Author: Vincent Diepeveen
Date: 20:01:41 12/19/99
Go up one level in this thread
On December 18, 1999 at 22:01:20, Robert Hyatt wrote: >On December 18, 1999 at 19:21:13, Tom Kerrigan wrote: > >>On December 18, 1999 at 09:15:26, Robert Hyatt wrote: >>>This has already been done. To prevent the oscillation you mention. You can >>>simply 'fudge' up the score for the first move, and adjust alpha/beta to match, >>>or you can leave the score for the first move alone, but search all remaining >>>moves with score+N, score+N+1. I tried it many years ago, but didn't like it. >>>If you believe your eval, then you believe that +.05 is worth something... >> >>It seems like you could save a LOT of time doing this in some positions... that >>isn't worth it? Do you have some statistics? >>-Tom > > >No statistics, just observations. We played a bunch of games using Cray Blitz >on a VAX vs various (at the time) commercial chess machines like the supercon >and so forth. shifting the window does make the search go faster, and it helped >in those positions where you are oscillating between two developing moves like >Nf3 and Nc3 (for example). But in middlegame/endgame positions, I often found >myself wishing (after playing over a game with the shift turned off) that it >had played the move that was only slightly better score-wise... > >Null-move already effects scores at the root enough... In DIEP i only see in the endgame now and then root score changes. Not so in the middlegame. No difference there.
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.