Computer Chess Club Archives




Subject: Re: gothmog's eval

Author: Tord Romstad

Date: 08:55:48 02/17/04

Go up one level in this thread

On February 17, 2004 at 08:13:45, martin fierz wrote:

>On February 17, 2004 at 07:36:25, Tord Romstad wrote:
>>On February 17, 2004 at 04:53:54, martin fierz wrote:
>>>hi tord,
>>Hello, and thanks for the comments and suggestions!
>>>two other things i noticed:
>>>-> gothmog 0.4.5 evaluates KRN vs KR as about +4 for KRN. it's usually a draw
>>>though... i didn't check, but you should, KRB-KR, KR-KB, KR-KN - all usually
>>>draws. this should be easy to fix
>>I think you must have confused Gothmog with some of your other sparring
>>partners here, or you are using a version much older than 0.4.5.  The score
>>in KRNKR endgames should never be more than about +1 pawn (usually less)
>>for the stronger side, unless I have some bug.  I tested several different
>>KRNKR endgames now just to be sure, and I didn't notice any problems.
>i'll have to check again in that case. i'm pretty sure that i have 0.4.5 and i'm
>also pretty sure it was gothmog and not frenzee. i'll let you know what i find -
>but if you have already fixed that it's not so important  :-)

I certainly haven't fixed anything like that; at least not recently.
I just checked my old sources and found that the KRNKR code has been
there ever since the pre-historic (and pre-release) version 0.2.4.
Nothing has changed since then.  I have seen lots of KRNKR, KRBKR,
KRKB and KRKN endgames during the last half year, but I never saw
any problems.  If you really observed a +4 score in KRNKR, it must
have been a bug, but apparently a very rare one.  It definitly does
not occur except in a very tiny fraction of the possible KRNKR endgames.


This page took 0.04 seconds to execute

Last modified: Thu, 07 Jul 11 08:48:38 -0700

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