Author: Dagh Nielsen
Date: 19:03:54 12/28/05
Go up one level in this thread
On December 28, 2005 at 21:13:25, Gustavo Bedoya wrote: >Hi all >I'm rather new here , and I know that Rybka is the >outstanding engine nowadays. >But Surprisingly I've read it lost a game with an (excuse me if I'm >wrong)inferior engine. I read that Rybka is not playing with full strngth >because if so, it would be so easy for Rybka. >I ask all the experts in this Topics if Rybka is really the strongest engine of >all time, an how can be so stronger than others, what is different in its source >code that makes it so lethal,( in my modest engine matches games, Rybka crushes >every engine one after one, excepting Fritz9 in the ChessBase interface +5=2-5). >And in the future it can be cloned and there will be stronger engines every >year, even every mounth? >I'll be happy with all your opinions >Best Regards Dear Gustavo Let me attempt some answers and comments: (1) Rybka lost to Spike. Spike may be an inferior engine, but it may very well also be very strong. It is a new version as far as I know, not public yet, so who knows how strong it really is? Only time and tests will tell, which leads me to... (2) One loss is hardly conclusive regarding the strength of an engine, neither Rybka or Spike. Upsets will allways occur, also to the current king of the hill ;-) (3) It may be that Rybka is playing "only" as the released beta version, but I don't think the reasoning behind it was anything like avoiding winning too easily!!! There may be many other reasons for not wanting to participate with a newer version that is still in its experimental stage. (4) Yes, Rybka is the strongest ever public engine, according to various ranking lists. I am afraid it is not simple to explain what makes it so strong, otherwise we probably would have seen Rybka-strength engines long ago... Many people say it is outstanding in middlegames. My own limited experience is that it handles advanced and/or passed pawns extremely well, rook intrusions on 7th rank also, handling of an active queen too, and many other things... Perhaps it has a very good and sophisticated positional evaluation function. Besides that, I have a personal theory that it has integrated a clever search feature that lets it build its search tree more effectively than other programs. (5) The source code is not public, so direct clones are not very likely, but maybe other program authors will take a sneaky look and get some inspiration :-) Probably there will be improvements every year for some time to come, why not? It's not like every idea and combination has been tried and optimized yet, I would think. Regards, Dagh Nielsen
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.