Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Knee jerk reaction!

Author: Vasik Rajlich

Date: 02:51:26 09/16/04

Go up one level in this thread


On September 15, 2004 at 10:09:33, Peter Berger wrote:

>On September 15, 2004 at 07:19:24, martin fierz wrote:
>
>>however, with humans just as with computers, that is just "masking" your
>>inability to come to grips with certain positions (e.g. the queen-pawn-borers
>>are usually tactically weak), and if you want to improve, you should rather try
>>to get comfortable in other positions too rather than trying to avoid them - you
>>(or your program) will become a more all-round player which is never wrong...
>>you can always go back to limiting your repertoire if it doesn't work out.
>>
>
>Agreed. I also think that when your program has many troubles with a line other
>computer players excel in, it shows areas to improve. E.g. the Najdorf often is
>mainly about tactics and king safety. When you choose to avoid the line, the
>principal problem still stays.
>
>Of course there are also practical considerations e.g. time availlable for book
>preparation, same as with human masters, but usually avoiding playing the
>sharpest and critical lines, especially with black, means much lower chances to
>play for a win.
>
>Peter

Now I am starting to understand why this was such a long thread ;)

I agree that it's good in theory to be able to play many different types of
positions, both for a human and for a computer. But - in practice - it's not so
easy.

One issue is the learning curve. Martin likes to change up his openings every
year or two. Well, I haven't changed my openings in about 10 years ;) Like most
chess players, I lose games mainly because of tactics, sometimes because of
strategy, and almost never because of the opening. I don't think learning some
new variations would really change anything.

For a professional chess program, though, learning curve is no excuse. That's
where the second reason comes in - sometimes you are just naturally better at
something. For example, in computer chess, the more selective your search, the
better you will be in wild tactical positions, and the (relatively) weaker you
will be in quieter positions. (Of course this is just a theory, and it may not
apply to bigger hardware, etc.) So it won't be just a question of fixing up some
holes - we're talking about a question of basic approach.

Vas



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.