Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Thinker 4.6b third after 1st round!

Author: Robert Hyatt

Date: 08:49:48 06/02/04

Go up one level in this thread


On June 02, 2004 at 03:46:27, Sune Fischer wrote:

>On June 01, 2004 at 18:50:22, Robert Hyatt wrote:
>
>>Please.
>>
>>I had a choice:  (1) continue the hand-tuned book approach I used in the Cray
>>Blitz days;  (2) develop an alternative.
>>
>>I simply chose (2) because I wanted to see what was possible.  Now you want to
>>punish any program that uses (2) but let the ones that use (1) continue to enjoy
>>they hand-prepared book.
>
>I don't want to punish anybody, I do however want to allow the user to do as he
>pleases without hearing me complain.
>
>At the point of release we relinquish all control of the engine.
>Expect the settings to be altered, expect the engine to play in all sort of
>weird environments.
>
>Don't complain about it saying that's not how it meant to be played.
>If you won't acknowledge that the user is boss at his end then don't release the
>engine to him.

That's a completely nonsensical approach to anything.  I designed Crafty with
one goal in mind, playing chess.  Crafty is a "package".  Its book.  Its
learning.  Its search.  Its evaluation.  Its pondering.  Breaking any of those
makes little sense since it becomes "not crafty" at that instant.


>
>>It makes absolutely no sense.  Why not play without books?
>
>There are problems with determanistic behavior of the engines.

And there are problems with non-deterministic behavior of books.




>
>>Why not play with a
>>common (bad) book?
>
>Done often.


And just as worthless of course, unless you are trying to determine where your
engine does well or poorly against a specific program.  But the result won't say
one thing about how you will do when you play the _real_ program in an important
game...



>
>> But certainly don't play with a book hand-tuned to program A
>>and program B might well do poorly with it.
>
>Why not, it might help you locate weaknesses.


A person playing a basement tournament is not trying to fix weaknesses.  That is
where this thread started.  Not on author testing, which is a different thing
entirely...



>
>>The only viable choice is "hands off". Let the program have the best
>>opportunity to choose book lines by itself.  That is the very essence of
>>computer chess anyway, _fully_ self-contained chess playing.
>
>That's certainly an interesting configuration.
>Does that mean all other configurations are worthless? No!

Fine.  Go drive your car with 3 wheels mounted.  Report back.  I'm sure everyone
is _really_ interested in that result.  Let's make it the left front wheel for
simplicity.  Compare your top speed to a 3 cylinder Geo Metro.

Then lets have a real race to see how well your testing worked.


>
>
>>>I'm afraid that's just the way it is, I do the things I find interesting and
>>>funny to do. If people download it and turn it off as the first thing, well
>>>it's none of my business.
>>
>>No, but I can make that as impossible for Crafty as it is for programs that
>>don't have it in the first place...
>
>If you make it impossible I think Crafty will be kicked out of a lot of
>tournaments.
>People want to control the engine, if they can't the engine becomes less
>valuable to them.


If turning off learning gets it kicked out, that's fine by me.  I didn't write
it to participate in oddball-configured basement events.  I wrote it as a
stand-alone system to play chess.




>
>-S.



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.