Author: Andrew Williams
Date: 12:01:46 06/21/03
Go up one level in this thread
On June 21, 2003 at 14:03:14, Andrea Griffini wrote: >Sorry for the funky topic... but i wonder how do strong program >evaluate advanced pawns. I added a bonus table for free pawns >and multiple free pawns that get up to 300 when they're on the >7th rank, but this makes the program quite happy to sac pieces >to get there. Sometimes it works (especially against humans in >blitz games) but sometimes it doesn't. For example in > >2kr1b1r/ppp1nppp/2n1pq2/3P4/8/2N1BB2/PPP2PPP/R2Q1RK1 w - - 0 11 > >my program was happy to play dxc6 (score +2.3) in a 3 0 game >against another computer even after analyzing just a few >plys. So may be in this position really dxc6 is a winning >move (really can't tell from my 1500 ELO rating :D... and >the chess programs I have don't agree) but this is not >the point. My program likes dxc6 with a score initially to score. A little bit of superficial analysis suggests that it wins. (Even if it doesn't, it's a lovely move). > How do program face this problem ? Another >similar case is > >8/7p/5k2/5p2/p1p2P2/Pr1pPK2/1P1R3P/8 b - - bm Rxb2; id "WAC.002" > >Rxb2 looks obviously quite interesting even for a 1500 player :) >but how do program get to play it ? Do they really analyze all the >pointless checks the king will undergo or they will just consider >"positionally" that two pawns that far are well worth a rook in >these conditions ? > >Andrea In this position you have to consider the fact that after: 1...Rxb2 2.Rxb2 c3, there are *two* pawns aiming to promote. A rook can't stop these without some help. Andrew
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.