Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Brute Force vs. Selective Search Re: Fernando & Jim

Author: Robert Hyatt

Date: 20:11:54 01/24/99

Go up one level in this thread


On January 24, 1999 at 18:41:13, Amir Ban wrote:

>On January 24, 1999 at 17:31:15, Melvin S. Schwartz wrote:
>
>>I contacted a technical advisor from Saitek in Hong Kong who assured me that
>>Brute Force was superior to Selective Search. His reason was that Brute Force
>>searches more extensively and therefore minimizes the risk of an occasional
>>oversight. Apparently the Selective Search is quicker but not as thorough.
>
>This can't be right in this day and age. Does Brute Force even exist any more,
>other than in the initial implementations of amateurs ?
>
>I estimate that by making a decision to be brute-force (no forward pruning, no
>extensions), you lose around 300-400 rating points to start with, meaning that
>you won't score more than 15% or so against a strong program.

I don't call brute force "no extensions".  IE the original chess 4.0, which
was the first successful brute-force program, had extensions for getting out
of check, etc...  as did we all back then.

I've always considered brute force to be 'no chess-knowledge-based forward
pruning'.  IE null-move is still a brute-force idea, although it stinks of
selectivity when you look at what it is doing.  But it is ignorant about the
game of chess, generally, and works just as well in other games that don't have
zug problems...

from that perspective, the Saitek guy was probably 1/2 right... because I think
the 'good commercial programs' are brute force early and selective later.
where 'early' is close to the root, and 'later' is closer to the leaves.

I call 'selective' a program like my original chess program where we did a _lot_
of plausibility analysis and weeded the ply-1 move list from N down to 6-9 moves
max...  ditto for every ply in the tree.  _that_ is selective.  :)  and _very_
dangerous too... :)




>
>
> So,
>>from this I would assume that IM Larry Feldman was wrong or misinformed. It is a
>>shame that he has not responded to my query. However, it does appear he was not
>>correct in what was stated in his review.
>
>What statement does this refer to ? Perhaps you mean IM Larry Kaufman ?
>
>Amir



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.