Author: Uri Blass
Date: 16:32:38 07/23/01
Go up one level in this thread
On July 23, 2001 at 11:25:17, Heiner Marxen wrote: >On July 23, 2001 at 10:10:41, Uri Blass wrote: > >>On July 23, 2001 at 09:50:27, Heiner Marxen wrote: >> >>>On July 23, 2001 at 02:25:50, Uri Blass wrote: >>> >>>>On July 22, 2001 at 22:44:46, Heiner Marxen wrote: >>>> >>>>>On July 21, 2001 at 15:02:43, Uri Blass wrote: >>>>> >>>>>4k3/8/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - acn 718366; acs 8; bm e4; ce 32756; dm 6; >>>>>pv e4 Kf8 Qg4 Ke8 Qg7 Kd8 Bb5 Kc8 Bc6 Kd8 Qd7#; >>>> >>>>Here chest is faster than chessmaster in finding the shortest mate. >>> >>>:-) >>> >>>>chessmaster6000 cannot see mate in 6 even after some minutes because of null >>>>move pruning unless I asked it to search for mate in 6. >>>> >>>>The problematic line from chessmaster's point of view is: >>>>1.e4 Kd7 2.Qg4+ Kc7 3.Qg7+ Kc6 4.d4 Kb6 5.Qd7 when black is in zunzwang. >>>> >>>>Usually chessmaster is faster than chest in finding the shortest mate but it >>>>does not know that the mate that it found is the shortest mate. >>> >>>Yes, and this is what makes Chest slow, most of the time: proving that all >>>those shorter mates do not exist. >>> >>>>I guess that the times of chest in every problem is for finding all the >>>>solutions and in this case chessmaster has a clear advantage because it is happy >>>>with finding 1 solution. >>> >>>Normally (for Chest that is with composed chess problems with just one >>>solution) this is not a big thing. The average solution time would reduced >>>by a factor of 2, or even less, since the non-solution moves often are >>>significantly faster to compute. >>> >>>Here, where we have 2 or sometimes even 5 solution moves, it could be a >>>factor of 4. That would change the times significantly. >>>OTOH, it is nice to know all solutions as a reference point. >> >>Yes >>Chessmaster6000 had no error in the first move >> >>> >>>>Chessmaster results in finding the shortest mate in the same problems in >>>>comparison to chest: >>>> >>>>-, 8 >>>>4, 190 >>>>2, 180 >>>>2, 129 >>>>13, 141 >>>>4, 157 >>>>17, 128 >>>>6, 182 >>>>2, 182 >>>>52, 250 >>>>25, 203 >>>>108,278 >>>>125,178 >>>>- ,186 >>>>24 ,216 >>>>6, 291 >>>>38, 197 >>>>87, 179 >>>>5, 230 >>>>25, 213 >>>>19, 200 >>>>103,357 >>>>71 ,324 >>> >>>Thanks for the info! >>>That doesn't look bad for Chest. What CPU/memory did you use for Chessmaster? >> >>PIII850(128 Mbytes hash tables) > >Thanks! So the times are (very) roughly comparable to my K7/600. > > >>>>Chessmaster also found mate in 8 in all the other problems that were posted >>>>except one >>>> >>>>[D]4k3/pp2p3/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1 >>> >>>That one is still busy computing :-) >> >>Chessmaster's results >>e3 13 seconds: mate in 13 >>e3 17 seconds: mate in 11 >>e3 61 seconds: mate in 9 >> >>Regards,Uri > >Chest just completed the above position: 6851.09 sec (114.2 min) (1.9 hrs) >It is a mate in 8 with 3 solution moves: e3, e4, d4: > >PV: e3 e6 Qh5+ Kd8 Nc3 b6 Ne4 e5 Qh7 b5 Nc5 Ke8 Bxb5+ Kf8 Ne6# >PV: e4 e6 Qh5+ Kd8 h4 e5 Nc3 b6 Nd5 b5 Bxb5 Kc8 Bc6 a6 Qe8# >PV: d4 e6 g3 b6 e4 e5 Bb5+ Kf8 Qf3+ Kg8 Bc4+ Kh8 Qf6+ Kh7 Qh6# > >Regards, >Heiner 1)I checked again one of the problems when chessmaster could not see mate in 7 in a few minutes(chest needed only 186 seconds to find it) 4k3/p4p2/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1 Chessmaster could not see mate in 7 in the first 15 minutes 2)In the position when chest found 3 solutions Chessmaster6000 needed 833 seconds to see that 1.e4 is mate in 8 [D]4k3/pp2p3/8/8/8/8/PPPPPPPP/RNBQKBNR w KQ - 0 1 Uri
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.