Author: Eelco de Groot
Date: 11:36:48 10/11/03
Go up one level in this thread
On October 11, 2003 at 14:07:00, Uri Blass wrote: >On October 11, 2003 at 13:39:13, Ed Schröder wrote: > >>On October 11, 2003 at 10:40:16, Nolan Denson wrote: >> >>>The below game was played between Rebel 12 and Crafty 1 cpu. >>> >>>Rebel announce a mate !! after 39. .... KG6 ????@@@@#####%%% >>> >>> >>>DEUTERONOMY (2734) vs. MrPeabody (2735) --- 2003.10.11 10:12:32 >>>Rated Blitz match, initial time: 3 minutes, increment: 3 seconds >> >>Nolan, >> >>You must be playing a special style, the default setting can recognize Bh5+ as >>the best move (34 secs) but no mate seen after 5 min (A/2000). >> >>Nevertheless a nice mate combination. >> >>My best, >> >>Ed > >Or maybe when he says "announce a mate after 39...Kg6" he does not mean in the >immediate move after it. > >I see that rebel used 13 seconds for Qf2 and I wonder if Rebel continue to >search even after finding mates only to find a shorter mate. > >Uri It was a bugreport. When the score goes up or down a certain treshold something seems to be going wrong in a lot of cases and Rebel produces a bad move, like in this case 39...Kg6 In ERT the move would probably show up with a 0,00 score. A mate anouncement is optional. You never know, if it isn't just a failure in the Winboard protocol implementation in Rebel but rather something in the engine this might be something that acted up in old Rebel autoplayer games, so I hope Ed and Lex can find what is causing it. On the other hand, it wasn't present in the Prerelease version so that theory isn't very likely. Eelco Eelco
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.