gnugo-devel
[Top][All Lists]
Advanced

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

RE: [gnugo-devel] Move valuation question


From: Portela Fernand
Subject: RE: [gnugo-devel] Move valuation question
Date: Sun, 19 Jan 2003 19:00:48 +0100

Arend wrote:

> > - the score (balance in terms of solid territory)
> > - the "power" (balance in terms of influence)
> >     the implementation currently uses the territory valuation,
> >     but this is possibly not the best choice for that purpose.
> > - and the game advancement (fuseki, chuban, yose)
> >     returned as a value between 0.0 (start) and 1.0 (game is over)
> 
> ...IMHO you should just use the score as calculated by the influence code,

My experiment in this new function is not meant to replace the current
score estimation function(s). My idea was rather to see if it would possible
(with cheap calculations) to get an idea of how much territory is "solid",
and more importantly, the balance of this measure. influence_score() is
interesting (although not used at all currently), but I'm afraid it mixes
different things, e.g. (yet definite) territory and what I called "power" by
lack of a better name. So I tried to code something different. The idea is
to be able at a later point to use them in strategic decisions or to
influence the engine somehow. An example (and nothing more than that):
- clear lead in power/influence ?
  => attack !
- ahead in territory, but behind in power/influence ?
  => invade !
- behind in both ?
  => err... attach, cut, overplay ! :p

Of course, this patch is far from being ready for such use. While the score
estimation part is rather ok IMO, I think the measure of the power balance
is way off. The last element, game advancement measure looks about right to
me. There are possibly better ways to do it, as you suggest.

> instead of creating new tunables with WEIGHT_TERRITORY etc. E.g.

These are used for the game advancement algorythm only.

> Certainly you would have to make it dependent on the number of handicap
> stones, too. I think
> influence_delta_territory(INITIAL_INFLUENCE, INITIAL_OPPOSITE_INFLUENCE)
> would be a useful criterion, too (for measuring how much the position is
> played out).

I didn't test my patch with handicap games. I will do soon and see how
well or bad it works.

Thanks for the comments and suggestions.

/nando




reply via email to

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