Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Mate in 17: Chessmaster 6555 solve in 1sec!

Author: leonid

Date: 17:03:10 10/20/00

Go up one level in this thread


On October 20, 2000 at 18:55:26, Bruce Moreland wrote:

>On October 20, 2000 at 16:21:13, Uri Blass wrote:
>
>>On October 20, 2000 at 16:00:37, Bruce Moreland wrote:
>>
>>>On October 20, 2000 at 15:19:32, Andreas St. wrote:
>>>
>>>>Hi,
>>>>
>>>>here a position from CSS 4/93:
>>>>
>>>>white: K f1 ; Q e2 ; R g1,h5 ; B c1 ; N e1 ; P a4,d3,e5,f2,h4
>>>>
>>>>black: K h8 ; Q b1 ; R d8,f8 ; B b7 ; N b4,c6 ; P a6,b6,e6,f5,g7,h7
>>>>
>>>>
>>>>FEN: 3r1r1k/1b4pp/ppn1p3/4Pp1R/Pn5P/3P4/4QP2/1qB1NKR1 w - - 0 1
>>>>
>>>>
>>>>wKf1,Qe2,Ne1,Bc1,Rg1,h5,Pa4,d3,e5,f2,h4/bKh8,Qb1,Nb4,c6,Bb7,Rd8,f8,Pa6,b6,e6,f5,g7,h7
>>>>
>>>>1. move: Rxh7+
>>>>
>>>>Mate in 17. I think, no Program is as good as Chessmaster 6555 here. Right?
>>>>
>>>>
>>>>Greets Andy
>>>
>>>This is a problem where moves by white force situations where black has very few
>>>replies.  You can solve this quickly by extending if black is in check and has
>>>few ways out.  Most programs have this extension or something similar, the
>>>differences are in the degree to which they try to constrain it.
>>>
>>>bruce
>>
>>Not all of the moves of white are checks.
>>The main line begins with
>>1.Rh7+ Kxh7 2.Qh5+ Kg8 3.Rxg7+ Kxg7 4.Bh6+ Kh8 5.Bg5+ Kg7 6.Qh6+ Kf7 7.Qf6+ Kg8
>>8.Qg6+ Kh8 9.Bf6+ Rxf6 10.exf6 and here black is not in check.
>>
>>I believe that chessmaster knows to extend after 10.exf6 because of the fact
>>that white threats mate in 1.
>>
>>Uri
>
>They don't all have to be checks, just almost all of them.  I am surprised that
>CM can find these and still operate reasonably in other cases.
>
>bruce

You are right! It is not all the time obligatory to have checking moves to be
revisioned in each ply. It could be just some fixed number of moves that will be
seen each time in each ply. Only when some ply contain checking moves, they goes
at the head of the moves chain.

Search for mate can be done in one billion variations where many of them will be
very close in efficency. One way to search is to see all moves in initial one of
few plys. Later search goes in selective way. This search will start with all
kind of moves and still will be very rapid. The other way is to see fixed number
of moves where checking moves represent only one part of moves. Other way of
search is to look into those moves that effect king's vital lines, and so
like...

Leonid



This page took 0 seconds to execute

Last modified: Thu, 15 Apr 21 08:11:13 -0700

Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.