Author: Robert Hyatt
Date: 07:07:10 10/29/03
Go up one level in this thread
On October 28, 2003 at 14:12:43, Uri Blass wrote: >On October 28, 2003 at 13:29:50, Robert Hyatt wrote: > >>On October 28, 2003 at 12:18:45, Sune Fischer wrote: >> >>>On October 28, 2003 at 11:57:02, Robert Hyatt wrote: >>> >>>>On October 28, 2003 at 11:27:30, Sune Fischer wrote: >>>> >>>>>On October 28, 2003 at 10:55:53, Robert Hyatt wrote: >>>>>> >>>>>>I don't think it has a thing to do with honesty. I'd never question >>>>>>Jeroen's honesty at all. >>>>>> >>>>>>It does have a lot to do with fairness. Bruce Moreland summed it up >>>>>>best: "why do I have to face the _same_ outstanding book twice in the >>>>>>same tournament when I don't ever face the same _program_ twice?" >>>>>> >>>>>>That's a good point. A good book can be a significant advantage. There >>>>>>are complaints if an amateur tries to use a commercial program's opening >>>>>>book. Why not if two different commercial entries try to use the same >>>>>>book? >>>>> >>>>>It's a grey zone, just like with Eugene's endgame tables. >>>>> >>>> >>>>Not quite the same. Everybody uses them. The info is identical for all >>>>users. Commercial books are not used by everybody... >>> >>>But what is fairness? >>> >>>Some programs run on dog slow machines and others on top end multiprocessor >>>systems, some "have an arrangement" with a book author to use his books, some >>>get paid to develop whilst others do it in their spare time. >> >>That is just a part of the deal. But if you go back to early WMCCC events, >>commercial vendors were allowed to enter _four_ copies of the same program. >>And play in a _swiss_ event. you talk about foul-play? look at some of >>the old WMCCC reports about thrown games, etc. :) >> >>I am willing to compete with a commercial programmer. I don't have a problem >>with that and never have. What I don't particularly like is to have one person >>working on three different teams. And yes, Virginia, the book is a major part >>of a chess engine and it is important to have a good one. Letting one person >>work on the book for three engines is no different than one person working >>on three different engines directly. It isn't allowed by the rules... > >I think that the main problem with a book for more than one program is that the >the book maker may help one of the program to beat another program. > >I do not suspect Jeroen for doing it but this possibility is a reason that is >good enough not to allow it. > >> >> >> >> >>> >>>Who's to say what is fair and what is not? >>>More importantly, is the question really interesting? >> >>The book part is. IE suppose _everybody_ in an event gets to use a high- >>quality book, except for _you_. You get thrashed game after game, before >>you even leave the opening book. Does that show anything interesting? Do >>you like the result? > >I think that this is overestimating the value of the book. >Better book is an advantage but if you know that the opponent has a better book >you practically still can do things to avoid losing by some killer line. > >I guess that the better book of the commercial programs gave them 40 elo >relative to Ruffian's book and these 40 elo were not important enough to prevent >Ruffian to win the tournament. > >Uri You have to think outside your statistical box here. Your 40 elo obviously has to apply to the effect on _many_ games. But in a tournament, that is not so important. A good book can produce a "1" score in a critical game, if you know who you are playing, what he likes to play, and you have time to prepare a particular line against him. I saw this in _many_ of the old ACM CC events. I did it myself multiple times. So even though such book preparation only wins one game out of every 10, that could be the key game that puts you in 1st position after the final round.
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.