Author: Robert Hyatt
Date: 09:04:39 09/22/00
Go up one level in this thread
On September 22, 2000 at 09:48:49, Kim Roper Jensen wrote: >Hi All > >I saw the post about variable selectivity, and wondered. > >How about having a "turbulence" ( this comes from an article in the book >Computer chess Compendium ) variable, that is defined by the number of moves in >the position, like high number of moves its a open position, the pieces have >more squares to go to, low number of moves the pieces have less squares to go >to. Then the selectivity could be adjusted so at high turbulence the selectivity >should be low( more tactical traps so we have to look at more continuations), >low turbulence the selectivity should be high( more positional traps so we have >to look at less continuations, but deeper on the positions) > >Is this something thats tried ?? > >With regards Kim That is actually backward from the way selectivity has to work to be effective. If you have a large number of moves to choose from, if you don't toss many of them, then your search will be way shallower than when you do. That is the catch-22 of selectivity in its many forms. The more you toss out, the deeper you go, and the more you see, and the more you also miss. The less you toss out, the shallower you go, the less you see, and the less you also miss. Which is better? :)
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.