Author: Chessfun
Date: 16:26:23 10/21/00
Go up one level in this thread
On October 21, 2000 at 19:14:14, Uri Blass wrote: >White is losing in this line and has no advantage but programs have the illusion >that white has advantage or at least equality and it pushes them to play the >wrong Bxc4. > >Some programs do not play Bxc4 at tournament time control because the illusion >is too deep for them to see it. > >The fact that tiger never considers Bxc4 as best may be a result of right >evaluation or result of not seeing the illusion because it is too deep. I think it is evaluation not as posted 1) but 2). >My experience with chess programs tells me that the common case with programs is >the case that the illusion is too deep for the programs to see. > >It can be proved by giving them the position before Ne5 and checking if they see >positive evaluation. Both Tiger's are negative before Ne5 and even forcing Bxc4 is a -1.00 swing from what it wants to play. Which is why I say 2). >Crafty17.11 does not see positive evaluation but it can see evaluation that is >close to 0 and it may push it to play Bxc4(I do not know,I remember that an old >version of Crafty could play the wrong Bxc4 after a long time). >Hiarcs7.32 can see positive evaluation only to fail low later and the same for >other programs. As I said Tiger and Gambit seeing it as a -1.00 swing they fail immediately and it seems to me the reason is they know by evaluation or knowledge that Bxc4 is wrong. Sarah.
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.