[Freeciv] Re: formula idea for ally counting

[Horn Gábor <Horn.Gabor@xxxxxxxxxxx>]

Hi!

Well the r-s in this example was not due to real numbers!!! I hadn't the
script when i wrote this example, i just wrote the numbers (i mean the
'he's better than X percent of players' part...and X is critical!!!) by
a guess. But now i have the modified script so i can test any real game!
Please provide me games which u think is good to demostrate alliances
(because this scripts extends the original by handling them it doesn't
alter the no alliance games' result much, just fixes them a bit. Old
script doesn't recognise dead players, and counts them too.)

In short: numbers weren't mean anything, just the logic. Now we can test
w/ real games as the script is updated not to use simply add points but
the new formula,

bye, hirisov


2003-10-09, cs keltez=E9ssel 19:31-kor Jason Short ezt =EDrta:
> Horn G=E1bor wrote:
>=20
> > v(i)=3Dp(i) + T*(w(ally1)*s(ally1) + w(ally2)*s(ally2) +...)
> >=20
> > v(A)=3D1400
> > v(B)=3D1200 + 1460*(1/20*0.1) =3D 1214.6=20
> > v(C)=3D1000 + 1460*(1*0.22) =3D 1321.2
> > v(D)=3D1100 + 1460*(0.66*0.05) =3D 1148.18
> > v(E)=3D940  + 1460*(1*0.16) =3D 1173.6
> >=20
> It is problematic that E's virtual ranking is pushed up above D's by=20
> this algorithm.  You should guarantee that this doesn't happen (somehow=
).
>=20