[Freeciv-Dev] Re: Comments on CMA 2.6

[Raimar Falke <hawk@xxxxxxxxxxxxxxxxxxxxxxx>]

On Sun, Oct 21, 2001 at 08:41:05PM +0200, Christian Knoke wrote:
> Am Sonntag, 21. Oktober 2001 19:51 schrieb Raimar Falke:
> > On Sun, Oct 21, 2001 at 07:26:15PM +0200, Christian Knoke wrote:
> > > Am Sonntag, 21. Oktober 2001 15:01 schrieb Raimar Falke:
> > > > On Sun, Oct 21, 2001 at 01:29:42PM +0200, Christian Knoke wrote:
> > > > > Am Sonntag, 21. Oktober 2001 09:08 schrieb Raimar Falke:
> > > > > > On Sun, Oct 21, 2001 at 12:54:23AM +0200, Christian Knoke wrote:
> > > > > > > The resolution for the sliders is to high.
> > > > > >
> > > > > > I don't understand.
> > > > >
> > > > > Left side sliders: You'll never need -2..-20 surplus for
> > > > > anything. Just allow negative surplus or not.
> > > >
> > > > Ack.
> > > >
> > > > > Why not: (negative, 0, positive) three possibilities, where
> > > > > positive is just (+1).
> > > >
> > > > No. You may need +2 or +3 to finish your unit in n turns. So a
> > > > slider is still needed.
> > >
> > > Yes, but really really seldom. You'd rather buy. Too much hassle
> > > with the agent. In high production phases, you would need to set
> > > +30 e.g.
> >
> > No problem of expanding the value range.
> 
> But a problem for the user to set the right value.
> 
> >
> > > > > Right side sliders: Just (0, 12, 25).
> > > >
> > > > I think more steps are necessary. Not 25 but more. Also note that
> > > > you may not be able to express everything with say 4 steps
> > > > (1,2,3,4).
> > > >
> > > > > I don't know what algorithm you used, are these numbers
> > > > > weights?
> > > >
> > > > Yes.
> > >
> > > Ok. Then I have to ask: what scale did you use? What is G=1 vs.
> > > P=10 e.g. Sorry that I ask, I should read the source, probably.
> > > I think there should be weights that behave like 1:2, 1:3, 1:4,
> > > and 1:5.
> >
> > Example with only two stats: food and shield. The city may provide
> > the following combinations (all surplus): (f=0,s=10), (f=2,s=7),
> > (f=11,s=2) and (f=12,s=-1). You now search the maximal sum of
> > sum=food_weight*food_surplus+shield_weight*shield_surplus.
> >
> > food_weight shield_weight best_sum  best_combination
> >     1             1           13      (f=11,s=2)
> >     2             1           24      (f=12,s=-1)
> >     3             1           35      (f=11,s=2) or (f=12,s=-1)
> >     4             1           47      (f=12,s=-1)
> >    10             1           120     (f=12,s=-1)
> >
> >     1             2           20      (f=0,s=10)
> >     1            10           100     (f=0,s=10)
> >
> > Was this clear?
> 
> Yes. So the maximal ratio is 1:25, i.e. give up 25 food for
> a single shield (f=25, s=1). That is pretty much. 1..10 or
> even 1..5 would be sufficient, don't you think so? I never
> want to loose 10 gold for 1 shield e.g. On the other hand,
> if you reintroduce the zero, one could claim the food (surplus)
> not to be weighted at all. Or do I miss something?

With say 1..5 you may not be able to set every ratio. For example the
current server ai has:

#define FOOD_WEIGHTING 19
#define SHIELD_WEIGHTING 17
#define TRADE_WEIGHTING 12

> > > > > Then what does the 0 means?
> > > >
> > > > You can't use the 0. 0 means no influence but this isn't what is
> > > > needed because every stat should have an influence.
> > >
> > > Oh, it's 1..25 not 0..25, sorry.

        Raimar

-- 
 email: rf13@xxxxxxxxxxxxxxxxx
 "Your mail could not be delivered to the following Address:
  VTCMC.VTLPR@xxxxxxxxxxxxx        ** Unassigned error message **"