Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Interview with Vasik Rajlich (The Fruit of The Rybka)

Author: Stuart Cracraft

Date: 16:33:16 12/20/05

Go up one level in this thread


On December 20, 2005 at 18:14:45, Dann Corbit wrote:

>On December 20, 2005 at 17:03:13, Stuart Cracraft wrote:
>
>>On December 20, 2005 at 17:00:01, Alessandro Scotti wrote:
>>
>>>On December 20, 2005 at 16:30:11, Stuart Cracraft wrote:
>>>
>>>>So what is so special about Fruit 2.1's search?
>>>
>>>It works. ;-)
>>
>>Makes no sense - hundreds, even thousands of programmers have labored
>>for 50+ years on computer chess, and along comes Fruit and Rybka?
>>
>>Some new special trick - I don't think just "well debugged code" is
>>the point.
>
>It's the most important difference between his code and that of many others.  It
>is solid as a tank.  His use of asserts is actually quite beautiful.  It shows
>his clear and clever thinking.

My first-ever programming course was with Bob Floyd at Stanford who pioneered
the point of provably-correct programs with the use of ASSERT and he emphasized
it a lot.

Do I place ASSERT's in my code? I do. I should do another run with them
activated to see if anything has been introduced to hurt the program.

I saw Fabien's ASSERT()'s but didn't think THAT much of them. It's an
established paradigm. Programmers should use it.

>
>>Some new special search tricks.
>
>His branching factor is about 2, just like Shredder.  That's obviously the most
>important thing.  Since he is smart enough to throw away the bad nodes and
>consider the good ones better than other people, his eval and search are clearly
>above average.

Better than mine which is 2-3.

>
>>Now our collective job is to find it and absorb it into the
>>collective Borg-style.
>>
>>It will be assimilated. Resistance is futile!
>>
>>Stuart
>
>It reminds me of the book:
>"One Jump Ahead : Challenging Human Supremacy in Checkers"
>
>Because Fabian will take another step, when we catch up to where he is now.

Yes - and unfortunately no source code then. One-time-gift. Still it was
a good whack on the side of the head and a great kick in the seat of the
pants for the community as a whole.

I liked Jonathan's book. It's in my collection. Schaefer's result is
non-trivial.

Marion Tinsley was remarkable. Nothing like him in any other sport ever as
far as I am aware.  Too bad the crown had to pass due to a health issue.
I guess Capablanca in chess and Shusaku in go come to mind as multi-year
invincible champion-types.

Stuart



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.