Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: is everybody but rybka just plain wrong; or can we not trust engines

Author: Dann Corbit

Date: 17:42:06 12/21/05

Go up one level in this thread


On December 21, 2005 at 20:34:15, Joseph Ciarrochi wrote:

>Based on these posts, It seems like all the programs except rybka and fritz 9
>think that white has a strong advantage. (rybka at depth = 17 as +.48, which is
>a typical white advantage, but not that strong).
>
>I think these non-rybka engines are getting it wrong....if you look at
>(http://www.chesslab.com/PositionSearch.html), black and white are close to
>even(33% white; 25% black) (granted based on only about 12 games). Dann, how
>many games was your "57% for white" based on?

57% points, not wins.

>Also, if you pit the engine against itself, there does not appear to  be a
>strong white advantage. i.e. white seems to only win about 50% ((i'll update
>this after 40 or so games of different engines playing themselves)
>
>How much can we trust engines to evaluate opening variations?

Not at all.  Try this one (and let your engine think as long as you like):
r1bqk1nr/pppp1ppp/2n5/b7/2BpP3/2P2N2/P4PPP/RNBQ1RK1 b kq - am dxc3; bm Nge7;

In a gambit, the engines are often as blind as a bat.

In a standard opening, they do very poorly.  See Kurt Utzinger's NBE (No Book
Experiment):
http://www.utzingerk.com/nbe_2003_english.htm

>best
>Joseph
>
>
>
>On December 21, 2005 at 20:05:33, John Merlino wrote:
>
>>On December 21, 2005 at 19:03:43, Dann Corbit wrote:
>>
>>>On December 21, 2005 at 18:32:52, Dann Corbit wrote:
>>>
>>>>Which results in this position:
>>>>[D]rnb1kb1r/pp2pppp/1qp2n2/8/3P4/2N2N2/PPPB1PPP/R2QKB1R w KQkq - 0 7
>>>
>>>White gets 57/5% of the points, in my database (well above average).
>>>
>>>Preferred moves seem to be Bd3, Bc4, Na4
>>>I see Rb1 played occasionally in my database, but it looks inferior.
>>
>>CM9000's opening book ends at 6...Qb6. CM9_R1 on a P4-2.4 likes the position
>>very much for White and says this:
>>
>>Time	Depth	Score	Positions	Moves
>>0:00	1/6	1.39	39467		7.Bd3 c5 8.Bb5+ Nbd7 9.dxc5 Qxc5
>>					10.Be3 Qb4
>>0:00	1/6	1.40	42928		7.Na4 Qd8 8.Bd3 b5 9.Nc5 e5 10.O-O
>>					exd4
>>0:00	1/7	1.29	58085		7.Na4 Qc7 8.Bd3 Nbd7 9.O-O e5 10.Qe2
>>					Bd6
>>0:01	1/8	1.19	122963		7.Na4 Qc7 8.Bd3 e6 9.O-O Bd6 10.Nc5
>>					O-O 11.c4
>>0:02	1/8	1.46	266962		7.Bd3 Qxb2 8.Rb1 Qa3 9.O-O Nbd7
>>					10.Rb3 Qa5 11.Qe2
>>0:04	1/9	1.27	634459		7.Bd3 e6 8.O-O Be7 9.Rb1 O-O 10.Qe2
>>					c5 11.d5 Bd6
>>0:11	1/10	1.44	1610449		7.Bd3 e6 8.O-O Be7 9.a4 O-O 10.a5
>>					Qd8 11.Qe2 Nd5
>>0:29	1/11	1.24	4567995		7.Bd3 e6 8.O-O Bd6 9.a4 O-O 10.a5
>>					Qc7 11.Qe2 Ng4 12.h3 Nf6
>>1:35	1/12	1.36	15184957	7.Bd3 e6 8.O-O Bd6 9.a4 O-O 10.Qe2
>>					Nbd7 11.a5 Qd8 12.Ne4 Nxe4 13.Qxe4
>>
>>jm



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.