Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Another example of things that could happen

Author: Robert Hyatt

Date: 15:58:19 12/10/03

Go up one level in this thread


On December 10, 2003 at 18:23:44, Omid David Tabibi wrote:

>On December 10, 2003 at 17:55:54, Robert Hyatt wrote:
>
>>On December 10, 2003 at 16:09:34, Omid David Tabibi wrote:
>>
>>>On December 10, 2003 at 07:02:14, Gian-Carlo Pascutto wrote:
>>>
>>>>On December 10, 2003 at 06:55:35, Omid David Tabibi wrote:
>>>>
>>>>>Another scenario:
>>>>>
>>>>>Sjeng isn't having a lucky day and in move 16 of a tactical variation in
>>>>>Sicilian, suddenly the amateur opponent engine plays a brilliant sacrifice
>>>>>resulting in a forced win. But your opponent frowns and realizes this is the CB
>>>>>GUI and not his engine (which doesn't support book at all). He requests to take
>>>>>back the move played by the GUI, disable book in GUI, and let the engine try to
>>>>>find the move on its own.
>>>>>
>>>>>Of course you know the engine can never find this mate on its own, so if you
>>>>>allow it you are saved and if you refuse you lose the chance for the world
>>>>>title.
>>>>>
>>>>>Do you consider it reasonable to allow him to do this? I DON'T!
>>>>
>>>>My question here would be who made the book.
>>>
>>>Let's look at it this way:
>>>
>>>The author created the book himself, but didn't write the access code. It is
>>>pretty much like EGTB, you use the EGTB but haven't written your own access code
>>>for it.
>>>
>>>The only question you will ask now, is whether the author has written the EGTB
>>>himself? No, but he does have permission to use it I guess. For example, if you
>>>get special permission to use the fritz opening book, you can use it. That's
>>>also the case with EGTB.
>>>
>>>So, basically, there is no difference between the interface playing from book,
>>>or from EGTB.
>>>
>>>
>>
>>Actually there is a _big_ difference.  Playing from an egtb is a deterministic
>>procedure.  There is no choice.  You just pick the moves at leads to the
>>shortest mate.  With a book, there is _plenty_ of room for creativity in
>>choosing a book line.  IE you have lots of information about a particular
>>book move:  wining percentage, number of times played (higher means the
>>winning percentage is more reliable), learning scores, maybe CAP scores,
>>maybe human comments (this is aggressive, this is passive, this is drawish,
>>this is sharp, etc.)  How you use all that information to choose a single
>>book move is much more creative than just looking up a position in an
>>endgame table.  And, in my case, I can actually choose a sub-set of book
>>moves and then search them to choose the best, if I want...  Which gets
>>the engine involved in choosing a book line also.
>
>Yes, but when using the popular interfaces like Fritz or Shredder, the engine is
>totally asleep in the opening phase.
>

For that interface, yes.  But we are talking about "interface" in the
general sense.  But even without that, the GUI makes several decisions in
choosing a book line.  That is different than a table base probe which is
just a direct index into a table to get a number.


>
>
>>
>>
>>
>>
>>>>
>>>>--
>>>>GCP



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.