Author: Dave Gomboc
Date: 20:32:19 09/28/01
Go up one level in this thread
On September 28, 2001 at 23:02:12, Christophe Theron wrote: >On September 28, 2001 at 11:13:40, Dave Gomboc wrote: > >>On September 27, 2001 at 20:01:28, Uri Blass wrote: >> >>>On September 27, 2001 at 18:00:39, Jorge wrote: >>> >>>>I was using Depth PLY 10 setting for ChessTiger14 when playing against CM8000. >>>>At some point in the game Tiger finds Mate in 8 moves, however every time >>>>Tiger's turn to move from then on, it took longer and longer to respond. When it >>>>came within Mate in 3 moves, it took over 40 minutes before I Forced the rest of >>>>the moves for Tiger. >>>> >>>>I am no expert in PLY search, but this does not appear to be normal (at least >>>>Fritz and Hiarcs do not behave this way). Has this happened to anyone else? >>>> >>>>jorge >>> >>>yes >>>it happens to me also in few of my ply games between Deep Fritz and Tiger14(not >>>more than 40 minutes for mate in 3 but certainly more than 3 minutes for mate >>>in2 at depth 9). >>> >>>Fortunately Deep Fritz often resigns when tiger gets a winning position so it >>>does not have to suffer for a long time. >>> >>>Tiger also does not behave in this way in most of the cases but there are cases >>>when it behaves in this way. >>> >>>I did not waste a lot of time about ply matches in the last monthes. >>>I started the nunn2 match at 13 plies for every program and the result between >>>Tiger14 and Deep Fritz is >>>8-6 for tiger14. >>> >>>Note that at this level games on PIII850 usually take few hours. >>>Thr longest game was adjudicated after more than 10 hours as a win for tiger14 >>>(tiger saw a mate in 11 and Deep Fritz saw 12.59 pawns against itself after >>>Tiger's move) >>> >>>Times in that game were tiger14 5:54:25 Deep Fritz 4:42:56 >>> >>>The biggest time that tiger calculated for one move in that game was 1:53:02 >>>The biggest time that Fritz calculated for one move in that game was 32:14 >>> >>>Uri >> >>Perhaps Tiger has a mate score instability bug? >> >>Dave > > > >No it's just that Tiger does not stop thinking in ply depth mode when it has >found a mate which is shorter than the current ply depth. > >Imagine you have selected fixed ply search, 10 plies. Even if Tiger has found a >mate in 2 (3 plies), it will keep on looking at the other moves for the 10 >plies, which is obviously a total waste of time. > >I do not think I'll fix this, because it's not what fixed ply is supposed to do >anyway. > > > Christophe Fair enough! I didn't realize they were not using it in "normal" mode. Reading too fast... Dave
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.