Hello Christoph,

I agree with you that this algorithm is not scalable if you use this 
definition, but : 

When people say that MC infinite scalability is mathematicaly proven, they do 
not refer to the definition you give, they refer to the definition I used.

If you want to do a theoretical analysis of MC scalability with your 
definition, you will first have to define exactly what you call "a better 
move". This might be tough. 

Even if you managed to do that, you would see that MC is not scalable with this 
definition, because obviously, more computing power can, in some occasions, 
provide a worse move (what ever you call a worse move).

Ivan

----- Message d'origine ----
De : Christoph Birk <[EMAIL PROTECTED]>
À : computer-go <[email protected]>
Envoyé le : Mardi, 22 Janvier 2008, 22h50mn 29s
Objet : Re: Re : [computer-go] Is MC-UCT really scalable ... is a troll

On Tue, 22 Jan 2008, ivan dubois wrote:
> in theory, infinitely scalable. For example, the folowing algorithm is 
> infinitely scalable :
>    Analyse the complete mini-max tree of the game. If enough time to 
> finish, returns the correct move, if not, return a random move.
> Now, is this algorithm really scalable, in practive ? Of course not.

I have to disagree. The above algorithm is not scalable in a sense
that more time leads to better moves. Ie. It returns a random move
given 1 minute, 1 hour or 1 day of (current) CPU time, while a
scalable algorithm should return a better move the longer it runs.

Christoph

_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/


      
_____________________________________________________________________________ 
Ne gardez plus qu'une seule adresse mail ! Copiez vos mails vers Yahoo! Mail 
http://mail.yahoo.fr
_______________________________________________
computer-go mailing list
[email protected]
http://www.computer-go.org/mailman/listinfo/computer-go/

Reply via email to