[Retros] A question on stipulations

[Renny Bosch]

Thank you Bernd, Mario, and Andy, for your thoughtful replies to my 
question.  But now let me make it more specific.  I have been working on a 
problem in Joost's Probleemblad post of 27 October, that goes like this:

R356 - Harry Goldsteen
8/8/7p/PP1P4/kr1rP3/BbKpRP2/S2PppP1/qSqRbB2 (14+11)
+---+---+---+---+---+---+---+---+

|   | . |   | . |   | . |   | . |

+---+---+---+---+---+---+---+---+

| . |   | . |   | . |   | . |   |

+---+---+---+---+---+---+---+---+

|   | . |   | . |   | . |   |*P |

+---+---+---+---+---+---+---+---+

| P | P | . | P | . |   | . |   |

+---+---+---+---+---+---+---+---+

|*K |*R |   |*R | P | . |   | . |

+---+---+---+---+---+---+---+---+

| B |*B | K |*P | R | P | . |   |

+---+---+---+---+---+---+---+---+

| S | . |   | P |*P |*P | P | . |

+---+---+---+---+---+---+---+---+

|*Q | S |*Q | R |*B | B | . |   |

+---+---+---+---+---+---+---+---+
Mate?

After much analysis, I have "proved" that the answer must be No, because I 
can't resolve the position (I could if the black pawn at h6 were at h5, but 
it isn't).  But here's my problem:  I don't feel secure that I have really 
found the solution.  I probably overlooked something and if I kept on 
working I might have a wonderful surprise -- and then I'd KNOW that I had 
the solution.  This way it's very frustrating (like sex without orgasm). 
Wouldn't you agree that in this case it is very unlikely that the answer is 
"No"?

Renny


----- Original Message ----- 
From: "andrew buchanan" <andrew at anselan.com>
To: "The Retrograde Analysis Mailing List" <retros at janko.at>
Sent: Saturday, November 21, 2009 5:57 AM
Subject: Re: [Retros] A question on stipulations



>A few more thoughts on this...

>

> (1) Most commonly when the stipulation is "Mate?" we are talking about a 

> conventional RA problem where the interest is in figuring out the history 

> of the game. It is an additional stylistic flourish if the game ends 

> dramatically with a mate, and the composer calls attention to this in the 

> stipulation. (It seems to me that a mere check in the final position is 

> regarded as a slight defect, but a mate paradoxically is not, because of 

> its finality.)

>

> (2) Unless there is an hint of "funny business" (e.g. e.p.) it does make 

> sense for the solver to explore first the more promising avenue that the 

> position is legal, and therefore hunt for a proof game or perhaps just a 

> way to untangle the position. I am not sure I like the word "assumption", 

> because it's a bit ambiguous - if I *assume* the position is legal, why 

> would I then have to *prove* it with a proof game? If I want to prove that 

> the position is legal, I just resort to my assumption!

>

> (3) Renny wrote:

>> In the first place a non-existence proof

>> is much more difficult than an existence demonstration (typically

>> requiring computer assistance to be certain), and in the second place

>> such a proof is much less satisfying than a game with some ingenious

>> twists and turns.

>

> A non-existence proof could be very short. E.g. 2 White kings. Or a parity 

> argument. Or an unpromoted piece having escaped from a pawn cage etc etc. 

> And such an argument could be quite satisfying. However the thing is that 

> (except for the specialist topic of Illegal clusters) retro enthusiasts 

> don't really like *totally* illegal diagrams. We would like there to be 

> some escape clause, some unlikely scenario which we are forced to conclude 

> is actually what happened.

>

> The famous quote from Sherlock Holmes applies: "How often have I said to 

> you that when you have eliminated the

> impossible, whatever remains, however improbable, must be the truth?"

>

> Regards,

> Andy.

>

>

> ----- Original Message ----

> From: Mario Richter <mri_two at t-online.de>

> To: The Retrograde Analysis Mailing List <retros at janko.at>

> Sent: Sat, November 21, 2009 8:35:03 AM

> Subject: Re: [Retros] A question on stipulations

>

> Hello,

>

> Bernd wrote:

>> the answer might be "No", because the side which seems to be mated could

>> have a right to capture en passant.

>

> Other possibilities are: 50-moves-rule, parity, retro-opposition, ...

>

> Reduced to their logical kernel, those problems simply ask

> the question: "Is this position legal with a specified side to move?"

> so they are in good company with similiar types like "Mate in one?",

> "Who can mate in n?" a.s.o.

>

> Renny> if ... the proof is an elaborate explanation of why every possible

> Renny> attempt at resolving the position leads to a dead end,

> Renny> would they publish such a problem?

>

> Why not?

> (Just a thought experiment: Replace "Is Black mated?" by "Can White 

> castle?".

> Would you still argue, that if an elaborate proof shows, that all attempts

> to resolve the position with preservation of White's castling right leads

> to a dead end, then the problem is not worth publishing?)

>

> Best wishes,

>

>  mario

>

> _______________________________________________

> 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