Author: John Merlino
Date: 15:05:31 11/04/03
Go up one level in this thread
On November 04, 2003 at 14:39:50, Dieter Buerssner wrote: >On November 04, 2003 at 11:06:27, leonid wrote: > >>Try the next position that have good branching factor and, probably, will be >>easy to solve by selective, or by brute force search. > >Hmmm. I used my chess engine (Yace), which in your terminoligy, you would >probably call selective. It still should be reliable (never announce a mate in >x, when there is no mate in x but perhaps only a mate in x+1/2/... Of course it >might announce mate in x, when mate in x-1/2/... is possible). I did neither get >a good branching factor, nor an easy solution .... > >> 1qkrr3/Rn2q3/PRPBqQ2/nqpbqB2/NQ2Nb2/1Q2Qq2/1Q2Qq2/1K2Qq2 w - - > >... actually got no solution at all in 30 minutes on my P4 2.53 GHz with 100 MB >hash and using about 50% of the CPU (another analysis, that I did not want to >stop is running at the same time with more hash). > >Perhaps, it might be easy by selective search, when extending more aggressively >(many checks in a row are possible). BTW. It took 17 seconds to finish the depth >1 search. I guess many engines will evene need longer (but did not try any). >Branching factor from depth 1 to 2 was indeed good (17 s, then 21 s). ply 3: >1:02, 4: 14:20. Hardly one could consider it good ... > >I don't complain at all, but personally (and probably I am not the only one >here), I find more "natural" chess positions for mate problems more interesting. >Of course, it was my choice, to test your position anyway (I was curios). >Especially, I would find those problems most interesting, where a mate comes >rather surprisingly after a rather balanced score. Seeing several iterations in >a row with rather high score from almost the beginning seems less interesting >for normal playing engines. Sure, when we have 12 Qs on the board already, a >rather high score may not mean much. > >You are probably aware, that the position cannot be reached in a chess games. > >My favorite mate problem (already discussed here several times) is: > >6r1/2rp1kpp/2qQp3/p3Pp1P/1pP2P2/1P2KP2/P5R1/6R1 w - - > >Regards, >Dieter Chessmaster 9000 had a very difficult time with this one, and with a 10-minute search using both selective and brute force (on a P4-2.4) was unable to announce a mate. I was able to show that there is no mate in 9, but that's it. Also, MY favorite mate problem is this position (also mentioned many times before): [D]4r1k1/qp1r1p2/2pb1Bp1/p6p/2PP1n1R/1P3P2/P4P2/2Q2K1R w Qxf4! is a Mate in 10, while the supposed solution of Rxh5 is "only" a Mate in 11. jm
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.