[Retros] reward for stalemate

Wed Jun 18 14:50:42 EDT 2014

Guus,

I have to admit, part of me does like where it sounds like your suggestion
might be going.
However, I'm probably an easy sale, because I have always preferred the
logic of "a posterori" retros; nevertheless, enjoying a particular logic is
one thing, it's quite another to advocate that this (or something like it)
should be the default standard -- I would never be so bold.

Frankly, I still wonder if special case rules (castling, en passant,
two-step pawn moves, promotion, 3R, and 50M) should be entirely removed
from the default standard, for the purpose of simplicity (regardless how
much I would regret seeing them relegated to a less orthodox existence).

In fact, I once wondered whether even pawns should be removed -- and keep
to directionless units.
I once took a giant database, and stripped away all problems involving
pawns, and castling (such that the board orientation had no impact).
Needless to say, I discovered that I had lost nearly every great problem
ever composed. Of the few remaining problems, greatness had sunk to a
profound rarity (but, in those very rare directionless problems, where
greatness could yet be found, it was always worthy of careful study).

Best,
 Kevin.

On Wed, Jun 18, 2014 at 9:34 AM, Guus Rol <grol33 at gmail.com> wrote:

> Hi kevin,
>
> If you set the bar too high, everybody leaves the contest.There is
> probably nothing in the world that spans the full base of problem chess
> except "the full base of problem chess". Andrey made some notes on its
> fantastic variety in an earlier post.
>
> I do have great sympathy for everyone who sticks to problems with
> "absolute" solutions. I started that way myself and it is still the core of
> the retro-analytical field. In fact I used to explain that retro-analysis
> is more orthodox than orthodox chess itself to any player giving me funny
> looks. Provability rules and ELO-ratings bear no weight!
>
> However, this is not to say that it is impossible to manage
> "uncertainties" in a satisfactory way. That is really an old school
> viewpoint. In the Logic domain, I know of at least 2 logics "Quantum logic"
> and "Fuzzy logic" which deliver solutions to handling "uncertainties".
> Justifiable solutions, not absolute solutions. What is missing in retro
> chess is "justification scenarios" as a basis for making "justifiable
> decisions". The acceptance of such scenarios is probally not universal but
> depends on what I call "emotional resonance". E.g. I would not expect
> anyone to join who hates matematical models and mathematical strictness.The
> justification scenario is always the "feeling connection" and makes someone
> stay or leave. As you and Andrey have both keenly observed "pretending to
> know about an uncertain system creates an illusion". The acceptance all
> depends on how real the illusion can be made. Just check out how
> the "Trekkies" are doing!
>
> Not to leave you with nothing tangible, here is a part of the RS
> justification scenario: To appreciate what is happening in an RS-solution
> you must move over from the player seat to one of a member in the audience
> who arrived during the game. You no longer "decide" but "observe". You see
> things happening and none of them will shock your expectations even when
> you haven't seen the whole game (note: seeing an e.p. move would shock you
> as an unexpected event unless it is clearly provable). At some point you
> have received sufficient information from your observations to see where
> the game is going and you can play it in your mind along with the players.
>
> What is being said in this scenario is essentially that RS-logic is also A
> Posteriori logic, i.e. based on observation. The perception of analytical
> uncertainty (which is a priori) has been replaced by the instruction to
> observe. It is only natural to you that nothing shocking happens on the
> board.and that the players know what game they are playing. Strange as it
> may sound, this scenario is very familiar in the form of reality shows on
> television. Everything that happens is as real as a proof game in chess
> but you know it has all been carefully edited to please you and not shock
> you out of your wits. If this story resonates somewhere inside you, chances
> are you might be an RS fan some day.
>
> Best wishes, Guus Rol
>
>
> On Wed, Jun 18, 2014 at 4:19 PM, Kevin Begley <kevinjbegley at gmail.com>
> wrote:
>
>> Thanks for that clarification, Andrey.
>> On this topic, I'm actually rather proud to have been anticipated by you,
>> while simultaneously a tad ashamed that I did not already better know your
>> position on the matter. :)
>>
>> Furthermore, I completely concur with your sentiments of consistency --
>> by whatever logic we decide to base our conventions, it is important that
>> they span the full basis of problem chess.
>>
>> To this end, I'm eager to learn what suggestions Guus might put in play.
>>
>> Best,
>>   Kevin.
>>
>>
>> On Tue, Jun 17, 2014 at 11:03 PM, afretro <afretro at yandex.ua> wrote:
>>
>>>  Dear Kevin,
>>>
>>> I agree with you.
>>>
>>> In fact our whole world may be nothing but an illusion.
>>>
>>> As to the similarity between a helpmate and a shortest proof game I
>>> mentioned it back in 1991, in the book on SPGs written by me and Gerd Wilts.
>>>
>>> Recently, Thomas Brand remarked in *feenschach* that it was
>>> inconsistent to refer to A>B problems as a non-retro variety while
>>> assigning SPGs to the retro category.
>>>
>>> As to the possibility to interpret backward play as forward play I also
>>> wrote about it, in an article on retrostalemate published in
>>> *feenschach*:
>>>
>>> “*By the way, while pondering on retro ambiguities and approaches to
>>> explaining them, the author was suddenly struck by the thought that
>>> retroanalysis could be presented as … a fairy chess type.*
>>>
>>> *Retroanalysis: a fairy chess form in which the ultimate goal of
>>> cooperative play is to reach the initial game array. In orthodox
>>> retroanalysis, the number of moves needed to attain the IGA is not
>>> indicated. Retros in which this number is indicated are known as SPGs.*
>>>
>>> *Direct play in retroanalysis, as a fairy form, is performed according
>>> to the following rules.*
>>>
>>> *White and black make moves in turn. In retroanalysis, a move (“forward
>>> move”) is “a mirror reflection” of a move by the same piece in standard OTB
>>> chess. When making a move, a piece can “revive” a piece of the other color
>>> on the square it is leaving, or turn into a pawn on the 2nd(7th) rank in
>>> appropriate cases, etc.”*
>>>
>>> Then I learned that similar thoughts had also been expressed by others.
>>>
>>> A retro problem with “backward” play can be seen as “forward play,” yes.
>>>
>>> But it is hard for me to see the backward plus forward span of the
>>> solution of e.g. a threemover with retro implications as a single “forward”
>>> span. I have to admit though that this can be done, too.
>>>
>>> The diagram position is a sort of “great divide” anyway; for a problem
>>> in which the stipulation is formulated in a “retro” way, like “Last N
>>> single moves?” or “Position after White’s Xth move; how did the game go?”,
>>> no conventions are needed to dismiss any possible line of retro play “for
>>> formal reasons”; moreover, in a retro problem with no additional conditions
>>> affecting retro play (like “Sb1 did not move”) a line of play other than
>>> the one intended by the author would always mean a cook; we must not
>>> dismiss such a cook on the basis of some formal convention. But as soon as
>>> a “forward stipulation” is added to a retro problem we have to immediately
>>> dismiss a lot of possibilities like triple repetition or 49 full moves
>>> occurring without pawn move and capture immediately before the position on
>>> the diagram has been reached. Indeed, there may be different “shortest
>>> mates” in a position depending on how many “pawnless plus captureless”
>>> moves have just been played consecutively – 46, 47, or 48. But we dismiss
>>> all these possibilities, while in a “release-the-position” retro we do not
>>> ever dismiss anything. This “dismissal convention” can be seen as a certain
>>> weakness, albeit perhaps an inevitable one.
>>>
>>> Yours,
>>>
>>> Andrey
>>>
>>> 18.06.2014, 09:20, "Kevin Begley" <kevinjbegley at gmail.com>:
>>>
>>> I don't fully agree, Andrey -- particularly as it pertains to
>>> proofgames.
>>> Attempt to differentiate between forward and backward play are always
>>> based upon an illusion.
>>> In fact, in terms of entropy, all chess problems can be shown to flow in
>>> the same direction.
>>>
>>> First, the proofgame is nothing more than helpgame. In fact, the
>>> proofgame reduces to a help A->B problem, in which A must equate to the
>>> game's arbitrary starting position.
>>> Now, compare this solving process to that of the standard helpmate, and
>>> you'll find that the only difference is that the helpmate solver must guess
>>> from among many possible end positions -- once the correct final position
>>> is guessed, the very same technique may be required to reach that position.
>>>
>>> Therefore, if there exists any "backward play" (or retroanalysis) in a
>>> proofgame, it must also exist, to some comparable degree, in every standard
>>> helpmate.
>>> Entropy flows in the same direction (always forward, toward the
>>> achievement of a final position).
>>> The primary difference is that the proofgame solver is given the final
>>> position, and this information constitutes an inherent cook-avoidance
>>> mechanism (the achievement of the one stipulated diagram is necessary, and
>>> by avoiding alternative mating scenarios, the composer is able to devise a
>>> more complex task).
>>> There is no element of "backward play" which inherently distinguishes a
>>> help-diagram (or help-game) from a help-mate.
>>>
>>> Furthermore, the entire notion of "backwards play," in any given retro,
>>> can be treated as the "forward play" of a fairy condition, running the
>>> opposite direction. For any problem, entropy runs in the same direction
>>> (from the chaotic circumstance of a diagram, toward the order of satisfying
>>> the stipulation).
>>> Therefore, this whole notion of "purity" in the retro domain is, well,
>>> purely an illusion.
>>>
>>> Best,
>>>  Kevin.
>>>
>>>
>>>
>>>
>>>
>>> On Tue, Jun 17, 2014 at 8:34 PM, afretro <afretro at yandex.ua> wrote:
>>>
>>>  Dear all,
>>>
>>> If an OTB player comes across this debate and reads it and tries to
>>> fathom on its basis what retroanalysis actually is, he/she may think that
>>> retroanalysis is all about e.p. captures and castling rights as well as 3R,
>>> 50 move remis rule and DR. These things, however, pertain to a smaller part
>>> of retro problems. They are only related to retros with a stipulation
>>> focusing on forward play. And mind you, *forward play is not
>>> retroanalysis*; retroanalysis is actually *backward play!*
>>>
>>> Are there any uncertainties involving e.p. capture, castling rights,
>>> etc, when one is to solve a retro with a stipulation like “Last N single
>>> moves?” or “Where was the black light-squared bishop captured?” or a
>>> shortest proof game in X moves? No; all the information that a solver
>>> needs to solve such a problem is “embedded” in the diagram. *No
>>> conventions are needed for “pure” retros that are devoid of any
>>> forward-play implications.*
>>>
>>> When one deals with a pure retro, *no legal proof game that may have
>>> led to the diagram position is discarded *on the basis of any
>>> conventions. The retro stipulation – “Release the position” or “Last N
>>> single moves?” – simply indicates – either in “a general form” or strictly
>>> – what fact(s) about the move(s) that may have been played to produce the
>>> diagram position the solver is to reveal. In doing so, the solver does not
>>> have to dismiss on the basis of a convention any of the possible legal
>>> games. For example, when the stipulation focuses on the last X single
>>> moves, these moves are usually unique (unless there are two or more
>>> solutions); the moves that may have been played before that are not
>>> “dismissed” but “neglected” on account of being unessential. But when a
>>> forward stipulation is added, like “Mate in 3 moves,” and when it involves
>>> e.p. capture or castling legality, then a convention “brutally” dictates
>>> the solver to disregard a multitude of games that may have led to the
>>> diagram position and select some line of play in line with the requirements
>>> of this convention.
>>>
>>> Please note that this in fact amounts to “retroanalysis under pressure”
>>> versus “pure retroanalysis.”
>>>
>>> Yours,
>>>
>>> Andrey
>>>
>>> 18.06.2014, 00:37, "Joost de Heer" <joost at sanguis.xs4all.nl>:
>>>
>>> On 06/17/2014 10:28 PM, Kevin Begley wrote:
>>>
>>>  Whether or not you agree with GM Short's solution, is of no consolation,
>>>  here.
>>>  The point is:  FIDE (a game federation) owns the FIDE rule book, and
>>>  they may do what they like with it, for the purposes of serving their
>>>  own unique charter.
>>>
>>>  They have no cause to be concerned how it affects a chess problem
>>>  community.
>>>  Chess problems predate their game, and should never have been impacted
>>>  by it.
>>>
>>> Indeed. I keep finding it bizarre that several fairy retractors use the
>>> game rule '3 fold repetition' for a genre that has absolutely nothing to
>>> do with game play.
>>>
>>> Joost
>>>
>>> _______________________________________________
>>> Retros mailing list
>>> Retros at janko.at
>>> http://www.pairlist.net/mailman/listinfo/retros
>>>
>>>
>>> _______________________________________________
>>> Retros mailing list
>>> Retros at janko.at
>>> http://www.pairlist.net/mailman/listinfo/retros
>>>
>>> ,
>>>
>>> _______________________________________________
>>> Retros mailing list
>>> Retros at janko.at
>>> http://www.pairlist.net/mailman/listinfo/retros
>>>
>>>
>>> _______________________________________________
>>> Retros mailing list
>>> Retros at janko.at
>>> http://www.pairlist.net/mailman/listinfo/retros
>>>
>>>
>>
>> _______________________________________________
>> Retros mailing list
>> Retros at janko.at
>> http://www.pairlist.net/mailman/listinfo/retros
>>
>>
>
> _______________________________________________
> Retros mailing list
> Retros at janko.at
> http://www.pairlist.net/mailman/listinfo/retros
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://one.pairlist.net/pipermail/retros/attachments/20140618/d952d0cb/attachment-0001.html>