[EM] unintended changes in pairwise preferences

[Andrew Myers]

On Mon, Jul 18, 2005 at 04:00:46PM +0200, Kevin Venzke wrote:
> --- Stephen Turner <smturner0 at yahoo.es> a écrit :
> > (2) Among the criteria we usually discuss on this
> > list, we do not have one on "stability", which should
> > mean something like: "a small change in the ballots
> > should change the outcome as rarely as possible". 
> > This seems desirable.  Has it already been discussed
> > somewhere?
> 
> I think, if such a criterion came to be valued, it would result in
> a lot of insensitive methods that behave more like Approval than any
> ranked method.
> 
> I'm lately interested in ranked methods with a strong approval 
> component, but I consider the insensitivity to be an annoyance
> rather than a virtue: Why use a ranked ballot at all if additional 
> voters rarely make a splash?
> 
> Kevin Venzke

I think stability is a very good property for a voting method to
have, and I don't understand the logic that says stability means
you don't want a ranked ballot. The point of stability is that it
gives you more confidence that you've obtained a strong consensus,
which seems to me to be a major goal of Condorcet methods.

I've noticed that in practice MAM -- and the deterministic variant I developed
for CIVS -- both seem to be much more stable than Schulze/beatpath winner,
though I don't have a good argument for why this is. It seems that it's
easier to upend the ordering by creating long, inobvious beatpaths than it
is in MAM.

-- Andrew