Author: James Robertson
Date: 13:33:08 06/12/99
Go up one level in this thread
On June 12, 1999 at 15:24:33, Frank Schneider wrote: >On June 12, 1999 at 14:33:42, James Robertson wrote: > >>These are some of the distressing things testers have said about my program. >>Does anybody else recieve emails with stuff like this in them? >Yes. > >> >>James >> >> >>"... that is great to analyze mistakes and other mystical things with JRCP." >> >>"Sometimes JRCP neglects the evolution of it's chessmen." >Constructive criticism isn't that bad. > >> >>"I can't believe my eyes [Rhg1??]. This is remembering the early days of >>computer chess." >... or the early days of a program > >> >>"There is certainly something wrong with King safety." >> >>"JRCP draws with LambChop. After analyzing the game, I saw JRCP gives stalemate >>in a total lost position. This is certainly a strange way to get a 1/2 point." >> >>"All other programs were playing Qxd8. YOUR program plays Kxd8, somewhat >>unexpectedly." >> >>"It seems JRCP does not know about King safety." >> >>"Today I have prepared a testsuite from blunder-games of JRCP." >That is the best that can happen to you. Could you ask him/her to prepare >a similar testsuite for Gromit? > >> >>"> Big bug in King safety fixed. >> >>Is it?" > >Frank All of the advice was excellent. The blunder testsuite was one of the best things that happened to my program. When I fixed the bad moves in the blundersuite, my program played much, much better. I am now asking people to instead of send me games that my program played, send me positions that my program goofed on. I think I will be able to improve my program more rapidly that way. 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.