Author: allan johnson
Date: 00:15:25 03/19/01
Go up one level in this thread
On March 17, 2001 at 22:55:14, Robert Hyatt wrote: >On March 17, 2001 at 14:48:05, Chris Hull wrote: > >>On March 17, 2001 at 13:38:31, Uri Blass wrote: >> >>>On March 17, 2001 at 13:04:29, Antonio Senatore wrote: >>> >>>>It seems incredible that a top chess program like Junior 6 (without the KPPKB >>>>tablebase but with the KBK and the KBKP tables) CAN'T SEE MATE IN ONE in the >>>>next position: >>>> >>>>[D]7k/4pK1p/7B/8/8/8/8/8 w - - >>> >>>It is not a bug >>>It is a design decision by the programmer. >> >> >>A bad design decision is even worse than a bug. A bad design is >>almost always much harder to correct than a bug in the program. > >If you count the number of games this "bug" loses, vs the number of >games it wins by letting the program avoid trading into an ending that >is drawn 99.99999999% of the time, you might decide that this isn't a >"bug" at all. It is a very reasonable way to handle a well-known problem. > > >> >> >>>Junior knows that king bishop with not more material cannot mate and the same >>>for king knight with no more material. >>> >>>It practically help Junior in games because Junior simply does not have to >>>search in KB vs KPP positions unless it evaluates the position as advantage for >>>the side with the pawns. >>> >>>I believe that you can change the default parameters if you prefer a weaker >>>program that can see the mates in the position that you posted. >>> >>>Experience tells that they never or almost never happen in games so the >>>programmer(Amir Ban) did not care about them. >>> >>>Uri >> >>Chris
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.