Author: Sune Fischer
Date: 01:46:59 02/20/05
Go up one level in this thread
On February 20, 2005 at 01:28:46, Dan Honeycutt wrote: >On February 19, 2005 at 17:08:23, Uri Blass wrote: > >[snip] >> >>The point is that using them in the discussion about the value of book is >>irrelevant if the public book of shredder is not considered by you as a good >>book. >> >>Uri > >I was not talking the public book. Recall my statement "This all supposes the >authors are willing to let their tournament prepared books out, which I doubt" > >Perhaps I created confusion by saying "you could test" since you can't test if >the book is not public. Let me rephrase. To prove or disprove Vincent's >statement you would need to test using a program with a tuned book such as Diep >or Zappa using Arturo's book or Shredder using Sandro's book. > >What Arturo thinks about Shredder's book has nothing to do with what I'm talking >about. My sole point is that, IMO, Tord did not test with the type book Vincent >described when he made his statement regarding the value of the book. Let's consider the four main types of book 1. no book 2. wide random book 3. normal book 4. tournament book If Vincent says 1 vs 4 is 700 Elo and we assume the engine without book has no learning, then I must agree with him. All you need is to repeat the last line the engine lost, this is in fact more than 700 Elo and it doesn't even require much of a book, just two lines of code. The other possible combinations won't give 700, especially not with learning, some say 100 others say 200 but it's in that ball park. -S. >Best >Dan H.
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.