Author: Uri Blass
Date: 15:34:05 02/19/05
Go up one level in this thread
On February 19, 2005 at 17:49:29, Arturo Ochoa wrote: >On February 19, 2005 at 17:08:23, Uri Blass wrote: > >>On February 19, 2005 at 16:33:48, Arturo Ochoa wrote: >> >>>On February 18, 2005 at 16:19:26, Uri Blass wrote: >>> >>>>On February 18, 2005 at 13:55:25, Dan Honeycutt wrote: >>>> >>>>>On February 17, 2005 at 17:48:18, Uri Blass wrote: >>>>> >>>>>>On February 17, 2005 at 14:42:41, Dan Honeycutt wrote: >>>>>> >>>>>>>On February 17, 2005 at 14:03:30, Tord Romstad wrote: >>>>>>> >>>>>>>>A couple of days ago, a well-known programmer and regular >>>>>>>>poster here on the CCC claimed that a good opening book >>>>>>>>was worth at least 700 Elo points. I thought this number >>>>>>>>looked completely outrageous, and decided to do a simple >>>>>>>>experiment. >>>>>>>> >>>>>>>>I am the author of a basic and minimalistic UCI chess engine >>>>>>>>called Glaurung. Source code and executables for Mac OS X, >>>>>>>>Linux and Windows can be found at the following URL: >>>>>>>> >>>>>>>>http://www.math.uio.no/~romstad/glaurung/glaurung.html >>>>>>>> >>>>>>>>Recently, I have played some test matches with Glaurung >>>>>>>>against the strongest engine I have on my compter: Hiarcs >>>>>>>>9.6. Not surprisingly, all such matches end in crushing >>>>>>>>victories for Hiarcs. The last match I played ended >>>>>>>>75-25 in Hiarcs' favor. >>>>>>>> >>>>>>>>As a crude test of the "good book=700 Elo" claim, I have >>>>>>>>now repeated the match with identical program versions >>>>>>>>and conditions, except that Hiarcs was now playing without >>>>>>>>an opening book. Assuming that Hiarcs' book is worth 700 >>>>>>>>Elo, the expected result of this second match would be >>>>>>>>something like 95-5 in _Glaurung's_ favor. >>>>>>>> >>>>>>>>The actual result of the second match was very close to >>>>>>>>the first match: Hiarcs won by 72-28. >>>>>>>> >>>>>>>>As far as I can see, this means that at least one of the >>>>>>>>following must be true: >>>>>>>> >>>>>>>>a) The statement "good book=700 Elo" is lightyears away >>>>>>>>from the truth. >>>>>>>> >>>>>>>>b) Hiarcs has an extremely bad opening book, and with a >>>>>>>>half decent opening book it would be several hundred >>>>>>>>rating points ahead of Shredder. >>>>>>>> >>>>>>>> >>>>>>>>Tord >>>>>>> >>>>>>>Hi Tord: >>>>>>> >>>>>>>I too think 700 is a number from the twilight zone. But the statement, IIRC, >>>>>>>was 700 points for a very good book vs no book. So you have to include the >>>>>>>possibility: >>>>>>> >>>>>>>c) Glaurung's book is no good. >>>>>> >>>>>>Glaurung's book is not the subject here. >>>>>> >>>>>>Tord simply comapred hiarcs book with no book. >>>>>> >>>>>>He let Glaurung with it's own book to play agaisnt Hiarcs with it's book and let >>>>>>Glaurung with it's own book play against Hiarcs with no book. >>>>>> >>>>>>Hiarcs with it's own book failed to perform even 100 elo better. >>>>>> >>>>>>Uri. >>>>> >>>>>All right then add the possibility: >>>>> >>>>>d) Hiarcs' book is no good. >>>> >>>>Tord already considered that possibility >>>> >>>>one of the possibilities that tord considered: >>>> >>>>"b) Hiarcs has an extremely bad opening book, and with a >>>>half decent opening book it would be several hundred >>>>rating points ahead of Shredder." >>>> >>>> >>>>> >>>>>However you take it, to prove or disprove the statement you are going to have to >>>>>test with a "good book". We know Arturo builds good books - you could test with >>>>>Diep or Zappa. >>>> >>>>Arturo did not make his books public so I cannot test them. >>>> >>> >>>I dont have reasons to make public my books. >> >>No problem with it. >>I only responded to "you could test with Diep or Zappa." >> > >I run my own private tests and I know what the advantages of a tuned book are. >Of course, I would be crazy or I would be very idiot to give such a book to bad >testers and/or unrelieable people. And the authors that I help would not be >happy with such idiot actions. > >> >> >> Neither Vincent Diepeveen nor >>>Anthony Cozzie would be happy about that :) Besides, why to make public a book >>>that is useless, boring and not required for the majority of the public >>>(specially persons as you? >> >>I did not ask to make it public. >> > >When you said that I had not made public any of my Tournament Books, you were >expecting that something like that might happen. Of course, it wont happen for >the know reasons...... No I did not expect it to happen. I only responded to the follwoing sentence of Dan "We know Arturo builds good books - you could test with Diep or Zappa." I only explained why I could not test it. I did not compain about the fact that I am unable to test it. > >>> >>>Ask Vincent Diepeveen or Anthony Cozzie if they are really willing to give such >>>a combination of software: their own engines plus my tournaments books. >>> >>> >>>> >>>>> We know Sandro builds good books - you could test with Shredder. >>>> >>>>I do not know that the book that you can give to shredder is defined as a good >>>>book by Arturu. >>>> >>> >>>My name is Arturo not Arturu. My books are tuned foor specific engines. >> >>sorry for the mistake in your name. >> >>Again I only responded to the poster and the subject was not your book but the >>question if sandro's book are good book. > >Well, I would doubt that Mr. Sandro´s books are bad. I would not defend this but >this question is quite..... > >You will know how the combination of a strong engine like Shredder plus a tuned >book for it is very successful. However, you cotinue missing and ignoring the >point. The time has showed who is wrong..... > > >> >>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. > >I am not talking about a specific case: You generalized about how useless are a >book is. Therefore, I have showed why this position is absurd and wrong. Note that I never claimed that a good book cannot help an engine to win a tournament. If people understood it from me then I did a bad explaining job. I will try to do better explaining job in this post. I said that I consider book as unimportant and I said that an engine that is strong enough has good chances to win even with 1.h3 but I never said that the engines that play in world championship are strong enough relative to their opponents The difference between the best and the second best is for years less than 100 elo and when there is a small difference of course a book can be the decisive factor about the question who is going to win. The point is that I believe in the following assumptions: 1)It is possible to make an engine that is significantly stronger than shredder9(improvement of more than 200 elo in nunn type match and I am not talking only about nunn position but also about other equal positions out of book). I believe that it is only a question of time and I expect it to happen in less than 10 years from today. 2)A programmer who is good enough to develop an engine that is significantly stronger than the opponents has good chances to win WCCC even without working about book (when the opponents of course have not his new engine that made a big improvement of more than 200 elo). Uri
This page took 0.01 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.