Re: [Bug-gnubg] gnubg "personalities"

[Christian Anthon]

On 7/7/06, Jonathan Kinsey <address@hidden> wrote:
> What do people think about altering gnubg play depending on some kind of
> "personality" setting?
gnubg having a personality is secondary to me, but I would like to
have some knobs to turn, both for chequer play and cube evaluations.
What I can come up with right know are

1) volatility (as measured by the std. deviation on a 2ply evalution??)
2) number of blots
3) number of points
4) number of specific points (home board, enimy board, low anchors,
high anchors midpoints...)
5) pipcount, pipcount difference
6) gammons, single games

likely many more.

> Ps. I think I saw this in a comment about the
> upcoming zbot backgammon program.

Yes, and likely zbot will have an advantage here, since from what I
have read, the zbot net is developed to be more human like, that is,
you can ask questions do you understand this position, is it
volatile...

Christian.