Website founded by Milan Velimirović in 2006
23:35 UTC
| |
MatPlus.Net Forum General All issues of StrateGems available online |
|
|
|
You can only view this page!
| Page: [Previous] [Next] 1 2 3 | (1) Posted by Kostas Prentos [Tuesday, Oct 4, 2022 07:21] | All issues of StrateGems available online Visit the StrateGems website (http://www.strategems.net/) where all issues from 1 to 100 are now available for download. For some issues (23, 68, 71, 74) the page order is set for printing (n, 1, 2, n-1, n-2, 3...) but I believe this will be corrected soon. | | (2) Posted by Rajendiran Raju [Tuesday, Oct 4, 2022 12:48] | Generous Offer and Great Benefit to Problem Community ..Thanks a lot..!! | | (3) Posted by seetharaman kalyan [Wednesday, Oct 5, 2022 12:27] | Really great offer and very quick completion of a huge task. Thanks | | (4) Posted by Hauke Reddmann [Wednesday, Oct 5, 2022 21:47] | Archiv Binge Alert!
Can some StrateGem oldtimer suggest some must-reads? | | (5) Posted by Peter Wong [Thursday, Oct 6, 2022 08:10] |
QUOTE Can some StrateGem oldtimer suggest some must-reads?
SG-58 has the article 'The quest for a King-only proof game' by François Labelle. It covers (among other things) the discovery of the 2-kings-only PG with a unique solution, which was published as an original in the same issue.
BTW, this instant classic was dismissed by the judge as "lacking artistic merit" and "only of interest to programmers", and didn't feature in the Award at all (SG-65). Fortunately, it was placed 2nd in the Wenigsteiner of the Year Award. | | (6) Posted by Kevin Begley [Thursday, Oct 6, 2022 18:46] | @Peter,
That article can also be found online: https://wismuth.com/chess/2kings.html
In my humble opinion, it is a shame that did not win First at Wenigsteiner. In all the decades of that contest, I highly doubt they received a more profound (or a more economical) entry.
As for the SG judge who got this SO INSANELY WRONG... sadly, that was a dear old friend who made that tragically bad call.
I don't want to pile on (I hope he returns to problem chess with the intention of fully redeeming himself for that mistake -- and I fully believe in him, if so), but I must admit there were many other tragedies in his awards (which were bursting with declarations -- how can I say this nicely? -- lacking the faintest trace of the humility and reverence that a good judge requires).
LOL. He used one award as an anvil to hammer an entire fairy condition, writing: "In my opinion, this condition [Duelist] does not make for interesting PG’s."
That was that -- as if the Supreme Court ruled "Duelist is inherently incompatible with Proofgames" (leaving no opportunity for appeal).
In his defense, I believe he made that nonsensical statement prior to Caillaud's PG23.0 (in 2019) -- a problem which undeniably puts his claim (which again, appeared in an award!) to shame.
But, even before that, he should have been cautioned that an award is no place for strutting such cavalier pronouncements.
This is a fine temperament to stir debate in a problem forum, but it has no place in an award.
When you see a judge scolding composers how not to displease his majesty, you'd think somebody would say something, but it takes a great deal to spark a needed revolution.
I would gladly give François the commendation I received from that same award -- and only regret that I failed to hand over a higher distinction (I was quite lucky to receive a commendation, frankly) -- except that it would be an insult to suggest he should accept a mere commendation.
The commendation I received has become an insult, and I should have the right to reject it.
I worked heavily with a computer to create my entry. What's the difference? What could justify awarding the mere user of a solving tool (who needed only a few hours), and discard the composer who programmed his own tool (which needed months to achieve this iconic masterpiece that nobody had managed before)?
The judge claimed in the award, "I solved all of them except Kevin’s, which I had already seen."
I can attest that the judge is a very talented solver, yet I highly doubt he actually solved the massacre proofgame. Had he solved this, had he even made an effort to solve it, some mention would have been made (and a greater appreciation might have been displayed).
When he says "... I discarded P0330 (Labelle) and L4 and L5 from Labelle’s article," it suggests to me the problem was never considered (an abdication which should have spawned an appellate process).
It is an understatement to suggest there were grounds to discard this judgement.
Note too that this judgement closed with the ominous, "See you all next judgment!" (which either reads like a boast of unbridled power, or a dare for anyone to challenge this decision).
Frankly, we need somebody who will judge our judges. Maybe their capacity should be evaluated by some outsider (to whom we could all voice those concerns we prefer to keep to ourselves -- knowing we may see that judge again and again). Many judges need training to remove an illogical bias against computers.
Francois hit a grand slam home run, and the umpire kicked him out of the game.
Even baseball never had umpires this bad (and they were compelled to evaluate their umpires).
Somebody really needed to say something, but nobody did.
I was in an ideal position to say something (I might have spoken to the judge, as a friend, when I became aware of the error in his judgement), but I confess that I said nothing.
That's the part that haunts me most. | | (7) Posted by Hauke Reddmann [Thursday, Oct 6, 2022 20:59] | Who judges the judgemen?
At least: time the avenger. | | (8) Posted by Juraj Lörinc [Thursday, Oct 6, 2022 21:20] | Peter, thank you for bringing this find by Francois forward. Of course, I remember it as one of the WJP judges, now I have looked up the table to recall the details. As I see, two best problems in year 2012 were much ahead of all the rest, both getting at least 60% of possible points (that's a lot in this joint award), both being selected by 5 judges out of 7 into their top 5. RJM's problem got just 1 point more than FL's. Curiously, RJM's problem was even more economical, having no piece on the board. Everybody can check year 2012 here: http://www.wenigsteiner.de/ | | (9) Posted by Kevin Begley [Thursday, Oct 6, 2022 22:25] | For clarity, I don't consider it a "shame" upon Wenigsteiner for not awarding that problem first place (note: they did bestow a very honorable distinction upon it).
I just think it's a shame it didn't win first place, because I consider this problem to be the most accomplished entry in their history -- and yes, their most economical (a two-man problem can be more economical than some one- or zero-man problems).
But, overall, I think Wenigsteiner does a very fair job.
To its great credit, it was established with MANY judges (such that the bias of one judge is never so injurious).
Of the seven Wenigsteiner judges, only one *might* have discarded the problem entirely (another gave it only reserve placement in the award).
I stress might, because we don't know whether it was discarded (or perhaps that judge only deemed it worthy of 7th place).
It would be interesting to read a short statement from each judge, immediately after the final verdict is known, as to their opinion on the final placement (explain why they overlooked problems that their peers favored, for example). It might be revealing to allow the judges to account for their scoring (especially if this is provided immediately after the fact).
In general, I suspect the plurality of judges gets most things right -- but, who am I to say? The future will render the ultimate verdict on our problems and our judgements.
We can definitively say that anti-computer bias did not show in this award, as three of the seven judges deemed this problem worthy first place, and one deemed it worthy of second place (it's a miracle another problem could overtake such an impressive scoring).
At the time, Juraj thought it deserved fourth place, and gave the eventual first prize problem nothing.
I consider Juraj to be a fair judge (for what that's worth), but in this instance, he didn't agree then with what he now suggests (only in hindsight did the two problems stand above the field).
We have a mechanism to evaluate the judges of Wenigsteiner:
The judge who most often aligns with the final placement by their peers will tend to be the better judge (first take away the impact of their judgement, then score them).
While this will not account for systemic failures, and specific bias would only show after a protracted test period, comparison against their peers can be a very useful mechanism to evaluate the merits of even a good judge (at the risk of complimenting Juraj, lol, I suspect he would do very well overall in that scoring).
Further, we may score them on their capacity to predict how future judges might score the given problem set (future judges can be seen as more reliable).
@Hauke, I agree -- time must be the avenger!
It's one thing to debate the subtle differences between first and second place. I have my own opinion, and there's no way to be certain I'm right about that. That's a defensible call.
It's quite another when we witness how the unfettered bias of a single judge can completely discard a masterpiece. That's not defensible (and we can not allow bad solo judges to continually get away with that abuse).
We can not appeal every judgement that doesn't go our way (and few will appeal the decisions they found excessively favorable), but there are some judgements that are so misguided, so injurious, they beg for some remedy.
The best remedy I can suggest would be to provide some evaluation of judges (put some questions to them, and ask them to account for some troubling discrepancies with their peers). | | (10) Posted by Joose Norri [Friday, Oct 7, 2022 09:54] | "BTW, this instant classic was dismissed by the judge as "lacking artistic merit" and "only of interest to programmers", and didn't feature in the Award at all (SG-65). Fortunately, it was placed 2nd in the Wenigsteiner of the Year Award." (Peter Wong)
The judge's view can fully be defended. The fact that something is unique, well nigh impossible to do etc., does not necessarily mean that it's problem art (and please let's not try to define that...). And there is no contradiction, a problem - if this even is a problem - can have merit in one competition but not in another. | | (11) Posted by Olaf Jenkner [Friday, Oct 7, 2022 10:44] | Could anybody translate this into German? I don't understand Kevin's posting.
First line:
"I don't consider it a "shame" upon Wenigsteiner for not awarding that problem first place"
Second line:
"I just think it's a shame it didn't win first place"
??? | | (12) Posted by shankar ram [Friday, Oct 7, 2022 11:46] | It's NOT a shame on the WJP...
means they are not to be blamed. They don't have to feel any shame.
It's a shame...
means it's a pity, or unfortunate (no blame being assigned to anybody!) | | (13) Posted by Olaf Jenkner [Friday, Oct 7, 2022 13:01] | Thanks, I understood, although it's not in German. | | (14) Posted by Neal Turner [Friday, Oct 7, 2022 14:11] | There's a great debate to be had concerning the relationship between problem chess and computers, but I'm not going to enter it.
What I will do is express my sense of relief that the Deepmind people didn't turn their attention to chess problems when they were developing their Alpha systems.
Playing the game of Chess is much more difficult than making chess problems, but they managed to get their AlphaZero program playing at a superhuman level after letting it run for just one day.
In another scenario they could have had a million two-movers sitting in memory the following morning.
Then the relationship between problem chess and computers would have been akin to that between Bambi and Godzilla: https://www.youtube.com/watch?v=c4oUuRBnEkE&ab_channel=ReadingMovies | | (15) Posted by Juraj Lörinc [Friday, Oct 7, 2022 19:09] | Another must read is the article by GM Petko Petkov on hs# in SG33, "Help-Compel Mate – A Stipulation of The Future". 17 years ago he summarized previous history and development of the genre up to that point, with some suggestions for further development. Since then, the genre did not stop its progress, a lot of fantastic new works appeared and the flow is much stronger than ever. For me the key moment when I got interested for hs# was just a few years earlier: although I have tried hs# in 90s, it was just curiosity, then during Portorose congress in 2002 late GM Bo Lindgren has shown me a few of his hs# and I finally understood this something different from both h# and s#. I am now in the camp mentioned in Petko's article as understanding hs# much closer to h# than to s#, but the difference between h# and hs# is crucial anyway. | | (16) Posted by Kevin Begley [Friday, Oct 7, 2022 20:22] | Of course the mode of play in a hs#n is predominantly Helper (not Selfer).
But, it's important to recognize this is only by the default convention.
Help-(=)-n (aka Helpstalemate) == Help is the mode of play, with Stalemate as the aim (and the aim is a terminating aim, requiring 1 ply).
Help-(Self-#)-n (aka Helpselfmate) == Help is the mode of play, with selfmate as the aim (and the aim is itself a recursive aim, requiring two ply).
Many, many years ago, I had some email conversations with Petko about this, and other stipulation possibilities, where I suggested the inner aim might be changed from s# in 2 ply (which is a default helpself-aim problem), to s# in 4 ply (or 6 ply).
I later discovered that my nested aim approach had "predicted" problems that were already published.
If you want to glimpse the "stipulations of the future," I encourage you to think about stipulations in a nested (or recursive) fashion.
I found that was the simplest path to a MUCH greater understanding (and I dare say it is a better predictor than even the great Petko).
By the way, a Self-(Help-#)-n (aka Selfhelpmate?) == Self is the mode of play, with helpmate as the aim (and the aim is itself a recursive aim, requiring two ply).
If you work out what that means, you will discover something astonishing (as I did MANY years ago).
Self-help-mate perfectly describes the semi-reflex-mate stipulation, and there's no need for a "reflex" fairy condition in semi-r(aim) problems, whereas a fairy condition can not be avoided in reflex(aim) problems.
In a help-(s#)-n, both sides help to achieve a position with a subproblem (white is on the move, and white has selfmate in 2 ply).
In a self-(h#)-n, white compels black (who resists the goal) to achieve a position with a subproblem (white is on the move, and white can help a helpmate in 2 ply -- what we would call a h#1 for black).
This perfectly describes the semi-r#n, and there is no need for the fairy condition ("reflex-goal") because black's motivation (black's mode of play) changes from self-play to help-play (just as the mode of play changes in hs#n problems), rendering the reflex condition redundant -- as redundant as if you added a fairy condition compelling black to give mate in 1 move, then stipulated a directmate (of course black is already so motivated, because this avoids white achieving the aim).
The semi-r#n can be described without any fairy condition, but this can not be achieved in r#n problems (because black will have an opportunity to provide white the chance to checkmate the black King, and that compulsion is contrary to white's goal -- thus, you can only achieve this by changing the rules of the game).
We have such a poor classification system, and such a poor understanding of the elements of a standard problem stipulation, that nobody even noticed that a semi-r#n is heterodox, whereas a r#n is not.
And we act like our failed system is too big to fail.
It was blasphemy to speak this truth years ago, and to this day, these terms remain completely misunderstood (even by the top composers, and the entire "classification project" pretense -- which to this day fails to provide a primer allowing anyone to decide what is a stipulation and what is a fairy condition when a new fairy element is introduced; it amounts to beautiful visuals of a taxonomical failure).
If you want an honest taxonomy, you must begin with honest definitions describing (and clearly disambiguating) the inner elements of a chess problem (including the stipulation).
If you want an honest classification project, it must provide any newcomer the ability to logically discern (by well defined terms, not by lookup table) what is a stipulation and what is a fairy condition.
If your definition says series-movers are a stipulation (that the stipulation is where move alternation is imposed), but you do not have a stipulation for progressive chess (nor for that other popular rules variant called Maharajah and the Sepoys), then you're frankly not serious about understanding the elements of a chess problem (and you can not classify what you refuse to understand).
I hope that all translates nicely into German, because a detailed explanation of these matters could fill a tome (for the benefit of those who seek to deny the obvious).
But, if you start by accepting the fundamental definition for what is a "fairy condition" versus "mode of play" versus "aim", this will inexorably lead you to appreciate the descriptive and predictive power of the nested aim. You can work this all out yourself (as I did), and suddenly you'll find yourself describing stipulations even Petko did not foresee (and you'll discover many of these stipulations were already published).
Petko can predict what Petko can predict (to his credit, that can not be undervalued or overstated; nevertheless, it is limiting).
But, Nested aims would enable anyone to predict stipulations of the future, that neither I nor Petko could imagine (you can legitimately invent novel stipulations, and have a system that will describe them in a standardized fashion).
If you want a full understanding what elements underlie the subatomic chess problem, just start by recognizing goals are aims, and combinations of aims, that may be recursively described by higher level goals (a sub-problem can function as the aim of a larger problem).
Whether or not our forever backward-looking community is ready to adopt better standards, this is a system which benefits every enthusiast with an eye to the future.
This is how you want to understand chess problems. This is how your problem database wants to search. This is how your solving tool wants to be tasked.
Ultimately, this is how an Artificial Intelligence wants to think about a chess problem.
It goes without saying that there are considerable benefits to a standard stipulation -- something capable of describing as many problems as possible (thus, we want to move from PGn to help-dia-n).
The stipulation must describe who moves first, who is tasked to achieve the aim, and when the stipulation stops (for example, if the inner aim is s#1, perhaps of a help-self#n problem, the default is to play through to all nested aims, down to the checkmate; but, the composer may want to task the solver to merely reach the compelled mate position at the final coup of a s#1 subproblem, and not play down to the mate itself, thus altering what moves should be seen in the solution, and what moves might be considered a dual).
The composer has every right to describe these conventions (e.g., what constitutes a dual), and nobody else (not even a group of GM composers) has any right to impose their conventional understanding of aesthetics upon the artist.
The good news is: there's a way to achieve ALL this, and more, with a standardized problem stipulation.
The bad news is: people don't want you to hear the good news. The will tell you I can not be understood. I used too many words (the same way Mozart used too many notes).
This is the future. There's no getting around it (it can only be delayed).
If Moses came down from the mountain today, he'd need only one commandment: thou shalt not halt the future (maybe Bob Marley was our Moses -- he said it better).
That's as true for valuing Artificially Intelligent composers as it is for recognizing it's long past time we adopt an honest (logical) classification system.
The reasons we need such a commandment are clear: people don't want a logical system to change how problems are categorized (because this might impact their title pursuits).
I have no care or concern for how editors want to categorize problems, and FIDE may group problems into any judgement bin they like -- just as a Thematic Tourney is free to do.
You are free to present fairy conditions in the same group with selfmates, if that is your preference. Just do not falsely declare it "Heterodox" (when it's clearly you can't even define the term).
I only care about a how problems are classified (so long as they are logically classified, where fairy conditions are not masquerading as stipulations).
Beyond that, you are free to do your worst in grouping problems for judging purposes, for publication purposes, and for the purpose of awarding titles.
Helpmates and selfmates were once grouped as fairies, but nobody should think they could ever be classified as a fairy condition.
Popularity is what determines grouping, and there's no logical formula to predict what will be popular.
Someday (soon, I expect) hs#n problems with orthodox rules will have their own section (due to their popularity, not due to their classification).
Some forward looking journals have already recognized the need to establish a separate section, and I encourage them to decide those matters for themselves (the same way I encourage FIDE to decide what sections they want to establish for judging problems).
Classification is another matter. We can not continue to classify problems based upon an illogical, inconsistent, backward-looking model.
It does not serve this community (which can define every conceivable interference pattern imaginable) to remain unable to define a fairy condition.
We are making it much more difficult than necessary to appreciate the beauty of this artform. If that trend continues, a day may come when every chess problem (and all our titles) will have lost all meaning. We owe it to ourselves to provide a more forward looking classification standard.
Now, I've taken us way off the topic. What year is this? They're all a blur, because no matter what we discuss, this always remains the core topic. | | (17) Posted by Olaf Jenkner [Friday, Oct 7, 2022 21:35] | I recommend a look at this masterpiece:
https://pdb.dieschwalbe.de/P1381235 | | (18) Posted by Kevin Begley [Friday, Oct 7, 2022 21:58] | @Olaf,
That is indeed a remarkable problem.
Bishops can do very strange things in help-aim problems (especially when they are required to lose a tempi).
I wouldn't say the two are directly comparable, but for an illustration of what remarkable lengths a Bishop will go to when it needs to lose a tempi, see: https://pdb.dieschwalbe.de/P0577055
In the hs#, there's the need to lose a tempo (not obstructing the path for the king), then the issue is the Bishop needs to be the other way around, which drives the remarkable circuit.
That's quite a nice motivation.
I wouldn't be surprised if we see even richer content possible from help-self problems (which is why this stipulation is becoming increasingly popular).
I wonder: has anyone yet composed a help-self-game? | | (19) Posted by Hauke Reddmann [Saturday, Oct 8, 2022 14:05] | @Neal: Teaching AlphaZero the composition of chess problems
would mean teaching AlphaZero what constitutes a good chess problem.
(In contrast to the trivial checking checkmate status of a game.)
I hope you'll enjoy judging 1000000 crappy problems :P | | (20) Posted by Marcos Roland [Saturday, Oct 8, 2022 17:08] | Thank you, thank you, thank you! | | Read more... | Page: [Previous] [Next] 1 2 3
MatPlus.Net Forum General All issues of StrateGems available online |
|
|
|