Author: Uri Blass
Date: 07:44:48 11/27/00
Go up one level in this thread
On November 27, 2000 at 08:46:37, Dan Ellwein wrote: >On November 27, 2000 at 04:24:02, Jouni Uski wrote: > >>If Rebel Tiger and Gambit Tiger are stronger and with Windows interface, why >>even install Rebel Century 3 for DOS? Just wondering... >> >>Jouni > >Jouni > >Ed and Christophe can answer this better then me... > >but i believe that Rebel Century 3 is designed to play best against humans... > >while Tiger and Gambit Tiger are designed to play best against computers... > >also... > >when giving Rebel Century 3 a given position... > >its assessment of the position (ie the score whether positive or negative) > >is regarded as being accurate and can be trusted... I almost did not try the last Century3 but I believe that the program is a preprocessor like many programs(I know that the previous version was a preprocessor and I believe that Ed will tell us if he changes it). It means that it is possible to see a move like 14.0-0 with positive evaluation for white at all depthes and after 14.0-0 see negative evaluation for white at all depthes and it means that Rebel can tell you contradicting things about the same position. >(obviously, there may be exceptions to this depending on how unusual the >position is) I do not think that preprocessing is exception in Rebel but Ed can tell us better. Many top programs do preprocessing and the interesting question is the size of the preproccesing. It is a question that only the programmer can find out because we do not know the static evaluation of the programs. The test should be done by the following way. 1)Take a random node that you evaluate(you can use database of the program's games for getting these nodes). 2)Calculate the evaluation of the node. 3)Calculate the evaluation of the same node assuming the root position is 2 plies earlier. 4)Calculate the difference between the evaluations. You get a number. Continue with this experience and get a lot of numbers. I will be interested to know the distribution of these numbers for all the top programs. You can do the same experiment with 1 ply and the reason that I suggest 2 plies is to check only preprocessing and not asymmetry in the evaluation. If you want to check also preprocessing that is a result of asymmetry use 1 ply Uri
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.