Author: Günther Simon
Date: 14:27:47 12/02/05
Go up one level in this thread
On December 02, 2005 at 17:17:50, Robert Hyatt wrote: >On December 02, 2005 at 15:49:11, Paul Jacobean Sacral wrote: > >>On December 02, 2005 at 15:19:51, chandler yergin wrote: >> >>>On December 02, 2005 at 14:56:48, Uri Blass wrote: >> >>>>Engines can say mate in 14 even when there is mate in 11 but not the opposite. >>>> >>>>Engines should say mate in 11 for white only when they checked all defences of >>>>black and found that there is no defence against mate in 11. >> >>>Engines evaluate all moves 1/2 move i.e. 1 Ply at a time. >>>Nothing more.. nothing less. >>>The deeper the search the better the evaluation. >>>You assert bugs when there is none. >> >>Excuse me, but with replies like that people will start to think that you are a >>troll. Please don't explain to Uri Blass how engines work: He has written an >>engine himself. >> >>Call it bug, mistake, error or whatever you like: If the shortest possible mate >>in a position is #14 and an engine displays #11, there is something wrong. Maybe >>a rare effect of nullmove pruning or something like that... but #11 is simply >>wrong and not a "better evaluation!" :) >> >>Yours truly Paul J. Sacral > > >That is completely correct. If an engine claims a shorter mate, and it is >proven the mate is longer than that, then the engine has a bug, pure and simple. > The inverse is not true, since many engines will find a deeper mate and quite >before finding the absolute shortest mate (some like Crafty won't do this except >when EGTBs are involved) and this is not necessarily a bug, but an artifact of >how the search is extended. > >I am not sure why this has become "disinformation central" all of a sudden... Bob, this is not 'all of a sudden'. It was always there in every post from CY since ever and in times he writes 20+ posts a day it simply looks like that over here... Guenther
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.