|(1) Posted by Hauke Reddmann [Monday, May 3, 2021 13:24]|
Selfmate where Black has only 1P...and promotes
If I recall correctly, KQR/KQ already is sufficient
for selfmate. I did not check for KR and other material.
Is it possible to have a position where Black to move
has only a pawn beside the King, promotes it, and must
give selfmate (somewhen!) regardless of the piece?
It mustn't be of equal length or be correct at all,
it would suffice that it is a "selfmate tablebase win".
(Do we have such tablebases, extending private efforts
of e.g. Torsten Linss?)
This should be doable, e.g. with a matrix Ke1 whatnot/Ke3 Pa2,
1...a1Q,R#, 1...a1B n.Bd1+ m.Rc3+ Bxc3, 1...a1S n.Rd1+ m.Sc2+ Sxc2.
Hey, have I just invented the selfmate study genre? :-)
|(2) Posted by Andrew Buchanan [Monday, May 3, 2021 14:23]|
There's nothing in PDB that matches your question, however I did find this in WinChloe:
(= 10+2 )
Sphicas, version Morse (or the other way round)
The Problemist 1985(v)
115 p.301, The Problemist (May 1987)
I guess the paradox is that the weaker the Black force, the easier it is to force them to selfmate. So only promotion to Q will give Black a chance to reach the full length of 59 moves. The same principle applies in over-the-board Schlagschach.
|(3) Posted by Frank Richter [Monday, May 3, 2021 14:45]|
somewhat like the 1st Prize here?
You may ask Olaf Jenkner for corresponding researches (and surely he will cook the given example above ...)
|(4) Posted by Rewan Demontay (Real Name: James Malcom) [Monday, May 3, 2021 16:07]|
Andrew, you're not looking hard enough in PDB. ;-) I know of this problem from prior searches: https://pdb.dieschwalbe.de/P1101367
Yaakov Mintz, 1635N U.S. Problem Bulletin 01-02/1989, dedicated to Ron Brain
(= 13+2 )
|(5) Posted by Torsten Linß [Monday, May 3, 2021 16:32]|
g='s#' and bpieces=2 and piecelist='kb' and sol='1='
Lists a few positions including some from tablebases.
|(6) Posted by Andrew Buchanan [Monday, May 3, 2021 17:39]|
Ha my bug was looking for piecelist='sB' rather than piecelist='b'. Randomly different syntax required in different parts of the UI!
|(7) Posted by Olaf Jenkner [Monday, May 3, 2021 18:11]|
Some years ago I tried to cook both examples. I ran Gustav days without any cook.
I'm quite sure that Mintz's problems are sound. Everybody is invited to show I'm wrong.
|(8) Posted by Frank Richter [Tuesday, May 4, 2021 15:18]|
So Hauke, what's your impression on given examples?
|(9) Posted by Joost de Heer [Tuesday, May 4, 2021 18:43]|
(= 10+2 )
Petko Petkov, Stella Polaris 1969
|(10) Posted by Andrew Buchanan [Tuesday, May 4, 2021 19:38]|
That s#4 is sublime
|(11) Posted by Olaf Jenkner [Tuesday, May 4, 2021 21:20]|
Olaf Jenkner, Phenix 306/307 2020
(= 6+2 )
4 model mates
For solution with autoplay see
|(12) Posted by Andrew Buchanan [Wednesday, May 5, 2021 09:14]|
Olaf that is extraordinary. How wonderful! :D How could one ever imagine and go about finding such a thing?
|(13) Posted by Hauke Reddmann [Wednesday, May 5, 2021 09:51]|
@Frank: I used the search string given in this thread to
dig around a bit in PDB and was flabbergasted how little
material is necessary, even for AUW :-)
But then, I'm no s# expert anyway and can't judge it.
|(14) Posted by Olaf Jenkner [Wednesday, May 5, 2021 18:07]|
I found the position using my laptop.
I put the kings and the black pawn in their places. All legal positions (black to play) with two queens (but at least one at 8th rank),
two rooks and knight were searched with Gustav. So there is a promotion key and a position without promoted pieces.
No more posts
MatPlus.Net Forum General Selfmate where Black has only 1P...and promotes