Author: Peter Berger
Date: 11:19:08 05/07/01
Go up one level in this thread
On May 06, 2001 at 01:48:05, Vincent Diepeveen wrote: >On May 04, 2001 at 18:46:19, Peter Berger wrote: > >Diep finds all pos except 8: d5, because it dislikes bad bishops >and i have no hope that it finds it before deep blue gets nullmove, >though it's just 0.09 or so difference. So Diep gets CCR 4 ? I am impressed - I think Tiger was the only one that got it so far . > >a4 position 10 takes some time to get (13 ply) so prob >doesn't get all points depending upon hardware. > >The Rfb8 move also is the victim of doubt. it wants to play it >from like 7 ply and on but then at 11 ply or so it flips >back to bc4 and then next ply rfb8 again etcetera. > >Wouldn't be surprised if it takes half an hour before it gets to rest >and plays Rfb8 only without ever again going to bc4 :) > >Note all these positions are strategic positions. giving just 30 seconds >to solve strategics things is nonsense. I disagree - take a look at CCR4 for example . Either your programs understands the concept of a minory attack or not - 15 plies won't let your program choose b4 here . For some of those positions 30 seconds is little - well : you can solve them in 2 minutes then and still get a point ;-) Another example is the one where you play a4 to hinder b5 ( not 22 ) - is it really possible to get it by search ? Dunno . Would _you_ find all the positional ones ( except 22 ) in 2 minutes ? - I bet you would ! > >we're talking about opening tactics. like 12 ply you need *anyway* >usual in opening. Especially that a2-a4 move. it's asking for >trouble with white with as only useful task for a4 to prevent Bb5 >getting exchanged. So in order to get awarded for that risk you need >first to calculate that it's tactical possible to do so! I completely agree to your assessment of CCR 22 - it's the most difficult one by far . I think Fisher or one of his team worked this out prior the game in analysis room - although I seem to remember Fisher claimed to have found it on the board . That's why I like to see the mainlines of the progs - if you can't find a4 in time 0-0 Bb5 is fine - or other lines with the Bb5 idea - that's a natural idea for black here IMHO . If your program doesn't know about good/bad bishops it will _never_ find a4 . It would be nice if you posted Diep's complete results . Cheers. pete
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.