bug-gnubg
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Bug-gnubg] Re: Race database and bearoff DB


From: Jim Segrave
Subject: Re: [Bug-gnubg] Re: Race database and bearoff DB
Date: Tue, 1 Oct 2002 11:38:19 +0200
User-agent: Mutt/1.4i

On Tue 01 Oct 2002 (07:21 +0000), Joern Thyssen wrote:
> On Tue, Oct 01, 2002 at 09:04:29AM +0200, ?ystein O Johansen wrote
> > 
> > Hi,
> > 
> > Yes, I really agree with Joseph here. Can anyone provide any race positions
> > where it misplays the positions? If this is a problem I think we should
> > rather make Joseph's osr (*) algorithm available for gnubg. 
> 
> I have ported Joseph's OSR code to a local copy of gnubg, but I never
> got around testing it.
> 
> 
> > This alogorithm
> > is very good, and it's not that slow. Basically, what it does is that it
> > rolls out each side independently to get a roll distribution for each side.
> > These two roll distributions are combined to give the cubeless winning
> > chance.
> > 
> > I have still not seen this routine make any errors in the race evaluation,
> > though I have seen the race network make some bad evaluations in tricky
> > positions. (I must admit I don't believe this bad evaluations by the race
> > network makes the overall strength of gnubg much worse. I really think the
> > overall strength still lays in improving the crashed network.)
> 
> It's fairly simple to extend the current one sided bear-off database up
> to at least 10 points. It takes around 6 hours to generate on my 1GHz
> box: 3268760 positions ( x 32 values x 2 bytes/value = 209MB). I'll also
> look into extending the exact bear-off database and have it included
> cubeful money game values.
> 
> Maybe it's possible to make another OS bear-off DB that goes even
> further if we restrict it to, say, 10 chequers per point. 
> 
> These one-sided bear-off databases can very easily be combined with the
> OSR code giving fast and reliable evaluations. Joseph also has an
> algorithm for converting OSR evaluations to cubeful equities.
> 
> I'll see if I can get it to work.
> 
> Unfortunately people would have to generate their own databases since I
> don't think we have the necesary band-width/web-space for such files.
> Alternatively, we could sell them on CDs from gnubg.org :-)

The advantages of working for an ISP:

I can easily arrange hosting of something like a 300Mb database on my
employer's public ftp server - ftp.demon.nl. If bandwidth is an issue
for other things, I can also probably hosting of the source and
Windows builds as well. Policy would prevent us giving accounts for
maintaining sites to non-employees, but I could either do mirroring or
scripted copying (scp/sftp preferred).  We are very well connected,
particularly for the Netherlands, UK and US.

Is there any idea how many gnubg users there really are? I don't think
it would be an issue - we've distributed some *huge* games software
updates before without any real problems (except having to move to a
100BaseT uplink as the 10BaseT was saturated for extended periods
after new releases). 

-- 
Jim Segrave           address@hidden




reply via email to

[Prev in Thread] Current Thread [Next in Thread]