Author: Howard Exner
Date: 08:59:47 01/08/99
Go up one level in this thread
On January 08, 1999 at 08:56:26, Robert Hyatt wrote: >On January 08, 1999 at 04:22:09, Howard Exner wrote: > >>b5k1/7p/4p1pP/4P1P1/8/8/8/N5K1 w >> >>Here is a position based on a game between Cm6000 and Fritz 5.32. >>( the recent game with the move d3 thread). >>I created this position to see if Chess programs have >>an accurate assessment of the relative values that the Bishop and >>Knight have. Material is equal yet all endgame authors I've read usually >>comment on positions like this in the following manner - "White has >>a huge advantage as all the pawns are on the same side of the board. >>Furthermore whites pawns cannot be attacked and once the powerfull >>white Knight reaches f6 the game will be over." >> >> >>How will programs eventually come to deal with assessing this position >>correctly? Will this be a kind of position that programs will never >>quite understand? Do any programs penalize the bishop here or >>conversely, give the Knight extra value in this position? > >I do. And I've been doing this for some time... my code simply says that if >pawns are on _both_ wings, the bishop is given a significant bonus over the >knight. If the pawns are on one side of the board, then this is not done. It >has worked pretty well, although I wouldn't say the 'value' is tuned perfectly >just yet... and it isn't always correct to do this. But it is right most of >the time... This is nice to see. I posted an example of Crafty's KKup2 game with Rebel where I was curious how programs would assess even this middle game. To my mind when I saw Crafty's knight plunk itself ,unmolested in the middle of the board I was wondering if Rebel would survive. It did but just barely since the endgame had pawns on both wings.
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.