Author: Vincent Lejeune
Date: 07:25:05 12/21/99
Go up one level in this thread
this is good example of human+computer cooperation in understanding a position : instead of giving dozens playing lines the human describe the strong moves (Rd3!: stalemate and Rc8 follow by a fork) On December 19, 1999 at 15:20:27, Richard A. Fowell (fowell@netcom.com) wrote: >On December 18, 1999 at 22:16:31, Vincent Lejeune wrote: > >> >>7b/8/kq6/8/8/1N2R3/K2P4/8 w - - 0 1 >> >>White play and draw ! >> >>Interresting to test Stalemate/zugzwang and repetition detection of programs >> >>(all my surprise effect has desappear now :'( ) > >Well, that's a tougher one that the first one, since it is one ply deeper. >I only tried this one on Chessmaster 6000 Mac on my 180 MHz 604e, since it >scored the best on the easier problem (tied for first at 15 seconds to score >the latter position "0"). > >Chessmaster 6000 Mac got your "good position": >7b/8/kq6/8/8/1N2R3/K2P4/8 w - - > in a hair over 70 minutes: > >*** Chessmaster 6000 Mac - 180 Mhz 604e **** >Time Depth Score Moves >-------------------------------------------------------------- >00:36:56 10 -3.63 d2-d4 a6-b5 e3-h3 h8-f6 a2-b2 b5-b4 h3-f3 >00:46:49 11 -3.66 d2-d4 a6-b5 e3-h3 h8-f6 a2-b2 b5-b4 h3-d3 >01:10:13 11 0.00 b3-d4 h8-d4 e3-a3 a6-b5 a3-b3 b5-c4 b3-b6 >01:12:15 12 0.00 b3-d4 h8-d4 e3-a3 a6-b5 a3-b3 b5-c4 b3-b6 > >Here's the ranking on your earlier problem: >7b/8/kq6/8/3N4/4R3/K2P4/8 b - - > >Chessmaster 6000 - score of "0" at 15 seconds (7 ply) >Chessmaster 4000 - score of "0" at 15 seconds (7 ply?) >HIARCS 7.0 - score of "0" at 190 seconds (11 ply) >MacChess 5.01 - score of 202 at 2 min 44 sec (13 ply) down from 680) >Screamer B42 - score of -0.6 at 22 min 13 seconds (13 ply) > (down from -8.22 ) > >*********************************************************** >**** Warning - solution follows *************************** >**** (in case you want to find it yourself ) ************** >*********************************************************** > > >- 5 - > > > >- 4 - > > > >- 3 - > > > >- 2 - > > >- 1 - > > >- Solution - > >The reason that Nb3-d4! draws is that it forces the >exchange of the Black Queen for the White Rook. >The move has these features: > >1) It unblocks White's rook on the 3rd rank, creating the > threat of Ra3+ followed by Rb3, forcing the trade of the > Queen for the Rook by skewer or pin. > >2) It guards the e6 square, creating the threat of Re6 pinning > the Queen to the king and forcing the exchange of the Queen > to the King. > >Most of Black's moves (e.g., any king move, any bishop move, or any >move of the Queen on the 6th rank or to the a,b,c files that does not >put the Queen en prise) are met by one of the above motifs - >Rook to a3, b3, or e6. > >The most obvious Black response - Qxd4, results in either draw by perpetual >check, as White shuttles the rook amongst a3, b3, c3, or by the stalemate >sacrifice of Rd3!! if Black moves his King to the d file. Since Rd3 then >pins Queen to King, and the reponse Qxd3 stalemates White, this scenario >is covered. > >The remaining try for Black is Qd8. The idea is to try to sneak the King >away from the barrage of Rook checks as the Rook shuttles between >a3 and b3 without allowing the pin/skewer of the Queen. >This try fails due to the power of the Knight and its forks >for most attempted exits. One can draw an iron wall of squares that the >King dare not step on: > >a4 - since the rook will always check at a3 whenever the King moves > to a5, the only entry to that square, since: >b5 - is covered by the Knight, and moving from the King from a5 to >b4 - is met by Nc6+ forking the Queen, and White actually wins >c5 - is met by Ne6+ forking the Queen and winning for White >c6 - is covered by the Knight >c7 - is met by Ne6+ forking the Queen and Black will be hard pressed to draw > >This only leaves c8 as a potential exit. But ... > >Kc8 - is met by Rb8+! Kxb8, Nc6+ forking the Queen, and when the board > clears, it is a draw, as Black has only a Bishop (insufficient mating > material). > >In conclusion, the two most interesting lines for: >7b/8/kq6/8/8/1N2R3/K2P4/8 w - - > >are: > >1. Nd4 Qxd4 2. Ra3+ Kb5 3. Rb3+ Kc4 4. Rc3+ Kd5 5. Rd3 Qxd3 * >1. Nd4 Qd8 2. Ra3+ Kb7 3. Rb3+ Kc8 4. Rb8+ Kxb8 5. Nc6+ Kc8 6. Nxd8 Kxd8 *
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.