Author: Robert Hyatt
Date: 08:19:43 09/02/04
Go up one level in this thread
On September 01, 2004 at 20:57:20, Jan K. wrote: >> >> 9 4.35 -1.01 1. Kf1 a5 2. Rxh5 gxh5 3. Rxh5 Re1+ >> 4. Kxe1 Nxh5 5. Qg5+ Kf8 6. Qh6+ Ke8 >> 7. Qh8+ Bf8 8. Qxh5 >> 9 4.82 +1 1. Qxf4!! >> 9 5.07 +3 1. Qxf4!! >> 9 5.50 +M 1. Qxf4!! >> 9 52.17 Mat06 1. Qxf4 Bxf4 2. Rxh5 gxh5 3. Rxh5 Bh6 >> 4. Rxh6 Qh2+ 5. Kxh2 Kf8 6. Rh8# >> time=52.17 cpu=98% mat=-1 n=85106981 fh=98% nps=1.63M >> > >hi, so it's alright(...) that it takes so long to find the mate? I thought that >this is some bug of my search that i can finish depth 9 with winning score in 3 >seconds and then depth 10 with mate score in 45 seconds. Although I extend only >1/2 ply for threat extension, here along with check extension it seems to blow >up the search. :) This is notmal. You do a search with a very narrow window and alpha/beta prunes away most deep mate branches. But once you widen the search window to include mates, bam. This also happens in endgames where the hash table allows you to search very deeply, pruning away variations where a pawn queens (the new queen can't be forcibly be created until say ply=40, but once it is forced, all the promotions at ply=10 and beyond suddenly can't be pruned based on score, and the tree explodes. It's an artifact of alpha/beta...
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.