[EM] Remember toby KD

[Kevin Venzke]

Hi Kathy,

--- En date de : Dim 29.5.11, Kathy Dopp <kathy.dopp at gmail.com> a écrit :
> > In the "easy to explain/solve but still Condorcet"
> category, I am okay
> > with Minmax(WV) but I think Condorcet//Approval (with
> implied approval)
> > is preferable because you don't need a defeat strength
> concept at all,
> > and it makes burial a pretty clearly unattractive
> strategy.
> >
> > Kevin
> >
> 
> Kevin,
> 
> Could you please explain in fairly simple terms how
> Condorcet/Approval works?
> 
>  I have no problem with Condorcet/Margins approach but it
> might be
> hard to explain to the general public. I like Condorcet in
> general.  I
> also like  Approval because it is so simple and could
> be done using
> today's voting systems (hardware) and ballots.

The way it works is that the voters will submit rankings. Anybody who
is ranked is considered approved. (I strongly recommend against making
approval something that is explicitly marked. If people want a method
like that, don't use this one.)

We will check to see whether there is a Condorcet winner. If there is,
he wins. That's phase 1.

If there's not, the approval winner wins. All rankings count exactly the
same, as one vote. I.e. everybody you gave any ranking to is getting 1
approval point from your ballot.

It is possible to limit the approval phase to candidates who are in the
Smith or Schwartz sets, but I'm not too concerned about that personally.

The most obvious downside to C//A is that, since phase 2 levels all
your rankings, the later-no-harm failures are worse: You are more likely
to regret ranking more candidates. This is like Approval of course.

But the phase 2 leveling (that is to say, the approval part) is also
why burial is deterred: It's undesirable to vote for candidates you don't
actually like, because you will be stuck voting for them (equal to
your favorites) if you succeed in forcing the method into phase 2 (which
would be the goal of burial).

Hope that helps.

Kevin