Author: Albert Silver
Date: 14:22:06 08/31/04
Go up one level in this thread
On August 31, 2004 at 15:47:46, Kurt Utzinger wrote:
>On August 31, 2004 at 15:46:02, Kurt Utzinger wrote:
>
>Hi Albert
>Are you sure that this solves your problem?
>If the GUI had loaded the default Rebel.eng
>then your test should have been identical with mine
>as discussed at SCW [http://f50.parsimony.net/forum200336/messages/3114.htm]
>but in the given position your Pro Deo played 1...Nxb3? and mine
>1...c4
>Regards
>Kurt
[D]1r1r2k1/4ppb1/7p/n1ppP2n/6P1/1PP1BN2/P4P2/3RR1K1 b - g3 0 26
True, but the 1...Nxb3 was played using a personality I have been working on. In
it I used the new function (from http://members.home.nl/matador/engine.htm):
[Pruning ? MISC_58] * Recalculate passed pawns
[M58_Table ? Normal] * normal|midg|endg|strong
I had set it to Strong, but even using Normal is considers 1...Nxb3? for a
couple of plies. The only thing not clear to me is whether this is because it is
over-evaluating its passed pawns after it takes the a-b-c or whether it is
unusually afraid of letting white push b4, in which case it might be
overevaluating the danger of white's passed pawns. Not that it isn't bad, but
giving the piece is worse.
It's worth noting that even making the correct changes in the ENG files, and
making sure they *all* have the Clear Hash = Always, I don't always see the
right ENG file loaded, so something is still not 100% right.
Albert
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.