Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: CCT5 planning started

Author: Sune Fischer

Date: 07:18:10 11/12/02

Go up one level in this thread


On November 12, 2002 at 09:48:24, Russell Reagan wrote:

>Maybe, but I don't think that would "prove" anything. The nature of SMP engines
>is that they are non-deterministic (IIRC), so if my program plays a different
>move from the program that I sent to the trustee, so what? I can say, "I am
>using a Unix machine at school that my professor let me borrow for the
>tournament and it has 64 processors." or I can just say that my program has
>complex AI and it makes decisions that are usually not reproducable. There are a
>lot of excuses I can use.

It is hard to reproduce things exactly, also because game senarios fills the
hashtable in a way that you can't easily reproduce with analysis on a single
move.

Suspicion of clones or time management cheating could probably be decided if one
were to have the binary.
I don't see how printing the pv or sending the logs (which can be edited
quickly) can in anyway reveal if someone is running their engine in some special
analyse mode and forces the moves manually at the right time.
I think that would be easy to do, actually.

-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.