[Election-Methods] Challenge: Elect the compromise when there're only 2 factions

[Abd ul-Rahman Lomax]

At 11:38 AM 8/22/2007, Howard Swerdfeger wrote:
> > A concrete example: true ratings are
> >    55 voters: A 100, C 80, B 0
> >    45 voters: B 100, C 80, A 0
> >
> > THE CHALLENGE: FIND A METHOD THAT WILL ELECT THE COMPROMISE (C)!
> >
>
>approval, would work if voters approval cutoff is below 80, as would
>honest range.

Of course. But that, of course, is not the problem he intends to present:


> > The fine-print: voters are selfish and will vote strategically...
>
>dang! I would suggest using one of the above methods and removing any
>knowledge from the voter of how any of the other voters will vote.

Use the handy-dandy Mind Eraser, patent pending.

Zero knowledge is, unfortunately, not part of election methods. It's 
a condition, not a method.

>No Media coverage.
>No predictions by political analysts.
>No talking with your family/friends/co-workers about who you will vote for.

No freedom of speech, no freedom of assembly, why bother with voting 
at all? If you have the power to do all this, why are you running an 
election? You just take a poll, at most, if you want to please your 
subjects, and, of course, you let anyone know that if they lie on the 
poll as to their real preferences -- and you have spies everywhere -- 
you will shoot them.

>This would limit the voters ability to vote strategically. i.e. if the
>group of 55 voters don't know they have a majority they would probably
>vote honest in order to get the best result.

Probably. Zero knowledge strategy with those utilities -- which don't 
have to be in common units, they are simply relative for each voter 
-- would encourage, probably, approval-style voting in Range 
including C. Or so-called sincere Range, assuming that the Range 
method had an accurate choice. (Range 2 (CR-3) would not, the 
"approval" vote would be more accurate).