Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: engines and openings: can engines guide human play?

Author: Dann Corbit

Date: 13:35:37 02/27/06

Go up one level in this thread


On February 27, 2006 at 16:19:42, Joseph Ciarrochi wrote:

>Hi folks,
>
>I am trying to learn a second defensive. system for black, other than the
>accelerated dragon. I have been trying out the scandanavian, and I do ok with it
>in practical blitz play (no correspondence yet).
>
>One way I learn about openings is to pit two good engines in it and look at the
>plans they come up with. In the case of the scandanavian, I will put a weaker
>engine in as white (spike 1.1.), and rybka in as black, just because i want to
>see how black wins.
>
>Well, with the critical scandanavian lines, black has tremendous trouble
>winning, even when it is a stronger engine. I tried giving Rybka more time (200
>to 300% more time), but still rybka has trouble against spike 1.1..   It isn't
>just specific to spike and rybka. I ran a little tournment with 20 games (rybka,
>fruit, fritz9, spike1.1. and ruffian 2.1) and white won 14.5 to 5.5. in the
>critical scandanavian line.
>
>The scandi scores well based on huge numbers of games on icc blitz (49% for
>black; but we don't know the strength level of the players), but does more
>poorly amongst higher level players in otb tournements.
>
>
>So....Should i base my opening choices on this kind of computer anlalysis....
>The computers think black has long term disadvantages in some scandi lines. Or
>are computers irrelevant to my choice.

For openings, I would definitely use a reference book like MCO or the like in
conjunction with the computer analysis.  Chess openings are the weakest point
for computer analysis. Especially for gambits, I would take any computer opinion
with a large grain of salt.



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.