Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: The importance of opening books -- a simple experiment

Author: Arturo Ochoa

Date: 16:33:01 02/20/05

Go up one level in this thread


On February 20, 2005 at 12:38:01, Uri Blass wrote:

>On February 20, 2005 at 11:07:06, Arturo Ochoa wrote:
>
>>On February 20, 2005 at 10:40:25, Uri Blass wrote:
>>
>>>On February 19, 2005 at 20:38:22, Arturo Ochoa wrote:
>>>
>>>>On February 19, 2005 at 19:32:33, Uri Blass wrote:
>>>>
>>>>>On February 19, 2005 at 18:46:53, Arturo Ochoa wrote:
>>>>>
>>>>>>>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.
>>>>>>>
>>>>>>
>>>>>>??????????? !!!!!!!!!!!!! Go to (*)(**)
>>>>>>
>>>>>>>
>>>>>>>
>>>>>>>I said that I consider book as unimportant and I said that an engine that is
>>>>>>
>>>>>>(*) unimportant = not meaning much, not having value or significance
>>>>>>
>>>>>
>>>>>I agree with the definition but the question what is unimportant is also a
>>>>>question of opinion
>>>>>
>>>>>reasons for me to consider book as unimportasnt are the following:
>>>>>1)not considering the target of winning tournament as an important target
>>>>>2)thinking that it is possible to improve engine instead of book and get better
>>>>>results
>>>>>If shredder9 with book is weaker than some future Shredder19 without book then
>>>>>it is going to show point 2.
>>>>>
>>>>
>>>>(1) Impossible that it happens simply. Shredder is always acompanied by its
>>>>tuned and tested book in official Tournaments. In the particular case of
>>>>Shredder, both the engine and the book have been improved and they also
>>>>constituted a pretty well tested piece of software. The Tournaments have showed
>>>>that the book of Mr. Sandro Necchi has also helped.
>>>>
>>>>Saying that the Shredder´s book has been unimportant is not true ( I would not
>>>>like to use "a big lie" since it is rude term).
>>>
>>>I agree that shredder will always play in tournament by book.
>>>
>>>The point is that even if it has 50% chances to win without book then it is
>>>still better to have 90% chances to win with book.
>>>
>>>I agree that we will not be able to test shredder19 without book against
>>>shredder9 with book so we will unable to test if shredder19 without book is
>>>stronger.
>>>
>>>>
>>>>
>>>>>>(1) Contradiction: "I did not claim that a book cannot help an engine...." ....
>>>>>>"I consider the book as unimportant..."
>>>>>>
>>>>>>(**) read the meaning of unimportant(*).
>>>>>>
>>>>>>
>>>>>>
>>>>>>>strong enough has good chances to win even with 1.h3 but
>>>>>>
>>>>>>(2) I have also suggested that 1. f3!! and 2. g4!! would be a lot better. :))
>>>>>>The tops engine are already prepared for all those idiot moves including 1. h3?
>>>>>>and 1. f3??
>>>>>
>>>>>This is only an example.
>>>>>
>>>>>The idea is clear.
>>>>>There are many ways to get the opponent out of book without lost position and it
>>>>>is not hard to find some line to take the opponent out of book with equality or
>>>>>almost equality with white.
>>>>>
>>>>
>>>>(2) An equal position may not be good for a chess program but great for other
>>>>chess program. An equal position can already mean a lost game for an engine.
>>>>That is one of the advantages of a tuned book: The engine that can get positions
>>>>where it will behave OK and the opponent will "feel" bad, it means the opponents
>>>>will make some mistake. The positions "equal" in chess is a term very relative:
>>>>If Engine A gets a position where it has a clear plan but the Engine B doesnt
>>>>know what to do, you know what the result will be.
>>>
>>>Of course but in order to know that the opponent will "feel" bad you need to
>>>know the opponent.
>>>
>>>If some strong engine is hidden by it's author and made a very big improvement
>>>then you cannot know it's weaknesses.
>>>
>>>>
>>>>If you also give such an advantage with initial idiot moves such as 1. h3?!, of
>>>>course, Shredder will smash anything.
>>>>
>>>>>>(5) Shredder, Junior and Fritz han showed this is not true over the latest years
>>>>>>in Official Tournaments. All of them use strong book tuned by hand. I have not
>>>>>>seen the first case from a no-book engine winning an official Tournament. Where
>>>>>>are the facts?
>>>>>>
>>>>>>AO
>>>>>
>>>>>The facts are that as long as the difference between engines is not very big a
>>>>>book may be important factor in winning tournaments. (******)
>>>>>
>>>>
>>>>Incorrect: Unless you call Blasstradamus, you cannot base your suppositions on
>>>>things that have never happened.
>>>>
>>>>facts <> things that have never happened
>>>>facts = things that have really happened
>>>
>>>
>>>I do not see what incorrect in what I said.
>>
>>Go to (******)
>>
>>facts <> things that have never happened
>>facts = things that have really happened
>
>Ok
>
>I understand what you mean.
>I wrote:
>
>"The facts are that as long as the difference between engines is not very big a
>book may be important factor in winning tournaments."
>
>It should be
>"the facts are that book was an important factor in tournaments in the past and
>the difference between engines was not very big."
>
>>
>>
>>>
>>>I guess that you also agree that a book can be a decisive factor in winning
>>>tournament when the difference between engines is not very big.
>>>
>>
>>Mr. Blass, must I repeat what I have said about one hundred times?!!!!
>>
>>
>>>I guess that even in case that there is going to be a big difference in playing
>>>strength between engines the best engine will use book because it is better to
>>>be sure in 99% in victory then to be sure in 60% in victory so not using book by
>>>the winner is something that I do not expect to happen.
>>
>>You have not discovered anything new that what I have been telling for over 40
>>messages. A tune and tested book is important and it can help the engine to wint
>>games.
>
>Yes
>I did not claim that I discovered something new.
>
>>
>>For the fith time: In my private tests from the 100% of the games won by Diep,
>>30% was a direct win from the book. Why dont you read? (Lack of comprension?!!)
>>
>>More ????!!!!!
>>
>>
>>>
>>
>>Well, you said a book is _unimportant_ (*). Now you say, it is important. What
>>is your position then? I put here some symbols ?!!!***???!!!!
>>
>>>>>>(*) unimportant = not meaning much, not having value or significance
>
>I say that it is dependent on what you want.
>Book is important to achieve what is significant for you.
>Book is not important to achieve what is more significant for me that is
>improving the engine.

I like to win and you dont plan to win: What is the sense to participate in
Tournament if you dont plant to win?


>
>I believe that at the level of movei winning CCT7 was practically impossible
>task for it and the best it could do with better book is maybe second place.
>
>If somebody volunteer to try to help it to get better place in tournaments in
>the future then (s)he is welcome to try to do it but I do not plan to spend much
>time about it.

After all your declarations, you have sowed your own tomb. Well, You had already
been buried yourself several years ago.

>
>I believe that other programmers also in most cases do not spend time on editing
>the book manually and let another person to do the work if they are lucky to
>find somebody to help them.
>

Good Authors know what a book can mean. You will learn that in 20 years perhaps.


>They do not say that book is unimportant(and I guess it was an unsuccesful
>sentence by me that may cause me problems to find volunteers for that task) but
>they also do not spend much time about book.
>

I doubt that you find people willing to help you after you have pointed out
here. I mean people who do a hard work with book: tuning by hand, testing every
variation, etc. Anybody can generate a random books, only some persons have the
patience to do a hard work. Well, see you in 20 years!





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.