Author: Howard Exner
Date: 19:10:34 10/11/00
Go up one level in this thread
On October 11, 2000 at 02:21:40, Jim Monaghan wrote: >[D] 2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - 0 1 > >A. Petrosian - Hazai, 1970 As Dr. Nunn states in his annotations Black's >position is quite bad and he tried Qa7-b6 last move and White couldn't resist >snapping off the queen ... and drawing. The correct scheme is to reject the >"gift" and play Qd2, Kb3, Nc3, Ka4, Na2-c1-b3 and Qxa5 with Black just watching. > I gave this to Crafty 17.13 ... > >20 4:33 5.32 1. Nxb6+ cxb6 2. h4 gxh4 3. Qc1 h3 > 4. gxh3 h4 5. Kb3 Kd7 6. Kc3 Bf2 7. > Kd3 Ke7 8. Qd2 Bg3 9. Qc2 <HT> > >Is there a program that "understands" the idea of a blockade, or is this still a >tough area for programmers ... where ply depth doesn't really help but something >in it's evaluation function? From a human standpoint the concept is not that >complex although A. Petrosian stumbled ... I'm glad it wasn't one of my heros, T. Petrosian. He would have spotted this in a flash. These are lines that could show up from your position after white wrongly plays Nxb6. I was curious if I could defeat a computer from the sequence Nxb6 cxb6 Qb1. [D]8/2k5/1p1p4/pPpPp1pp/2P1Pp2/P4PbP/K5P1/1Q6 w - - bm h3h4; Only move to draw is h4. Bh4 will lose to Qh1 or even the immediate g3. Any king move and white plays h4 (computers should detect this h4 move). The entire point is to get a pawn break so the Queen can infiltrate. Here's a sample win. 1. Qh1 Kd7 2. Kb2 Ke7 3. Kc2 Kf6 4. Kd2 Ke7 5. Ke2 Kf7 6. g3 fxg3 7. Qg2 Ke7 8. Kf1 Ke8 9. Qe2 Kd7 10. f4 * Here's an easier one that tests if programs will see the pawn break idea. [D]8/3k4/1p1p4/pPpPp1pp/2P1Pp1b/P4P1P/6P1/3Q1K2 w - bm g3
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.