Author: Dann Corbit
Date: 11:46:25 12/29/05
Go up one level in this thread
On December 29, 2005 at 05:25:55, Tord Romstad wrote: >On December 28, 2005 at 18:39:28, Dann Corbit wrote: > >>I would test it in your eval for benefit. Tord does not calculate queen >>mobility, for instance. > >This is incorrect. I do use queen mobility, in Glaurung Mainz as well >as in the most recent development versions. You probably misunderstood >a post of mine in a different sub-thread, where I wrote that I haven't been >able to measure any difference in strength when I switch queen mobility >on or off. Because queen mobility doesn't seem to hurt, I have chosen >to leave it enabled. > >By the way, I would be interested in learning how to compute the kind >of mobility I use in Glaurung by using bitboards. I haven't found >anything faster than counting the non-zero bits of a bitboard, which >is actually slower than my current code. You can precompute, so that there is no counting to be done. Hard to get faster than that. >For each piece on the board (except pawns and kings), I look at all >squares this piece attacks, and divides them into three classes: > >a) Empty squares and squares occupied by enemy pawns or pieces. >b) Squares occupied by friendly pieces >c) Squares occupied by friendly pawns > >I then compute the mobility bonus for the piece by a table lookup: > >mobility[side] += > BishopMobilityBonus[(squares of type a) - (squares of type c)]; > >In other words, I use a negative mobility increment for attacking >squares occupied by friendly pawns. > >Tord
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.