Re: [Bug-gnubg] A newbie question - GNUBG and neural network...

[Erez]

Thanks Christian for your answer.
So if I understand correctly, and correct me if I'm wrong, GNUBG doesn't 
'think' what would be the best move but chooses a move that is in it's 
database?
I would assume that if this is the case it will need a very very big 
database...

Thanks,
Erez


----- Original Message ----- 
From: "Christian Anthon" <address@hidden>
To: "Erez" <address@hidden>
Cc: <address@hidden>
Sent: Friday, February 06, 2009 9:07 AM
Subject: Re: [Bug-gnubg] A newbie question - GNUBG and neural network...


> There is three readily available backgammon programs that play at
> approximately the same level: Snowie, GNU Backgammon and BGBlitz. The
> two first probably play a bit better, but BGBlitz steadily improves.
>
> No the net only gets better with intensive and educated training, and
> nobody does that at the moment, but in most common positions there is
> little room for improvement.
>
> Christian.
>
> On Thu, Feb 5, 2009 at 6:38 PM, Erez <address@hidden> wrote:
> > Hello guys,
> >
> > I have a very basic question about the level of GNUBG.
> > A friend told me that GNUBG won some sort of computer BG tournament so I 
> > got
> > interested, read a bit and found out that it uses a neural network. Does 
> > it
> > mean that my downloaded GNUBG will get better and better as long as I 
> > play
> > it, or does the program gets better in time and my downloaded version 
> > will
> > get better when it will be updated?
> > (I guess it's the latter one but I'm not sure)
> >
> > Thanks
> > _______________________________________________
> > Bug-gnubg mailing list
> > address@hidden
> > http://lists.gnu.org/mailman/listinfo/bug-gnubg