Author: William Bryant
Date: 08:07:19 10/25/99
Go up one level in this thread
On October 24, 1999 at 23:45:47, James Robertson wrote: >On October 24, 1999 at 22:54:00, William Bryant wrote: > >>On October 24, 1999 at 12:55:22, James Robertson wrote: >> >>>r4q1k/p2bR1rp/2p2Q1N/5p2/5p2/2P5/PP3PPP/R5K1 w - - bm Rf7; id WAC008; >>> >>>My program has always solved WAC 8 easily. But last night I noticed the latest >>>version chose Nf7+ instead of Rf7. I looked at the position with Crafty 16.18, >>>and it gives mate in 7 after Rf7. After Nf7+, Crafty gives mate in 10 on ply 12, >>>so this value may not be trustworthy. >>> >>>Can anyone look at the position and find out how far Nf7+ really is from mate? >>> >>>James >> >> >>Jumping on this thread late, >> >>James, mate positions in the hash table can be tricky. My first problems were >>with the program never choosing the shortest path to mate. Bruce is an advocate >>of only storing bounds -not- distance to mate in the hash table. Bob, just the >>opposite. >> >>If you were solving this before your latest 'change'. See if the change affects >>your hash table and therefore starts leading you for a longer mate. >> >>William >>wbryant@ix.netcom.com > >I generally test WAC with 5 seconds per position. On my P233, nobody finds this >mate in that amount of time. :) > >James 5s a positions. That can make it quite a test. I just tested the latest Screamer at 1 minute per position. I am getting 264 positions in less than 10s (I will check to see how log I took to see position number 8 -- but it was less than 10 seconds). -284 positions is less than 30s -287 positions in less than 60s Thats a majority in less than a minute -- much better than before. Now I'm curious about how long I took to find #8, the "No One" gauntlett has been tossed :))))) When I get home I'll look this up and let you know. William wbryant@ix.netcom.com
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.