Author: Mike S.
Date: 12:45:17 01/26/04
Go up one level in this thread
On January 26, 2004 at 11:28:38, Uri Blass wrote: >On January 26, 2004 at 09:57:51, robert flesher wrote: > >>Ruffian plays NF8?? when Nf6 holds to draw? >>[D] 1k6/1P5n/2K1B3/8/8/8/8/8 b - - 0 84 > >I do not think that it is a bug. >It is probably a design decision of the programmer to use null move pruning >even in simple endgames. Yes, I think this is just the common nullmove/zugzwang problem. My general impression is, that the "old established" engines have an improved mullmove implementation that can avoid that problem often (maybe a kind of zugzwang detection, I don't know), while amateurs and neo-profis don't have that improved nullmove algos yet. Ruffian 1.0.5 does not have problem here (with, but also not without 4-piece tbs.): Analysis by Ruffian 1.0.5: 1...Nf6 2.Ba2 Nh5 3.Bb1 Ng3 4.Bd3 Nh1 5.Kb6 Ng3 6.Bc2 Nh1 7.Bb3 Ng3 8.Ba2 Nh1 9.Bb1 Ng3 10.Bd3 Nh1 11.Bc2 Ng3 12.Bb3 Nh1 13.Ba2 Ng3 14.Bb1 = (0.00) Depth: 1/1 00:00:00 = (0.00) Depth: 26/39 00:00:05 2818kN without tbs.: 1...Ka7 +- (3.33) Depth: 1/3 00:00:00 1...Nf6 2.Bf5 Ne8 3.Kb6 Nc7 4.Ka5 Kxb7 5.Bc2 Kb8 6.Bb1 Ne8 7.Ba2 Nc7 8.Ka4 Ne8 9.Ka3 = (0.00) Depth: 1/3 00:00:00 = (0.00) Depth: 16/24 00:00:04 2378kN 1.Nf8? Bf8 (=only winning move!) creates zugzwang and mates quickly, which Ruffian 1.0.5 sees immediatly. The older Fritz version 5.32 encountered such nullmove problems sometimes (IMO more often than recent ones), but not here. So I think you may be right with the assumption that some engines turn nullmove off when only very few pieces are on the board. Regards, M.Scheidl
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.