Author: Jonas Cohonas
Date: 00:28:02 10/15/05
Go up one level in this thread
>Hello Jonas,
>it is great that you try new settings. But i think you changed your settings to
>conservative. To raise the value to 102 changes nearly nothing. Of course i only
>know how these settings are calculated in fruit 2.1 but i thing this hadn´t
>changed in fruit 2.2. in fruit 2.1 the calculations are as follows:
Thanks for the feedback, it might not change a whole lot and that was pretty
much my goal as i don't think the default values are way off, but in practical
play it seems to have a positive impact ie 10 - 2 over Shredder 9 with only one
loss, but i will not make any final conclusions before i have enough games for
the error margain to be more balanced.
I will however try and highten the values later on in v02 when testing for v01
is over, which was my plan anyways.
>PawnStructureWeight = (option_get_int("Pawn Structure") * 256 + 50) / 100;
>pawn_eval=((opening[White] - opening[Black]) * PawnStructureWeight) / 256;
>for example:
>PawnStructureWeight = (102 * 256 + 50) / 100;
>=> PawnStructureWeight = 261
>opening[White] is the evaluation for the white pawn structure.
>opening[Black] is the evaluation for the black pawn structure.
>say opening[White] = 60 and opening[Black] = 10
>then pawn_eval = ((60 - 10) * 261) / 256 = 50.97 => pawn_eval = 50
>thats the same score as the default value Pawn Structure 100
>so my suggestion would be to raise the value much higher than only by 2.
I am affraid that my observations that led to changing the values in the first
place was based on my understanding of the game as a human and basic chess
principles, math always bored me to tears so the above completely escapes me :)
Regards
Jonas
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.