Author: Robert Hyatt
Date: 10:16:26 10/28/02
Go up one level in this thread
On October 28, 2002 at 00:12:47, Jeremiah Penery wrote: >On October 28, 2002 at 00:10:00, Robert Hyatt wrote: > >>On October 27, 2002 at 21:03:59, Robin Smith wrote: >> >>>On October 27, 2002 at 20:56:58, Robert Hyatt wrote: >>> >>>>On October 27, 2002 at 17:14:22, Ingo Lindam wrote: >>>> >>>>>Hello... >>>>> >>>>>I just thought about some questions from Bob that lead me to a problem. >>>>>I know the engines should have as much knowledge as the programmer can give to >>>>>it. But ofcourse in may searchtree may appear some positions I can't apply a lot >>>>>(or in worst case any) of that knowledge: material is equal and all the other >>>>>knowledge doesn't fit to current position (in the tree). Will my score 0 (or >>>>>atleast very near to 0)? >>>>> >>>>>If yes...I just ask me whether this could lead the computer to optimize the best >>>>>move sequence into complete knowing nothing about the position the sequence of >>>>>moves end in... as more as I evaluate the positions for my opponent the same >>>>>way? >>>>> >>>>>Shouldn't I better substract something from my score and add something on the >>>>>opponents score in case I know nothing about the position? >>>> >>>> >>>>Any such idea is very dangerous. It is amazing how a full-width search to >>>>reasonable depth will mangle your evaluation if you have a big hole in it. >>> >>>Bob, >>> >>>But isn't this exactly what you do in Crafty for locked pawn positions .... >>>penailze Crafty for being in such a position? >>> >>>Robin >> >> >>Yes.. But the point is that Crafty has specific code to recognize that kind >>of position, and a specific (and appropriate) score for such. >> >>That is quite a bit different than saying "I don't know how to evaluate this >>so I will just call it zero. Or something less than zero." > >I think they were saying not to call it zero, but to take the normal score you >would give for that position and lower it some degree corresponding with your >non-understanding of the position. So if the evaluation is already -9.00, it >would become -10.00 (or worse) if there are parts of the evaluation you don't >understand. This way, the search will try to AVOID those parts of the tree more >than normal. I interpreted it as "if I don't know anything at all about the positional features, take the positional score +/- some constant as the score." "not knowing about the positional features" is very dangerous, because your search will tend to seek out those positions where your understanding is wrong and the scores are higher than they should be..
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.