Author: Andrea Griffini
Date: 16:10:40 07/22/03
Go up one level in this thread
On July 22, 2003 at 18:48:34, Rick Bischoff wrote: >I use hash tables, yes. No null moves. A big error that remained for quite a bit in my program was that when interrupting a search because the time was over there was no check preventing saving the interrupted search results in the hash table. This was (no surprise) source of terrible moves once in a while because the results for interrupted nodes were of course bogus, but when playing next move they were considered trustable until they were going to be searched deeper than the nominal depth stored in the entry. As it's normal that in a no increment game you use a decreasing amount of time it's not impossible that on next move you're not able to do a search deep enough to discover bad nodes. This is just one of the many possibilities... I think it's really impressive how many bugs can be present in a chess program and still leaving it "somewhat" working. HTH Andrea
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.