Author: Mike S.
Date: 15:26:21 06/22/02
Go up one level in this thread
On June 22, 2002 at 17:22:02, Robert Henry Durrett wrote: >(...) >If worse came to worst, maybe Fritz could be called upon to annotate them, but >what about positions which are a problem for Fritz? Really need for a GM to >flag the errors? To find positional mistakes, or things like missing the correct plan, you'd need to be a strong player. OTOH, sometimes those things aren't that difficult to spot... but it will most always be a matter of evaluation (of the human), not 100% proveable. It's easier with cominative motifs: Just search for games, where 1. the moves contain the search information (depth/eval/time) 2. the loosing program had quick change (reduction) of it's eval within 2 or 3 moves (2) means that probably the previous move (where the program still evaluated good) was a mistake. Misjudgement of the opponents reply. An example (from an experimantal game, not the normal starting position): Crafty 18.15 (cb-native) - Fritz 7 [D]8/5kpp/5p2/R1Bbp3/1Pq2n2/4QP2/6PP/6K1 w - - 0 42 (no test position, just to show how the interesting situation emerged) 42.Ra7+ 4.03/9 8 Kg8 1.50/11 7 43.Qd2 (h3) 4.04/9 8 43...Qb3 (Se2+) 1.22/10 8 44.Ra1 3.84/9 7 h6 (Sd3) 1.25/11 10 45.Rc1 3.90/9 7 Bxf3 (Lc4) 0.03/10 6 46.Rf1 1.12/9 18 Be4 0.00/10 4 47.Qd8+ (Tf2) 1.33/8 6 47...Kh7 -0.31/10 4 48.Qd1 1.12/9 6 Qf7 (Da3) -0.25/11 8 49.h3 (Dg4) 0.85/9 6 49...Nxg2 -0.72/11 10 50.Kh2 (Dg4) 0.45/9 8 50...f5 -1.09/10 6 51.Qe2 0.29/9 10 Nf4 -1.19/11 5 52.Qe3 0.00/10 5 Qa2+ -1.28/11 6 53.Qf2 -0.13/11 5 Qb3 (Da6) -1.56/11 4 54.Qg3 -0.47/11 5 Qc2+ (Sd3) -2.13/10 4 55.Qf2 (Lf2) -0.22/10 8 55...Qd3 -2.78/11 2 56.Be3 -3.43/12 5 Ng2 -2.25/11 4 57.Bxh6 -3.40/10 5 gxh6 -2.25/9 1 58.Qa7+ -3.18/10 5 Kg6 -1.84/11 3 59.Ra1 -3.05/10 5 Bd5 -2.66/10 10 60.Ra6+ -3.14/9 5 Kg5 -3.19/11 8 61.Qb6 -3.63/9 11 Kf4 -3.31/11 3 62.Qf2+ -3.79/10 6 Qf3 -3.53/13 8 63.Qxf3+ (Kg1) -4.12/11 4 63...Kxf3 -4.22/12 1 64.Rf6 (Kg1) -5.05/11 12 64...f4 (Sf4) -4.28/12 1 65.Rxh6 -6.17/12 8 Ke2 -5.19/12 2 0-1 Here, the challenge was to see 45...Bxf3!! coming. The evaluation of 45.Rc1 +3.90 was wrong, next moves eval was +1.12 which indicates the problem, in addition to the change of Fritz' eval (with more thinking time, Crafty will replace the 45th move i.e. by Rf1). Tactics are always very time- and speed related of course, more than positional things. Also, if a certain plan can be found or not, is also time & speed related (not only knowledge dependant). 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.