Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Has Christophe's 0.01% chance actually occurred?

Author: Sune Fischer

Date: 00:54:57 09/26/02

Go up one level in this thread


On September 25, 2002 at 20:34:04, pavel wrote:

>Sune,
>      I still think some of the reactions has more to do with jealousy than
>sketicism. I mean if you look at the speculations (and some accusations) I
>listed in my previous post, and if you think about some of them, you will notice
>how dubiously far-fetched those speculations and accusations are.
>
>I would expect someone to take this is as an example, and work harder to improve
>his engine, not stratch their head and try to find faults that isn't there.
>
>IMO the previous clones were much too obvious, and clones are much too easy to
>catch in general.
>
>cheers,
>pavs ;)

Well, I just don't think it's anything to get upset about. Any new and strong
engine must pass the clone test, that part is pretty much standard procedure,
nothing personal or insulting in that what so ever in that as far as I'm
concerned.

I agree with anyone saying that for a new engine it is suspiciously strong,
isn't that the truth? However a few things, like the WB and UCI support make it
unlikely in this case.

I think if one really wanted to make a clone, the approach should be a little
different. Take a strong engine and weaken it, release it with lots of silly
bugs and ask a few standard questions about hashing. Start removing the bugs and
renaming the functions and variables (I don't if those can be grep'ed?), alter
the extensions and evaluation parameters just a bit. Over a 6 months timespan
make a few releases while you slowly remove the inserted bugs and put it back up
to full strength.

You now have a clone and no one will be the wiser :)

-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.