[EM] obvious possible CS fix

[MIKE OSSIPOFF]

Of course one way to maybe fix the CS fault that I described yesterday
would be to require that if a ballot indicates a candidate as its
favorite, it must also vote for that candidate as if it _is_ favorite.
So, if it's an Approval ballot, it must vote for that "favorite"
candidate. If it's a ranking, it must rank that candidate in 1st place,
though it could strategically rank someone else there too. If it's
a CR rating ballot, it should give the maximum points assignment to that
favorite candidate.

So in Approval+favorite CS, if some practical joker indicated some
notoriously unpopular character as their favorite, and voted for him,
and also for candidates chosen to make the ballot as close as possible
to the other voters, that "favorite" would be at at least a slight
disadvantage with respect to the other voters' candidates, because he
would have a distance-point for each of the other voters, and would
be the only candidate with a distance-point from all the voters but
the joker. But I don't know that one couldn't write an example in which
the unpopular candidate wins anyway.

If CS uses rankings or ratings, of course the unpopular candidate's
distance from all the other voters would be especially great, the
full maximum distance possible in one dimension (his own dimension).
For instance, in CS with 0 to 10 CR, that candidate would be 10 points
distant from all the other voters in his own dimension. The fact that
that's a maximum distance in that dimension from everyone else would
matter all the more strongly if Pythagorean distance were used. The
results with Pythagorean and city-block distance in my example suggest
to me that Pythagorean distance might turn out to be best.

Can it be shown that, by requiring voters to vote for their indicated
favorite as if s/he is really their favorite, it wouldn't be possible
for someone to make someone win whom no one but him votes for? Can it
be shown how popular, by some measure, a candidate must be to win in
CS? To alleviate concern about the CS fault that I described yesterday.

Anyway, not knowing that, I've said all I can about CS.

It's all hypothetical now, anyway, since CS is too complicated and
unconventional in its definition, and probably too unconventional in
its behavior, to be proposable till the public take more interest
in single-winner reform, after Approval, CR, etc., has been adopted.

Mike Ossipoff











































_________________________________________________________________
MSN 8 helps eliminate e-mail viruses. Get 2 months FREE*. 
http://join.msn.com/?page=features/virus

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em