Website founded by
Milan Velimirović
in 2006
21:06 UTC
| |
MatPlus.Net  Forum General SPAM in MPR |
| |
|
|
|
|
You can only view this page!
| | Page: [Previous] [Next] 1 2 3 4 | | (41) Posted by Dan Meinking [Friday, Mar 19, 2010 05:28] |
Here's a perfect example to demonstrate that HOW a problem is 'discovered' does not determine its worth:
P0549870
Michael McDowell & Edgar Holladay
=1st/2nd Prize
Chess Life 1985
 (= 3+4 )
h#4 2 solutions (3+4) C+
Solutions:
1) 1. Sc5+ Kf2 2. Ke5 Ke1 3. Kd4 Kd2 4. e5 e3#
2) 1. Dd8 e3 2. Sd6 e4+ 3. Ke5 Ke3 4. Df6 d4#
When this problem first appeared, columnist David L. Brown noted that he'd received the identical setting from the 2 authors independently. Quite a coincidence!
Let's assume:
(1) Edgar 'mined' this as described in my prior post.
(2) Mr. McDowell did NOT use any 'mining' techniques.
If these statements are true, then the same problem was 'discovered' by completely different methods. We conclude, then, that the METHOD of discovery does not influence the value of that which is discovered.
| | | (42) Posted by Steven Dowd [Friday, Mar 19, 2010 05:41] |
It is always dangerous to repeat "someone told me," but...
I understand one of the solving programs - maybe it was VSACH, has the ability to assist in the composing process, and that was the reason for some of the wonderful echo problems "found" by one composer.
Of course, what one is told and what one remembers....
I used Edgar's method to a certain extent when I started composing with varying results. I think he said he set up pieces on the board first, played around with them until he started seeing things (like the mates and stalemates Dan mentioned) and then started feeding things into the computer, shifting things around there.
It sounds to me like there is method in the madness there, although my problem as a beginner was of course assuming what was correct was also good. Edgar knew what was poor; I did not, although I had some pretty good guides who stopped me from doing too many stupid things.....
I am beginning to learn that the number of correct problems I DO NOT publish is probably a better indication of my (limited) talent as a composer. I am to the point where I sometimes have 20-40 versions of a problem, trying everything that works but not keeping it all for publication.
I've been moving towards working more systematically, with certain people helping me in organizing an approach similar to Petkov's PAPS system - which must be the polar opposite of Edgar's method.
The short story is that even with machine help, there is still composing going on for most of us. We may have better tools, but doesn't that encourage us to do more, not to be lazy? I know the increased solving capabilities of Gustav have encouraged me to look for the longest form possible with the most content. Not simply the longest form, but, and I repeat because this is important, with the most content. And I am only scratching the surface of its, and my, capabilities so far.
I do get the impression from the Roycroft article that he does think studies somehow magically appear from the tablebases. You have to have some idea before you begin composing - it may be a bad one, but a bad one is better than none at all. And you still work from there.
Sergiy came off very strong, very antagonistic here and it really disturbed me, I will admit, because I always thought of him as the rarity, "a great composer AND a great guy," but actually because he is such a great guy he had the cojones to challenge Roycroft, whom we all see as one of the most important figures in study composition. And once I learned the back story, I see why he did it.
| | | (43) Posted by Georgy Evseev [Friday, Mar 19, 2010 08:42] |
Well, my post was intended to avoid double standards.
One can ignore EGTB. In this case he has to equally accept any endgame, ignoring the way it was created.
Sergiy had no right to write
"The hunt for dual-free lines as the one that Georgy describes as his best result gives no prospects to a miner who intends to be a study composer. I don’t think this position is a study. Maybe it has some value for endgame theory, but a study requires more."
until he really analyzed the position I've shown. He definitely have not done it (as I expected - and this is exactly why I have not given any comments to variation shown).
Also, be ready that good programmer may become a renowned study composer. As I've said, it is reasonably easy to "mine" the positions with specific thematic contents. The much discussed example is the 1st Prize of Gurgenidze JT. For me it was obviously "mined", using special tools not available for everyone. But the author (who is not endgame specialist, BTW) did not explicitly said it and many study composers have found a lot of words to support him. It is normal, but they should use the same words of support when the author openly says that EGTB was used.
In fact, I think, study composers in the depth of their heart consider use of EGTB as cheating. But instead of saying this openly, they prefer to criticize endgames which are known to be "mined" without even seriously looking at them.
| | | (44) Posted by Jacques Rotenberg [Friday, Mar 19, 2010 10:09]; edited by Jacques Rotenberg [10-03-19] |
Georgy,
" In fact, I think, study composers in the depth of their heart consider use of EGTB as cheating."
Obviously you speak of yourself, dont't you ?
It is quite funny, btw, to see you speaking of "cheating" - and condamn it -, when in another post you were a defender of lying, stealing and cheating !
However, the point seems here to be as follow :
- some people seem to consider the value of a problem mainly on the scale of how hard it was to work it out - the price of the sweat - so to say
- some other mainly consider the result on the scale of how pleasant it is to be considered (I do so).
In any case the words used by John Roycroft are most excessive. The reaction of Serguei is in mirror.
What ? will you say that now, with the EGTB the Saavedra study has no more any value ??
Will you say, that in case it was an original of nowadays it should have been discarded ?
| | | (45) Posted by Miodrag Mladenović [Friday, Mar 19, 2010 10:22] |
Jacques wrote:
QUOTE
It is quite funny, btw, to see you speaking of "cheating" - and condamn it -, when in another post you were a defender of lying, stealing and cheating !
Can you be more specific? I am following up on all messages and I never saw Georgy defending lying, stealing and cheating.
| | | (46) Posted by Georgy Evseev [Friday, Mar 19, 2010 10:48]; edited by Georgy Evseev [10-03-19] |
No, I do not speak of myself.
I am for example speaking of Sergiy, who denied any value of computer-found position without analisys. The composer who knows what a hard work is does not want to believe that anyone can achieve similar results automatically or semi-automatically. The easiest way to do it is to deny the quality of any such results.
I cannot say anything about Saavedra study (the history cannot be changed), but I have a simple question to you. Are you ready to objectively evaluate the position from my post 30?
Until now I have two opinions from study composers (Rasumenko and Didukh) - both independently say that it is not a study. I consider this opinion wrong and conclude that they do not _want_ it to be a study. I can tell you that I was myself shocked when I found how interesting the contents of this position really is. I'll wait for any answer from Sergiy and then write my own comments to the position from the point of view of positively minded commentator.
| | | (47) Posted by Sarah Hornecker [Friday, Mar 19, 2010 11:25] |
Here's the third opinion: It is a genuine study. The final zugzwang position is nice (although it seems to be no mutual one, reducing paradox).
| | | (48) Posted by Jacques Rotenberg [Friday, Mar 19, 2010 12:26] |
@ Miodrag :
http://www.matplus.net/pub/start.php?px=1268997661&app=forum&act=posts&tid=501&fid=gen&page=5
106 (point 6)
113
114
| | | (49) Posted by Marek Kwiatkowski [Friday, Mar 19, 2010 16:24]; edited by Marek Kwiatkowski [10-03-19] |
In my opinion, this discussion has a sense, if it shows all aspects in reliable way.
Above, I mentioned about the Shredder's page, where it is very easy to get (online) scores and solutions of all 6-men (<=) positions.
Lots of studies were earlier tested by weak engines, like Fritz 8. An effect of this is visible in the latest Fide Albums.
Now, we have a very strong engine – Stockfish 1.6, which has algorithms to play endgames without tablebases (EGTBs). This free (!) engine is probably the best tool to analyze (solve) studies.
The following example shows the huge Stockfish’s possibilities:
??
 (= 3+5 )
win
Stockfish can quickly (a few minutes on PIII) find: 1.Qc7!! Rbc8 2.Qh2+ Kxg7 3.Qb2+ Kh7 4.Qb7+ Rg7 5.Qxc8 Rf7 6.Qh3+ Kg7 7.Qh6+ Kg8 8.Kxg6 +- (if 1.Qh1+ and 2.Qxa1 black plays 8…Rf6+! =)
| | | (50) Posted by Gerhard Josten [Friday, Mar 19, 2010 17:36] |
This study recently has been awarded:
 (= 4+3 )
White to play and draw
Solution:
1.c6! Rxh7+
Now the EGTB are reached and White draws.
Is it an art to discover such positions and to add a short introduction? However, the general problem is the existence of judges and not that of EGTB. De gustibus non est disputandum! Please imagine that judges would have to decide between a Pablo Picasso and a Leonardo da Vinci. Without judges we wouldn't discuss on their decisions and enjoy those compositions which touch our hearts.
Which is now the origin of awards in the field of chess? If we for once ignore the Arabic and medieval bet games, because they were solving competitions, then shortly after the London chess tournament from 1851, a turning point in the chess history of the modern period, for the first time we meet the concept of a beauty prize. It was awarded in 1876 to the Englishman Henry Edward Bird for his game against James Mason in New York.
One didn't have to wait very long for the first composition tournament. After Manchester 1857, Birmingham 1858 and Cambridge 1860 there followed the more well-known London chess tournament of 1862. The 1st prize was awarded to Conrad Bayer (direct mates), Arnold Pongracz (selfmates) and Bernhard Horwitz (studies).
When will this relic of former times finally be abolished?
| | | (51) Posted by Marek Kwiatkowski [Friday, Mar 19, 2010 17:47]; edited by Marek Kwiatkowski [10-03-19] |
 (= 2+3 )
searching by Stockfish (P4 2 threads)
depth 27; score +0.56; time 01:00; nodes 91267k;
Rd1 Nf3 Ke4 Ng5+ Kf5 Nh7 Rg1 Nb3 Ke6 Nc5+ Kf7 Nf8 Rg5 Ncd7 Rg3 Ne5+ Kxf8
Faster than by EGTBs, only 1 minute!
| | | (52) Posted by Sergiy Didukh [Friday, Mar 19, 2010 20:12] |
Again from the very fresh start.
1) The detailed thematic search (as Gerhard suggests and suspects) requires that every stored position gets the tree of moves (the main line … and variations). Only then CQL or something alike can be applied. Have you ever imagined the size of data? I’m not sure how long we’ll have to wait for results: a hundred thousand years or only one thousand.
That’s why I say that ‘mining’ interesting positions is impossible. Gerhard, this is fiction. So, 6-piece studies are the result of a composer’s creative work (or luck) and there’s nothing wrong with awarding them.
2)Another unsubstantiated accusation: some composers develop introductory play (build up dual-free lines) using EGTB and different tools. I hope you agree that this has nothing to do with the process of mining described above. Everyone is invited to try to develop something artistic using this method! Oh, this time I suspect that it’s going to be very-very hard. You see, I can suspect too. And I am almost sure that some composers (for example, C.Poissson and J-M.Lousteau) do it with very limited material. Once I saw C.Poisson’s artistic study. Probably, one hundred or more of his previous attempts (printed around the world) failed. They were soulless mechanical constructions!
3)My comments and actions never showed that the method how this or that study was created could influence my evaluation of it. I consider the result on the scale of how pleasant it is to be considered, like J.Rotenberg does. So, Georgy's conclusion on my account is absolutely wrong and probably it is based on my 'praise' of his 5-man piece. Georgy, I don’t like the 1st Prize-winner of the Gurgenidze JT either (the idea is good but badly developped), but your effort looks bad to me too. There’s still a chance that I missed some points in it. Did you mine it, find it by chance or develop it using EGTB? Siegfried, thinks that it is a study. I often criticize him too for making something like that. But he is still my friend.
| | | (53) Posted by Jean-Marc Loustau [Friday, Mar 19, 2010 20:35]; edited by Jean-Marc Loustau [10-03-19] |
Allow me to write some words a little out of topic because Georgy (and just now Sergiy) speaks about a study which I am the author (Gurgenidze JT 1st pr) and for which he makes false assumption: Yes of course I used EGTB, but No, I had no other tool used to produce this study, just EGTB, and no search tool. Oh yes… I used my mind, I mean I worked in organized way, with method… Let me speak about my experience…
I first found by "manual" search some interesting elements which seemed to be worthy of being presented in a study (ie 1 of the stalemates and a positional draw): "to be worthy" means "beautiful" to my tastes. After some work (I cannot quantify any more, but at least several hours), based on retro-game but also mainly on choice of some lines of play rather than others, I have almost reached the published position: one might call this phase "building phase" for there are choices to be done, based on criteria of personal taste (examples of possible choices: choice of the longest line, or choice of the “prettiest” moves, or choice of some echoes, or choice of simplicity and clarity, or on the contrary difficulty [often not a good choice I think in this context], and so on: my own choices are first based on what I consider to be beautiful to my tastes, thus deeply subjective). This is probably news to those who have never used seriously EGTB: you can still sometimes (I would even say always if well used) speak of "building".
After identifying this position I went to a 2nd phase, which I consider mandatory: understanding of the position. This 2nd phase of understanding took me about 300 hours, which is much more than the average time that I take to compose a problem. If I could spend this time is because I was "lucky" to have at this time a serious illness that forced me to quit my job. This does not necessarily mean that 300 hours are needed to understand the study, fortunately not. This time was for identifying some basic principles of the material used which is somehow quite "new," and also of course for understanding lines of play themselves. I must say that if I recognize that the published solution does not allow a satisfactory understanding (by the way this writing is not completely mine, and the first introductory line was in my own writing only a by-play variation); I attached when I sent the study to the tournament a synthesis of my analysis in 3 or 4 pages (never published of course, the lack of place, this is a generic difficulty for studies requiring long explanations) which should allow a first level of understanding quite satisfactory for a reader; moreover today I have produced several studies with this specific material and I'm much more comfortable with, and I think that I should take probably between 10 and 20 hours for a very good understanding of this study and writing a synthesis for what I call a 1st level understanding. Some composers (as Sergiy I suppose) can estimate that if 3 or 4 pages are required for basic understanding, then the study is "analytic", which is not for some tastes; for me the analytical aspect is not incompatible with beauty, but this is purely subjective opinion, and anyway the subject is not there; by the way, when Sergiy says the idea is badly developped, it's also a matter of tastes: I am very sorry he doesn't like the study, but the fact David Gurgenidze liked it is to me a big joy...
I read here and there about the merit of the composer? I am not interested by this approach, I do not claim any merit, not more for this study as none of my problems. I'm just saying that the work given for this position was around 300 hours for understanding (but probably 90% for a better understanding of the material and 10% for the position itself) and perhaps a dozen for “study identification” and “building”; which certainly does not mean that I have any merit, I just wanted to deepen my understanding of chess, and this is also one of the strengths of EGTB for those who want to do this work, the possibility to progress in the knowledge of chess. The difficulty that arises here is that this material has not been studied and there is nothing written to facilitate understanding; perhaps I will write an article once to approach it with more ease, and I also intend to reproduce some of my studies with notes to address them more easily (which I did for 2 of them in Phenix) but, sorry guys, I am not sick any more and I'm too busy with my job today, so it will be certainly not before 1 year or 2.
Sorry for spamming the forum with my personal case.
And by the way, Georgy: surely I am not an “end-game specialist”, at least as a composer, but I am nevertheless interested in endgames for many years, and I have a quite rich library in this field!
| | | (54) Posted by Sergiy Didukh [Friday, Mar 19, 2010 22:27] |
Jean-Marc, thanks a lot for this presentation of work at your study! It demonstrate well the new method of development of an idea using EGTB (mostly applied when working at difficult ideas, I think). You described it with right words: ‘building’ plus ‘learning’ and ‘understanding’.
You also guessed right that analytical lines don’t appeal to me. My motto is ‘never let analytical lines spoil the main idea’. This artistic criterion has been established long ago but nowdays it is most actual. Composers use powerful computer programs and other tools like EGTB in the calculation of lines and they often forget this principle. Software and composing methods are not guilty here at all. Only composers’s tastes and their technique can be blamed. The new method is perfectly legal but it is hard (as Jean-Marc shows) and it often doesn’t let obtain satisfactory results.
| | | (55) Posted by Jean-Marc Loustau [Saturday, Mar 20, 2010 09:57] |
@ Sergiy: In the work to be done during intensive use of EGTB there is also the "presentation", ie show (isolate) the main play(s) in a way attractive by emphasizing the "ideas" or themes, and supply apart for those who want more detailed explanations (and by the way if there is no need of any explanation the use of EGTB should not be a problem!): this work is also quite complex (and there also comes a degree of subjectivity). I also think it could be that in the future (not tomorrow) the aesthetic principles of appreciation change, which would not be surprising because we are dealing with a new field of investigation ... You speak of "new method": it would not be surprising that different methods lead to different criteria. It is also obvious(at least to me) that all methods are valid and respectable, and one of the difficulties today is that the results produced by different methods are evaluated (judged) together (If there were not this need to judge and compare, there would be no difficulty, everybody publishing what he wants without bothering others).
It is conceivable in contests or album Fide different sections (A: EG mainly produced with EGTB, and B: traditional studies or using EGTB marginally) but the difficulty lies on the border, meaning that studies which have one foot in the 2 sections are very numerous and the choice of this section or this one is subjective and highly debatable. That reminds me, the time (some 50/60 years ago) when helpmates, selfmates, retros and all fairy problems were judged together in the same section! So I have no solution ... and perhaps also it is too early; may ba also am I wrong...
| | | (56) Posted by Sarah Hornecker [Saturday, Mar 20, 2010 10:50] |
QUOTE
Siegfried, thinks that it is a study. I often criticize him too for making something like that. But he is still my friend. (SID)
Actually, my girlfriend also often tells me I'm silly (humorously, I guess). But I guess that doesn't belong here. :-)
So what makes this a study in my opinion is that there is a unique line and some artistic content. Surely, it might be easy to understand with EGTB but the content is not the EGTB analysis (which also might or not be possible with human analysis - here I guess it might). The study consists of the themes "refusal of capture" and "zugzwang", having (using the Friedgood/Levitt GPFD theory) in my opinion geometry, depth and a bit of paradox.
Something seems to have went wrong here, too. I was talking about the study with R vs two knights but now everyone seems to talk about a different study? I consider the R vs BBP studies of Jean-Marc more as (important!) theoretical endgames, also due to their difficulty.
QUOTE
Sorry for spamming the forum with my personal case. (JML)
That is no spam! At least not if you don't offer us composing-enhancing drugs for 20 percent off. :-)
| | | (57) Posted by Sergiy Didukh [Saturday, Mar 20, 2010 10:53] |
I must dissagree that the presentation of the solution with isolated main line could emphasize ‘ideas’ or ‘themes’. All the lines of the study should be in harmony and if they are not given it doesn’t mean that they don’t exist. They are a powerful means. They can ‘sing praises’ to the main idea of a study, and at the same time they can ‘bury’ it in the granite coffin. A perfect study is a story where every line has some role in the plot. I hope artistic principles won’t change much.
It is also wrong to divide studies depending on the use (marginal or intensive) of EGTB (or chess programs). If somebody wants to make a division he has to take into account only the final product: artistic studies and analytical studies. It doesn’t matter how many pieces they have or what tools were used. This section for analytical studies can contribute to the development of the endgame theory. But I’ll stay forever with artistic studies.
And rethinking ‘the new method’ I am now inclined to believe that it is not that new. I sometimes compose approximately the same way but use chess programs instead of EGTB for the preliminary evaluation of prospective lines and then turn on the creative hunt for ideas to be inserted. I can only suspect that some modern method is used by C.Poisson because his output features the battle of two knights against two bishops. So, I’d like to hear if he indeed downloaded TB for this material and creates dual-free lines with Wilhem. But I doubt that someone really likes such studies, if they can be called that. That's why this different method looks doubtful even if it is applied.
| | | (58) Posted by Sergiy Didukh [Saturday, Mar 20, 2010 12:05] |
Jean-Marc, you also often use limited material: R vs BB+P. Can you swear that you don’t use TB + ‘Wilhem’ for this material so as to build-up lines (I don’t think that C.Poisson will reply and I’d like to have some reliable answer to my question, just for the record). You say that you got the initial ideas for your study after ‘manual’ search. Do you mean you found them by chance, that is, you placed the pieces on the board and then surfed through the lines? Why did you choose this material?
| | | (59) Posted by Gerhard Josten [Saturday, Mar 20, 2010 13:09] |
@ Jean Marc:
There would be a simple and absolutely certain way to differ EGTB studies from EGTB-free studies by building two categories: A) Studies ending with less than 8 pieces and B) studies ending with 8 or more pieces.
@ Sergiy:
You distinguish between artistic studies and analytical studies. Do you think that the last ones can’t be regarded as an art?
| | | (60) Posted by Juraj Lörinc [Saturday, Mar 20, 2010 13:53] |
Gerhard, this line (less than 8 vs. at least 8) is only temporary, with every new tablebases it would change. Therefore bad criterium.
| |
Read more... | Page: [Previous] [Next] 1 2 3 4
MatPlus.Net Forum General SPAM in MPR |
|
|
|
ISC 2024
