Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Junior-Crafty hardware user experiment - 6th game

Author: Robert Hyatt

Date: 20:48:39 11/28/03

Go up one level in this thread


On November 28, 2003 at 23:39:04, Robert Hyatt wrote:

>On November 28, 2003 at 06:33:53, Jarkko Pesonen wrote:
>
>>There is definitely somehing wrong
>>
>>4k3/p5R1/7p/r6P/6P1/1B6/4K2n/8 b - - 0 50
>>Nxg4 ??
>>sacrafice for nothing, to get even worse end game which crafty thinks the
>>as good position than before sacrafice.

This looks dead drawn to my eye after Ng4 Rg4 Rh5.  the bishop and
knight can not possibly mate without the king's help.  It seems that
white has to bring the king up, provide a block with the bishop to
get the king past black's rook, all the time avoiding letting the
black rook get back to the second rank breaking the bind...

I just asked a GM to take a quick look at the position and asked "what
about Nxg4?"  He looked a couple of seconds and said "probably a draw".
Whether it really is is a good question, of course...


>>
>>Many games have turned from draw to loss because of lines below.
>>crafty let's enymy king so close that mate threat close and rook falls for
>>bishop.
>>
>>Jarkko
>>
>> ----------------------------------------------------------
>>|                                                          |
>>|   if one side is a piece up, but has no pawns, then that |
>>|   side can not possibly win.                             |
>>|                                                          |
>> ----------------------------------------------------------
>>*/
>>  if (WhiteMajors == BlackMajors) {
>>    if (TotalWhitePawns==0 && WhiteMinors-BlackMinors==1) can_win&=2;
>>    if (TotalBlackPawns==0 && BlackMinors-WhiteMinors==1) can_win&=1;
>>    if (can_win == 0) return(can_win);
>>  }
>
>While I won't argue that it may sometimes turn a draw into a loss, it has
>also turned _many_ wins into wins, rather than turning them into draws.  IE
>I have seen _many_ programs, commercial and amateur, that after winning a
>piece, trade down to (say) a KRN vs KN thinking they are a piece ahead.  They
>are, but it is a draw.
>
>My point is that the above is right 100X more than it is wrong...
>If I take it out, it will be wrong 100X more than it is right...



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.