MatPlus.Net

 Website founded by
Milan Velimirović
in 2006

17:41 UTC
ISC 2024
 
  Forum*
 
 
 
 

Username:

Password:

Remember me

 
Forgot your
password?
Click here!
SIGN IN
to create your account if you don't already have one.
CHESS
SOLVING

Tournaments
Rating lists
1-Jan-2024

B P C F





 
 
MatPlus.Net Forum General Oops they did it again
 
You can only view this page!
(1) Posted by Sven Hendrik Lossin [Saturday, Aug 31, 2013 10:15]

Oops they did it again


I just cooked the Problemists S2453 (s#175).
I have written some notes to those ultralongmovers here: http://matplus.net/start.php?px=1346533580&app=forum&act=posts&fid=tt&tid=993&pid=8810#n8810

There are some things to add:
In some cases the strategy is to give black the a- and the b-pawn and in some other cases black is left with two of the doubled pawns on the d-file. In both cases they can easily forced to mate the white king.

I don't know exactly why the Problemist ignored my concerns about these works, at least they could send them to Olaf or me before they publish them. Although I am sure that this is not necessary in this special case when it is completely clear that black can never ever survive 175 moves.

I only can recommend to anybody willing to publish these ultralongmovers to play chess games against novice chess players with selfmate rules. Then you can get a feeling how many moves somebody can survive in this kind of position.

And maybe somebody like Torsten Linss can answer the question how many moves white needs at maximum to force black pawns d7,d6 to mate the white king with white having Q, R, N in a random position.
 
(Read Only)pid=10855
(2) Posted by Michael McDowell [Saturday, Aug 31, 2013 10:42]

It's a composer's responsibility to make sure his problem is sound before submitting it for publication, not the job of the editor to do his work for him. Maybe you should issue a general invitation to composers of long selfmates offering your services as a tester.
 
 
(Read Only)pid=10856
(3) Posted by Sven Hendrik Lossin [Saturday, Aug 31, 2013 11:55]

Hi,
I agree to some extent.
But with these works I mentioned there is simply no chance that it is sound.
And everybody having a slight idea about playing chess on one hand and selfmates on the other hand can see that at once.

Have a look at the position and think about how many moves it might take to win all the black pieces but two d-pawns. Rough estimate 30 moves maximum. This is where you need the idea about playing chess.

How many moves it might take to force the twopawns to checkmate the white king is then part of the selfmate reasoning. Maybe it is 20 or so but never more than 30.
 
   
(Read Only)pid=10857
(4) Posted by Olaf Jenkner [Sunday, Sep 1, 2013 09:23]

Here is the mentioned problem:

http://www.softdecc.com/pdb/search.jsp?expression=probid=%27P1271289%27
 
   
(Read Only)pid=10861
(5) Posted by Diyan Kostadinov [Sunday, Sep 1, 2013 11:30]

Sven, about your: "...at least they could send them to Olaf or me before they publish them" - as an editor of KoBulChess (and director of some other tourneys) usually I receiving also Selfmates in more than 20 moves (most of them miniatures) which is difficult (impossible) to me to testing them... How you do it? Can I send some of them before publication to you (or Olaf Jenker)?
 
   
(Read Only)pid=10862
(6) Posted by Olaf Jenkner [Sunday, Sep 1, 2013 22:13]

Yes, but my resources are limited.
 
   
(Read Only)pid=10864
(7) Posted by Steven Dowd [Monday, Sep 2, 2013 07:53]

Diyan, I would be happy to help you test as well.
 
 
(Read Only)pid=10868
(8) Posted by Sven Hendrik Lossin [Monday, Sep 2, 2013 09:44]; edited by Sven Hendrik Lossin [13-09-02]

Dear Diyan,
my hardware is somewhat old but I would also support you. If it is about miniatures in selfmate then also Frank Müller and Torsten Linß may be worth an e-Mail as they know so much more about selfmate miniatures than me. Olaf and me are not magicians, the answer often is: "This one is probably sound." This was for example the answer when we checked the correction of the wonderful P1233664 ( http://www.softdecc.com/pdb/search.jsp?expression=probid=%27P1233664%27 ) in the PDB which regretfully received no distinction because the old version was cooked.

But I'd like to make a difference between two things: If you receive a miniature selfmate with ~20 moves then I can understand that an author and an editor have difficulties to check the soundness. But the ultralongmover of Marks I mentioned and those of Ugren about a year earlier don't need that check because it is simply impossible that they work (and these aren't miniatures). The strategy often is: Leave black with minimum mating material and force selfmate then. It seems as if some of our problem chess friends have completely no idea what is possible and what is not.
The problem is that with orthodox moremovers nobody would get the idea to show for instance a King, Bishop, Knight vs. King mate in 64 moves because nowadays everybody can have a look into databases and see that there is a shorter way. This is what needs to be done for selfmate as well. We need an answer to questions like: black has doubled pawns on d7 and d6, white has a queen, a rook and a knight: how much moves will it take at maximum to force selfmate. My estimate is that white is always able to force selfmate with very little exceptions and that it should be no more than 30 moves. The same question arises for an a-, b-, g- or h-pawn vs. queen, rook, bishop (a bishop is always helpful in that case).

"How you do it?"
I usually play forced moves that win black material and then one of the best cooking software nowadays, Olaf Jenkners "Gustav", does the rest for me. Although this "rest" took me some hours to find the right parameters and another four days of calculation for my old notebook in the case of Udo Marks selfmate. The parameters were roughly "one black move with four exceptions, white is only allowed to promote to bishop, black does not mate with a promoted piece" after I played six moves that win the knights and the rook on a1.
 
   
(Read Only)pid=10873
(9) Posted by Hauke Reddmann [Monday, Sep 2, 2013 11:08]

Ha! I would double and say "We need MOAR faily Nalimov! Whatever fairy!"
There is so much to find (I just point to an old Schwalbe article
about KS/KS h# - and that was ages before today). Even orthodox 4 men
on different board size will be interesting (e.g., when will R draw
against Q when the board is supersized? Will he draw at all?).
It can't be that hard to pimp existing software...

Hauke
 
   
(Read Only)pid=10874
(10) Posted by Juraj Lörinc [Monday, Sep 2, 2013 11:27]

Regarding Q vs R - Marc Bourzhutschky on initiative of John Beasley have found that on 15x15 Q wins (with exception of trivial positions of immediate mate, stalemate, capture or perpetual check). On 16x16 board there are positions where Q cannot force win. See http://www.jsbeasley.co.uk/besn/2004b.pdf

(This info via V. Kotesovec e-book Fairy chess endings on an n x n chessboard - http://problem64.beda.cz/silo/kotesovec_endings_on_an_nxn_chessboard_2013.pdf - very valuable material)
 
   
(Read Only)pid=10875
(11) Posted by Hauke Reddmann [Tuesday, Sep 3, 2013 11:34]

And that's why I love MPF so much - where would you find THAT
life-changing information elsewise?

Hauke
 
   
(Read Only)pid=10896
(12) Posted by Geir Sune Tallaksen Østmoe [Tuesday, Sep 3, 2013 20:13]

Wow. I never thought I would live to see a mutual zugzwang with Q vs R. I just wish I had a slight idea of why that is a mutual zugzwang,
 
   
(Read Only)pid=10900
(13) Posted by Sven Hendrik Lossin [Tuesday, Sep 3, 2013 21:50]; edited by Sven Hendrik Lossin [13-09-03]

"I just wish I had a slight idea of why that is a mutual zugzwang,"

That is exactly what I thought.

One year ago I was excited to see a mutual zugzwang with 2q vs. 2q, see
http://losso.blogger.de/stories/2082892/
 
   
(Read Only)pid=10901
(14) Posted by Dejan Glisić [Saturday, Sep 7, 2013 08:46]

Some solvers just like C- problems, some of them are cook-hunters! One older solver from Zagreb said: Computers have killed the joy of solving - why to solve if I can't find any cook or dual? :-)
 
 
(Read Only)pid=10911

No more posts


MatPlus.Net Forum General Oops they did it again