Author: José Carlos
Date: 05:49:58 09/02/02
Go up one level in this thread
On September 02, 2002 at 07:59:40, Sune Fischer wrote: >On September 02, 2002 at 06:25:00, José Carlos wrote: > >>On September 02, 2002 at 04:24:21, Sune Fischer wrote: >> >>>On September 01, 2002 at 19:12:01, Andrew Williams wrote: >>> >>>>Hi, >>>> >>>>On Friday I got back from a business trip and saw the discussion going on about >>>>the effect of the opening book. Too tired to think of anything else to test, I >>>>started a match between postmodernist with its big book (made up of Dann >>>>Corbit's games collection) versus postmodernist without a book. To encourage a >>>>bit of randomness, I told both versions to randomly vary the time they used for >>>>searching the first few moves out of book. I have no position at all on the >>>>debate that was going on, but here is some more data anyway: >>>> >>>>I played 200 games at 5 minutes + 2 seconds increment. Both sides played with no >>>>pondering on the same Athlon 1200. There were no duplicates amongst the games >>>>(according to Scid; I haven't checked by hand). >>>> >>>>The final score from the point of view of the version with the book was: >>>> >>>> +71 =78 -51 >>>> >>>>With a final score of Withbook 110 - 90 Nobook. >>> >>>As you and others have demostrated the (good) books have a positive effect on >>>the score, but is it because the books leaves them in a superior position, or >>>simply because of the time they save playing directly from book? >>> >>>I'd expect you might also see a score close to 110-90 if you give one program 5 >>>min 2 sec inc and the other only 4 min 2 sec inc. It certain is a factor to be >>>considered... >>> >>>-S. >>> >>>>Andrew >>>> >>>>PS Irrespective of the significance of the figures given here or elsewhere, I >>>>won't participate in any tournament where an opening book is used without >>>>explicit permission from its creator to use it in the tournament. >>> >>>PS >>>No matter the effect of the book, it isn't the _engine_ playing.... >> >> According to _your_ definition of engine. > >From what I've heard there are GUIs with protocols that handle all the book >moves and never even asks the engine for 'advice'. >In this case the distinction is clear, it is not the _engine_ playing. > >It is the combination of hardware, engine and book that makes the package >strong. > >-S. > >> José C. There're GUIs that resign, there're some that check legality of moves. Some even offer draw! Oh, and some of them control the EGTBs. And some clear the hash tables, or the learning tables, or... In summary, you can make any definition you want. In my case, my engine does all of this itself. I define a GUI as a "graphical user interface". No more, no less. An interface just and only takes care of communication. So my book code is part of my engine. Your definition of engine is "everything but what a GUI can do". Well, then an engine is nothing, because a GUI can search as well, why not. José C.
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.