Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: q-search question

Author: James Robertson

Date: 19:40:53 08/02/99

Go up one level in this thread


On August 02, 1999 at 19:41:14, Landon Rabern wrote:

>How much should adding a q-search to my program improve its performance?  When
>playing games against the version without the q-search it seems about even.
>Both are set at 7 ply.
>
>Thanks,
>
>Landon

Qsearch should beat non-qsearch 100%. For instance, say you only have a 4 ply
search and no qsearch. From the starting position, white cannot take anything if
black does not want him to in 4 plies, so the pv might look like:

1. e4 e5 2. Nf3 Nf6

But with 5 plies and no qsearch, black thinks that no matter what he will lose a
piece, because at every 5th ply white can take something, and black can't
recover it, so he _will_ play a stupid move. 6 plies always thinks black wins
material (because he has the last move), 7 always thinks white does, and so on.

If you have some sort of recapture extension, play will be much better, but the
depth searched will collapse (it is like doing a qsearch searching every move
instead of just captures). I will conservatively estimate that qsearch will
_still_ win more than 90%.

James



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.