[Top][All Lists]
[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Bug-gnubg] luck analysis suggestion
From: |
Joern Thyssen |
Subject: |
Re: [Bug-gnubg] luck analysis suggestion |
Date: |
Fri, 6 Dec 2002 20:16:16 +0000 |
User-agent: |
Mutt/1.4i |
On Thu, Dec 05, 2002 at 09:20:38PM +0000, O.M. Riordan wrote
>
> I have a suggestion about luck analysis, namely to have an option to
> increase the depth of the analysis used (probably 1-ply analysis of each
> roll rather than the current 0-ply is most likely to be useful).
>
> (Better luck analysis would be very useful for comparing gunbg with other
> programs - the quickest way to get stastically significant results is to
> use unbiased luck analysis on matches between them; the current analysis
> is not good enough to give quick convergence, as shown by analyzing sample
> matches and comparing the appropriate sum/difference of luck and error
> rates with 50%. Using deeper analysis, together with results from Tony
> Lezard's Dueller, it should be practical to compare gnubg with the rather
> more expensive Snowie 4; since the difference in strength seems likely to
> be small, this would be good for gnubg.)
>
> Oliver Riordan
>
> P.S. I noticed that the changelog says that luck analysis has been changed
> to use cubeful analysis, whereas this is not the case. (There is still a
> FIXME in analysis.c about thsi.)
Should be fixed now.
I've added a new set of commands: "set analysis luckanalysis ..." that
can be used to specify the evaluation parameters for the luck analysis.
For example,
set analysis luckanalysis plies 2
set analysis luckanalisys cubeful on
The default settings is cubeful 0-ply.
Jørn