:
The user inputs the position of a study and expects to get a complete solution, consisting of one main line or several variations. Ideally, the solution would completely coincide with the solution indicated by the author of the problem. If the tool cannot identify the main line, it provides several candidate lines. The mode is intended for solving correct studies that do not have duals in the main line.
3. Checking soundness:
The user knows the solution of the study, but wants to check its correctness, i.e. whether the solution is unique and have no duals in the main line. This mode would be useful to check old studies or for judges to check that the solution presented by the author is correct.
4. Extended search:
Checks two levels. For example, if the user selects a White move leading to a draw, the tool displays all Black moves leading to a draw and additionally provides several White's responses for each Black move.
If after some move by Black there is only one response, then we will know that the solution is unique in this branch. This mode is suitable for assistance in both solving a study and checking its correctness.
|
|
| (46) Posted by Mu-Tsun Tsai [Wednesday, Aug 23, 2023 03:14] |
@(45)
Sounds like a great direction. I believe I will work on something like that after I finish the other tasks at hand. Meanwhile we can continue brainstorming.
|
|
| (47) Posted by Viktoras Paliulionis [Wednesday, Aug 23, 2023 13:27] |
I think the 1st and 4th cases are quite easy to implement. 2nd is the most complicated. At the beginning, it would be enough for the "study solver" to find an author's solution in a one-line study.
To create a good study-solver, you need to have more knowledge about this genre of chess composition. The solution of a study is not only a unique sequence of moves. Usually, the solution demonstrates some study idea or a combination of several ideas. The number of ideas (study motifs) is finite. They are all more or less known and can be classified in different ways. If the sequence of moves doesn't show any idea, it's probably not a study solution. Such continuations do not make any impression on the solver from an aesthetic point of view.
|
|
| (48) Posted by Mu-Tsun Tsai [Thursday, Aug 24, 2023 09:55] |
@(47)
Makes sense, only that I am far from capable of implementing such a thing. Always feel free to contribute and make GitHub pull requests though!
|
|
| (49) Posted by shankar ram [Wednesday, Feb 21, 2024 18:38] |
Mu-Tsun Has made a further enhancement: Fen-tool now supports reading and writing .olv files of Dmitri Turevski's Olive!
Edit (22-Feb-24): The user had to do the fairy piece mapping. Mu-Tsun has further improved fen-tool so that the mapping is handled automatically!
|
|
No more posts |
Page: [Previous] [Next] 1 2 3
MatPlus.Net  Forum General FEN Tool for chess composers and solvers |
ISC 2024
