Author: Inmann Werner
Date: 07:41:28 02/19/99
Go up one level in this thread
On February 19, 1999 at 08:47:20, Steffen Jakob wrote: >Hi Werner! > >On February 18, 1999 at 18:01:43, Inmann Werner wrote: > >>Hello! >> >>After looking at the results of the other programs and mine, I got really >>disturbed. After ply 8, my program calculated 2.454.000 positions, and the >>others do around 700.000! >>I checked my extensions and put out my check extension, and suddenly my program >>only needed around 600.000 nodes. >>I think, i do something very very wrong! >>I extend on any check that comes when there is only 2 or less distance till >>quiet search. This seems to blow up the tree extensively. I do this extension >>up to 5 times within one search, so sometimes in a 8 ply search , I search to >>ply 13. >>So my question is. >>When should I extend on check? > >You have to make sure that you don't use extensions which are > 1 ply. Otherwise >your tree will explode. Another thing which is useful is to reduce the >extensions when you are searching very deep (e.g. if(ply > 2*root_depth) >extension >>= 1). > >Greetings, >Steffen. Thanks Steffen! I again found some big bug. The bugs don´t end.... I wanted to extend on beating sequences on the same field, which do not change material too much, but as bug is, I extended on all beatings possible!!! You can imagine the tree explode? Sometimes I am confused, why Inmichess with all this bugs play chess, and not so bad chess. In the last 2 weeks this is th 3rd big bug, I found. So i am looking forward to test my new version! I hope Hossa will eat the bugs.. :-) Werner
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.