Website founded by
Milan Velimirović
in 2006
14:05 UTC
| |
MatPlus.Net  Forum General Last move in a self-mate problem. |
| |
|
|
|
|
You can only view this page!
| | Page: [Previous] [Next] 1 2 | | (21) Posted by Kevin Begley [Tuesday, May 8, 2012 17:39]; edited by Kevin Begley [12-05-08] |
Black mating moves in s#n are generally superfluous -- but, not always.
But, regardless, many s#n composers specifically construct to achieve a single mating move.
Others may not.
If you ease the standard, you may find many previous constructions will appear less than economical (if not worse)!
Thus, even if a composer's black moves appear superfluous in a s#n, singular-mating moves may have been an important constructional constraint, that was imposed for a reason, and should not be dismissed.
Therefore, I'm not advocating that we ease the standard -- I'm advocating that we allow the composer to CHOSE their own standard (formally).
And, judges should be required to appreciate problems according to the constructional constraints that composers have imposed upon themselves.
We require a formal stipulation mechanism which provides for these alternative standards.
I've suggested one possibility for this -- one which would resolve a number of issues (including dual-suppression in CapZug), and would cover a wider variety of stipulations (such as Michel's h4#2); but, I'd welcome other suggestions (a more simplified, more familiar stip-format would be nice).
I realize this suggestion will lead to some confusion as to what constitutes the "Orthodox" form of s#n.
I presently have no opinion as to how we might want to define that...
Perhaps Orthodox will require that the primary and ultimate aims be the same...
Perhaps it will require that the ultimate aim is iterated down to finality (#, =, x, etc).
However, it is worth noting that nobody (presently) is able to provide a valid definition for the Orthodox/Fairy division.
Perhaps this will help to achieve a clearly defined division; or, perhaps we will conclude that such a division is artificial.
Frankly, I think this is a battle that we should have no interest to engage in (it really comes down to a jockeying for awards/titles).
There are alternative means by which we can divide s#n problems which use Circe/Madrasi/Grasshoppers from problems from those which follow entirely the FIDE rules of Chess.
But, this all depends upon our formal stipulation being allowed to override Dead Reckoning implications of the Codex.
So long as Dead Reckoning constitutes a Codex provision, then I don't see how anyone can argue that black is even permitted to play the final mating moves in s#n.
It doesn't matter how new a Codex rule may be -- an old law is no stronger than a law which goes into effect today.
Do problemists now require a separate rulebook (from chess players)?
| | | (22) Posted by Dupont Nicolas [Tuesday, May 8, 2012 18:25]; edited by Dupont Nicolas [12-05-08] |
Ok, ending a s# with a black move may be of interest. It remains that it is not consistent with other fields (plus the fact that such a move is illegal regarding DR): stalemating the black side is exactly the same as forcing this side to play an auto-check. Thus the stipulations =n and s!+n are identical (where !+ stands for auto-check). But an =n problem ends with a white move, we never verify that each possible black answer is an auto-check (although it might be as interesting as verifying that each possible black answer is a checkmate in the s# context). On the contrary, the same problem but with the s!+n stipulation should end with the list of black auto-checks!
| | | (23) Posted by Kevin Begley [Tuesday, May 8, 2012 20:01]; edited by Kevin Begley [12-05-08] |
Formal stipulation has to take precedent over Dead Reckoning.
I think DR can only apply to studies.
If I have K+B vs K, in a h=n, it is unreasonable that DR should declare the position dead, and nullify all moves.
DR presumes a particular objective (win/draw/lose), which is contrary to the objectives of any formal stipulation.
This would likely impact some DR problems (potentially any employing a formal stipulation).
It would seem highly ironic if we must resolve this by declaring DR (an orthodox rule) to be a fairy condition.
Maybe K+B vs K is h=0 (as good as stalemate itself!), in the DR Condition. :)
| | | (24) Posted by Diyan Kostadinov [Wednesday, May 9, 2012 07:08]; edited by Diyan Kostadinov [12-05-09] |
In my opinion selfmates should be finish with black move exactly as they are now. Without it the problem looks somehow unfinished. All we know that there are so many great S# with nice black last moves which add to the problem more beauty. The black last moves can be also a central part of the authors idea as was presented in the previous post by Frank Richter.
Actually I think that sometimes the authors do not pay enough attention to the last black moves in selfmates and this is mistake, because repetitions, duals etc. on the mating move can really destroy the good impression of the problem...
| | | (25) Posted by Kevin Begley [Wednesday, May 9, 2012 08:05]; edited by Kevin Begley [12-05-09] |
There are a number of examples, in the Win Chloe db, where white promotions are considered duals, and those duals result in problems marked as "cooked."
Here is just one example (which I have shown on MPF before):
Uri Avner
Prize, Israel Ring Tourney, 1981
 (= 14+11 )
s#3 (14+11) C- (according to Win Chloe db)
1…Rxb3 2.Rd6+ Kf5 3.Se3+ Rxe3#
1…Rc2 2.Qc3+ bxc3 3.f8=Q+/R+ (promotion dual)
1…Sxb3 2.Se4+ Ke6 3.Sc5+ Sxc5#
1…Sc2 2.Ra6+ Ke7 3.Qxb4+ Rxb4,Sxb4,Bxb4# (unmarked: multiple mating moves in the set play)
1.Qc2! (> zz)
1…Rb3,Bd~ 2.Qc3+ bxc3 3.f8=Q+/R+ (promotion dual)
1…Rxc2 2.Se4+ Ke6 3.Sc5+ Rxc5#
1…Sb3 2.Ra6+ Ke7 3.Qc5+ Sxc5#
1…Sxc2 2.Rd6+ Kf5 3.Se3+ Sxe3#
Are these promotions really duals, which render this problem unsound?
I don't believe this problem is cooked -- I suspect very few would approve of the C- label here.
There are other examples, in the WC db, where multiple mating moves, which occur in the play, are not considered cooks (though, some people argue they should be) -- I'll try to locate some...
The fact is, there is no universal consensus on the issue of multiple mating moves in s#n problems (even FIDE Judges have differing opinions as to whether this constitutes a dual).
So, personal preferences as to what this convention should be (or should remain), amount to little more than subjective dogma.
Despite what Diyan says...
We don't have the luxury to argue whether our present convention should continue (or be altered).
We don't have any codified convention which covers this issue.
What we do have, on the other hand, is problems which exhibit both of these possible interpretations.
So, regardless which interpretation our delegates may decide to codify as "orthodox" (if ever they can muster a definition for such a term), we will still be required to provide some formal mechanism to express the alternative interpretation.
In my view, the alternative interpreation can be best expressed by permitting the composer with some formal means to indicate what they intend as their ultimate aim.
a) s<#>0.5 = evaluate recursively down to the ultimate aim <mate> (and consider multiple mates to be a dual), or
b) <s#0.5> = evaluate only to the achievement of a <s#0.5> position (and, since multiple mates do not occur in the solution, they can not be considered a dual).
There is little value arguing here which of these should be considered "orthodox."
There is, as yet, no valid definition for the term "Orthodox" (nor "Fairy") -- even our Codex provides us no definition, despite using these terms.
It's like the definition of "heat" -- everybody thinks they know what it means, because everybody feels its effect; but, try to define it for yourself (you'll probably find it requires some understanding of the statistical movement of small particles).
Unlike heat, however, you'll find no viable scientific definition for the terms "Orthodox" or "Fairy."
These terms are really only useful for the purpose of creating a synthetic favored status (generally for the purpose of awards/titles).
The question here is not: which one of these (or both) deserve the blessing of "Orthodox" recognition.
The first question we must answer is: how do we formally provide for both interpretations?
After we produce a viable formal mechanism to describe both interpretations, we might gain the understanding necessary to adequately define Orthodox/Fairy (in a logical way, based upon fundamental stipulation components -- rather than in an element by element, thumbs-up/thumbs-down, fashion).
| | | (26) Posted by Kevin Begley [Wednesday, May 9, 2012 08:33]; edited by Kevin Begley [12-05-09] |
Richard Schurig
Schachzeitung, 1851
 (= 7+5 )
s#3 (7+5) C+ (according to Win Chloe db)
1.Bd5+! Ke5 2.Bg8+ Ke4 3.Rh3 ...Qg7#,Qxg8# (not listed as a dual)
There is no universal convention here.
And, even if you want to bless one interpretation as Orthodox, you have to provide some formal mechanism which describes the alternative.
The simplest way to do this -- which would provide coverage for a number of s# derivatives (hs#n), which would provide a formal dual-suppression mechanism for like-minded inventions (CapZug), and which would provide formal coverage for a number of fringe stipulations (such as Michel's h4#2) -- would be to alter the taxonomy of our formal stipulation, in such a manner that the ultimate objective may be explicitly stated as something other than the primary goal.
What alternative is there?
Are we going to create a fairy condition which alters the nature of dual interpretation?
I remind you that such a fairy condition would not alter (in any way) the rules of play -- therefore, this would not constitute a valid fairy condition.
We require an alteration in the task laid out by the composer -- therefore, it should be expressed within the stipulation.
My suggestion is to instantiate both the <ultimate objective>, and the (primary goal), in the stipulation (at the very least, do this when it is necessary).
Any s#n problem can be written as: O(s#0.5)n -- where "O" = oppose/direct play, "(s#0.5)" represents the primary goal, and "n" provides the deadline (a limit on the number of white moves).
The difference in the two interpretations of s#n, depend upon which ultimate objective is intended by the composer.
In the above problem, the ultimate objective is the same as the primary goal -- just reach a <s#0.5> position.
This can be written as: O(<s#0.5>)3 -or- O<s#0.5>3 -or even just- <s#0.5>3 (where O=direct play is implied).
This says, upon reaching a position with <s#0.5>, the solver should successfully cease, w/o listing any further moves.
But, in other s#n problems (particularly when black's mating moves are important), the primary goal is intended to be recursively resolved, down to the terminal aim (in this case, mate: <#>).
This can be written as: O(s<#>0.5)3 -or- (s<#>0.5)3 (where O=direct play is implied).
Obviously, in one of these alternatives (I suppose whichever one we consider the more orthodox), the ultimate objective can be implied.
Thus, the "orthodox" implication would either require that we solve recursively unto a terminal aim (#, =, x, 00, ep, etc), or it would require that we solve no further than the primary goal.
It doesn't matter much to me... what matters to me is only that the codex provides a logical, formal, universal mechanism to express an intended task.
And, if somebody has a better idea how to formally stipulate these things (perhaps something more economical / more familiar), I would certainly welcome any suggestions.
For the record (and elaboration), it's worth noting...
r#n is really a s#n + fairy condition ("W/B are obliged to #1, if possible").
However, semi-r#n, may not require any fairy condition.
Explanation:
Rather than attempting to reach a (s#0.5) position (as you would with a s#n/r#n), the primary goal of a semi-r#n may be (alternatively) written as: (h<#>0.5 for black) -or just- (h<#>0.5-b) -or just- (<#>1-b)!
Upon forcing the primary objective (a position with <#>1-for black), the solver simply recursively applies this objective as a stipulation to the position achieved.
In this case, it is not necessary to consider any of black's alternative moves (the solver simply iterates, from the achieved position, to solve how black checkmates in 1).
This shows that semi-r#n can be recursively instantiated, without using any fairy condition.
And, hopefully, it demonstrates how one can map the progression of fundamental stipulations, using recursive layering techniques.
Another example: suppose you have a babson #4, which has duals on white's 3rd move.
Do duals render this incorrect?
We can ignore this question, by instead insisting upon a correctly stipulated problem (with no duals).
Whereas a normal #4 can be written as "O(<#>)4", one with duals on white's 3rd move can not.
Instead, this problem should be stipulated as: "O(<#2.5>)2".
Here, the ultimate objective is no longer mate, instead, white has two moves to achieve a position where white can mate in 2, following any black move (or #2.5).
Now that the problem is correctly stipulated -- and there are no duals! -- the judge may shift attention to weighing the redemptive merits of the required unorthodox stipulation.
| | | (27) Posted by Sven Hendrik Lossin [Thursday, May 10, 2012 23:36] |
I don't understand what this is all about. In problem solving contest you never write the black mating moves because - as you said Nicolas - the stipulation is already achieved.
Nevertheless the mating move may play a role when judging about the artistic impression of a selfmate.
The other question: In selfmates there are no black duals for black not being responsible for fulfilling the stipulation. Moreover I would be annoyed if a problem of mine would be judged lower because black has more than one mating move.
| | | (28) Posted by seetharaman kalyan [Friday, May 11, 2012 07:44] |
I agree. Question of duals should arise only for the side responsible for fulfilling the stipulation. That is why in helpmates alone dual-free moves are required for both black and white. In direcmates such an issue should not arise for black moves.
| | | (29) Posted by Kevin Begley [Friday, May 11, 2012 13:41]; edited by Kevin Begley [12-05-11] |
I appreciate people expressing their views, but these discussions need to get beyond subjective preferences.
Without naming names, suffice it to say, some well known selfmate experts disagree on this particular issue.
I'm talking about FIDE Judges and titled problemists -- opinions in this forum are not likely to change their views.
Is it necessary to have a consensus?
Not absolutely -- I'm often reminded that despite plenty of differing opinions, the sky has not yet fallen.
Of course, this reminder happens to be a non-deductive argument (one that Hume would say commits a logical fallacy -- the equivalent of concluding that the sun will rise tomorrow, same as it always has).
But, even if you believe in scientific induction (and presume our "problem sky" is uniform), you would require measured data to quantify the likelihood of Chicken Little's hypothesis.
What is the measurable impact of having no consensus?
In this case, it undermines the universality of some of our most fundamental properties (soundness & duals).
If these properties need not be universal, then our Codex had no business attempting to define them!
Furthermore, these properties are fundamental to any legitimate judgement (awards/titles depend upon their consideration).
To guarantee that your problem is judged fairly (according to the standards that you set out for yourself), you would require a database which connects your standard to like-minded judges (presuming you have the luxury of knowing who the judge is, and/or the option of various publications).
That is a long way from the fair sky that we should expect -- it falls quite heavy on the backs of new composers!
When you actually look at measurable data, Chicken Little's hypothesis begins to look like a well supported theory!
On important matters (especially soundness, anticipation, rules, genre divisions, etc), a lack of consensus is measurably eroding the foundations of this art form.
We are no longer tethered by any elementary axioms, because we have neglected to promote the intended meaning of our most important terms: aim, goal, stipulation, fairy condition, fairy chess, orthodox, dual, soundness, etc.
We carelessly surrendered these axioms, without any appreciation of their intended benefit.
I only became interested in chess composition a little over 10 years ago.
In that period, I've seen more than my share of these discussions.
Anticipation -- I've seen arguments as to whether all positions in the EGTB constitute a study publication.
Soundneess/duals -- I've seen disagreements concerning the severity of promotion duals (and they've never even addressed the issue of fairy promotion duals).
Axiomatic terms -- (Aims/Stipulations/Fairy Conditions/Fairy Chess/Orthodox/etc) -- I've seen carelessness undermine the meaning of all these terms, to the point that the Codex usage of these terms is beyond any hope of clarification.
Rules -- I've seen a vast number of disagreements concerning fairy chess rules; and recently, it became clear that there is not consensus on the Orthodox Rules (particularly concerning the impact of Dead Reckoning).
And, I could go on, and on, and on...
A consensus, on highly important matters (such as multiple mates in s#n problems) would have substantial benefits.
In any genre, consensus provides the necessary stability to welcome new adherents, fairly.
If this is achieved logically and universally, it makes ALL of problem chess more accessible, for everyone.
My point is, we have to get beyond simply stating our own subjective interpretations.
We need a plan to achieve some universal consensus, on critically important issues.
We need a plan to restore the meaning of our terms.
It's not enough to win each argument -- regardless which sides prevail in these matters, we will need to provide for alternative standards (and alternative interpretations).
The plan I would suggest begins with defining the fundamental elements of a chess problem.
I believe we start by defining our fundamental aims (#, =, x, etc), and provide some combinational logic, which enables the construction of more complex aims.
We don't allow any rules to be part of an aim (##/== : may be an aim, but it can not be allowed to alter the rules of play -- that must be done explicitly, using a fairy condition!).
Using these aims, we define what an elementary stipulation consists of (type of play in relation to the aim, deadline to reach the aim, format of solution, what have you...).
One again, we don't allow any rules to be part of a stipulation (a stipulation such as r#n may not alter or constrain the rules of movement -- only an explicit fairy condition can do this).
Once we have axiomatically established what a stipulation is, we can allow the stipulation to be an aim itself (e.g., in a hs#n, the first layer aim is to achieve a s#0.5 stipulation, the second layer aim may be # itself).
Thus, our standard stipulation must include some information as to the ultimate objective -- a problem may be intended to resolve all the way down to an aim, or it may terminate prior.
This also provides a dual-suppression mechanism (where the composer can effectively dictate, formally, their own standard for duals).
Once you have a complete foundation for formal stipulation, you can begin to address more complex issues.
Like establishing a complete set of default rules, for chess (and for any families of variant conditions).
Then, you can begin to define what constitutes fairy/orthodox, and you can begin to logically determine the genre divisions.
But, you'll never reach a consensus without differing to a fair, authoritative arbiter (one committed to maintaining consistency).
Only the vote of delegates could achieve this... and, only the Codex could carry their decisions.
If you don't like my solution, that's fine... but, it's not enough to quibble over fine points.
We need a full restoration plan... or else, we can go back to ignoring the visibly gloomy sky.
| |
No more posts | Page: [Previous] [Next] 1 2
MatPlus.Net Forum General Last move in a self-mate problem. |
|
|
|
ISC 2024
