[EM] Market-Based multiwinner system II
Kristofer Munsterhjelm
km-elmet at broadpark.no
Thu Feb 18 07:51:59 PST 2010
Warren Smith wrote:
> Here's a different rule set which might be better for the internet
> age. It pretty much depends on modern technology (fortunately, or
> not).
>
> W-winner election with C candidates, 0<W<C. We assume there is a
> public "bulletin board" on which the "total" for each candidate is
> announced. These public totals are updated in real time. Each voter
> has 1000 pseudodollars and can allocate them in any way to the C
> candidates, and can change her allocation at any time until the
> pre-announced final deadline. When the deadline comes, the W
> "richest" candidates win.
>
> This system automatically seems to obey a "proportionality theorem."
>
> It is very simple (provided you are willing to trust the technology).
>
> It also elects parliaments that are "optimal" in the sense that if there is any
> reasonably obvious change that enough voters feel would be an
> improvement, they can make it happen?
>
> A criticism is: if not-very-good candidate B is currently "rich"
> enough to win a seat,
> and better candidate C isn't, voters may be afraid to move their
> "money" from B to C
> because to do so they'd have to pass through an intermediate state
> where neither B nor C was winning, but (therefore) ultra-bad candidate
> U was winning. So they'd stay with B
> for fear of "wasting their vote." This -- if it happens -- would be a
> vicious cycle causing initially-ahead candidates to stay ahead
> regardless of their actual quality, and thus causing initial
> perceptions from propaganda and big money to have too much power.
>
> My original rules discussed at
> http://rangevoting.org/MarketBasedVoting.html
> seem to be better in that respect (seems to be no such thing as a
> "wasted vote"), and do not require high technology.
A possible way to make this more like STV would be to have a rising
barrier. As the method nears the deadline, candidates below a certain
quota are excluded and the ballots are renormalized. If the voter wants
to change his vote at any point, he may do so.
Being more like STV, it would somewhat fix the "move your money from B
to C" problem. However, it would be much more chaotic than your system.
More information about the Election-Methods
mailing list