[EM] Addendum re: unselfish nonmajoritarian sincere-rating conditions
Michael Ossipoff
email9648742 at gmail.com
Fri Nov 1 07:47:33 PDT 2013
For a late and abrupt increase in the magnitude of the weighting of
negative ratings, I guess a hyperbola would be even better than an
exponential function:
y = -A/(x+B)
...where x is any negative rating-value, and y is its
altruistically-magnified value,
...and A and B are positive constants.
B is the distance between the hyperbola's asymptote and the y axis.
But, to pass the function through (0,0), an additional term is needed:
y = -A/(X+B) + A/B
To avoid y having less magnitude than x, that should be:
F = -A/(X+B) + A/B
y = min(x,F)
Where the ratings-range is[ -N to +N], B would be chosen to be
slightly larger than N. A and B would be chosen for what seems the
right combination of fairness and practicality.
Of course nommajoritarian unselfish sincere-rating conditions are
farther away than any of the other conditions that I've defined, and
so these #4 methods don't now have practical relevance for public
political elections.
Michael Ossipoff
More information about the Election-Methods
mailing list