Author: Peter Fendrich
Date: 10:33:36 09/22/04
Go up one level in this thread
On September 22, 2004 at 10:55:46, Anthony Cozzie wrote: >On September 22, 2004 at 08:40:00, martin fierz wrote: > >>On September 22, 2004 at 07:53:38, Drexel,Michael wrote: >> >>>On September 22, 2004 at 06:56:45, martin fierz wrote: >>> >>>>[snip] >>>> >>>>>>it's definitely not nonsense. i agree that the engine won't find a better (or >>>>>>the best according to the book), but a weak engine will make real blunders in >>>>>>the opening while a strong engine might just play a slightly inferior move. >>>>> >>>>>Strong engines don´t just play slightly inferior moves in the opening. >>>>>They play often complete nonsense without book. >>>>>I think we have discussed enough examples here in the past. >>>> >>>>of course - but you are guilty of selective perception. >>> >>>I´m not guilty of anything. >>>I was referring to: >>>"If the engine is strong enough to find better moves by itself then opening book >>>is not needed". >>>Engines generally won´t find better moves by itself than Top GMs in home >>>preparation. No matter how strong they get. >>>The opening book is needed as long as engines have no clue about long term >>>strategies. >>>Some simple development rules are not enough since there exist by far too many >>>exceptions. >>> >>>you have noted the >>>>examples discussed here. if you took 1000 opening positions from somewhere, and >>>>looked which moves GMs play, i believe that strong engines would play the same >>>>moves very often (ie. 80-95%). weak engines on the other hand... >>> >>>I also played 80-95% GM moves in the past. >>>Unfortunately the 5-20% "non-GM moves" made the difference. >> >>duh! your 5-20% are probably real blunders. top engines will not play serious >>blunders that often... >> >>cheers >> martin > >Kurt Utzinger did a test of some strong commercial engines without book a while >back. The results were pretty ugly if I remember correctly :( > >anthony I think that the question "no book vs (big) book" is not as interesting as a "tiny book vs big book". If you have a small wide book helping the engine to come out and start playing, combined with simple learning, it will probably work fairly well in long time controls. The book earns time and that's weighted higher in shorter time controls. /Peter
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.