Author: Robert Hyatt
Date: 19:52:59 09/26/04
Go up one level in this thread
On September 24, 2004 at 23:14:54, Stuart Cracraft wrote: >Bob wrote: > >>>Initially I do a q-search after making each root move, and use this score to >>>order the moves at the root. If I am pondering I make sure that the move from >>>the PV is put first of course. >>> >>>After each iteration, I re-order the move list keeping the best move first, and >>>ordering the rest based on the size of the sub-tree (number of nodes) each >>>produced when it was searched. > >Compared with other types of move ordering for root ply 0 you did, >how did the above compare? Can you provide a general sense of the >effect/improvement you obtained. > >Did you do it with Cray Blitz as well? If not, how did it order ply 0? > >Stuart My intent here was that if the best move is not first, that it be second, or third, rather than Nth where N is large. CB used static eval + SEE as earlier versions of Crafty did. It then used node counts after iteration 1 was completed, just as Crafty does now. It was marginally better, but seemed to be more intuitively correct since the q-search is way more accurate than pure SEE applied to all pieces on the board as I was doing in CB days... It is just "a reasonable way of doing things" rather than "the best way to add 200 rating points to your program way of doing things." Not all reasonable ideas translate into significant numbers of rating points.
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.