Author: Normand M. Blais
Date: 15:11:19 07/22/03
Go up one level in this thread
On July 22, 2003 at 14:38:35, Dieter Buerssner wrote: >On July 21, 2003 at 23:53:48, Normand M. Blais wrote: > >>After the game, I decided to let the program analyse the position again just for >>fun. But then I notice that, after finding the correct move, the pv line at >>depth 6 was ending in a checkmate while the score at that depth was 0.02. > >It is a typical consequence of using quiescence search. The details may depend >on several subtle points of the qsearch implementation. As an example, assume a >"normal" qsearch, that will return a value, when the eval > beta. But you might >be in check at this time, and actually, it might be mate already. If you don't >check for this, such things will happen. Whether checking espescially for such >things (and therby avoiding such stupid looking mainlines) will help an engine >or not, is a different question. > >Regards, >Dieter I think you are right. If I don't update the PV in the Qsearch then the move given checkmate is not print. NMB
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.