Author: Dann Corbit
Date: 15:41:15 08/16/00
Go up one level in this thread
On August 16, 2000 at 18:37:12, leonid wrote: >On August 16, 2000 at 16:57:08, Jari Huikari wrote: > >>[D]3R1B1K/5p2/2p2p2/b1rPNP1n/q1rkNP2/1Pp2Q2/1n6/5B2 W - - >>[D]1B1N4/p2n4/3P1P2/2nk1BP1/P7/N7/K3P2P/1Q6 W - - >>[D]kbb2rr1/1p6/p3p1q1/PN1pB3/1n6/R3Q3/5PB1/6K1 W - - >> >>These were quite hard positions for Nero. >>It needed 143 s. for the first one, 262 for the second and the third >>wasn't yet solved in 12 minutes. >> >> Jari > >Thanks for good positions in good graphics! Much easier to see and solve. > >You forgot to say what is your computer and your program. > >Your first position is mate in 4 moves. Can be solved by selective search 4 >moves ahead but not by simple one. By selective search (my computer is AMD 400. >Mate solver in LLchess) position was solved in slightly less that one second. By >brute force search it took 4 seconds. > >Second position. Mate in 3 moves. Could be solved by selective search as well. >Can't say exactly the time needed to solve it. Too tiny. Both ways (selective >and brute force) take less that 0.055 of one second. This is very interesting, because I have only seen mate in 4. Can you show the correct pv for mate in 3? >Third position. Mate in 8 moves. Can be solved by selective and brute force >search. Both ways it take less that 0.055 of one second to solve. > >To give you practical idea how long it will take in real program to solve all >those position, will give you numbers from my program. My program have mate >solver that work before every new move. It solved first two positions in less >that 0.055 and missed the last one. It missed the third one because it take only >as far as 7 moves ahead. > >You could be surprised that default mate solver took first position in less that >0.055 sec. Reason for this is that it looked more moves ahead that this position >demand as minimum. It solved it in 7 moves ahead. > >Leonid.
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.