Author: Angrim
Date: 13:49:54 10/26/01
Go up one level in this thread
On October 26, 2001 at 13:02:55, Heiner Marxen wrote: >On October 24, 2001 at 19:27:04, Angrim wrote: > >>On October 23, 2001 at 13:12:38, Andrew Williams wrote: <snip> >>>[D]3r1rk1/2n1nppp/1pp1p3/p2qP2Q/P1RPN3/2R5/1P2NPPP/6K1 w - - bm Nf6 >>> >>> >>>Regards >>> >>>Andrew >> >>Interesting position to feed to the mate solver programs, since it >>comes from a real game, and is a rather deep win. Here are the >>results from PN^2 search on Athlon 1.2ghz. >> >>proved that move e4f6 wins, 17 turns >>PN2:215756548 evals, 11265632 expands, 407.84 seconds >> >>Angrim > >Hello Angrim, > >I'm not sure what "17 turns" does mean, but I'm sure it is not the solution >depth in plies. Even after 1.Nf6+ gxf6 Chest does not find a mate in 10 >(28.4 hours on a K7/600 with 350 MB hash). > >Heiner 17 turns means 34 or 35 ply. Seems it was far from the shortest win. I store the values internally as ply/2 to save 1 bit, so my result is often off by 1 ply. I was wondering how black could hold off the mate for so long, so I had it play the mate out, here is the main line. 1.Nf6+ gxf6 2.Rg3 Kh8 3.Qh6 Ng6 4.Rh3 Qxg2+ 5.Kxg2 Nf4+ 6.Nxf4 Rg8+ 7.Kf3 Rg7 8.exf6 Rd8g8 9.fxg7+ Rxg7 10.Nh5 Ne8 11.Nxg7 Kg8 12.Nxe8 <any> 13.Qg7# Interestingly, it had a score of mate in 15 ply stored after 11.Nxg7.. but actually playing it out it found a much faster one. Angrim
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.