[EM] Rob: Condorcet's strategy problem

[Rob Lanphier]

Hi Mike,

Comments below.

On Thu, 2005-09-15 at 02:53 +0000, MIKE OSSIPOFF wrote:
> You wrote:
> 
> Incidentally, this particular scenario presents some nasty strategy
> problems for Range and Approval as well.
> 
> I reply:
> 
> Nasty enough to make voters bury their favorite?
>
> Nope. That incentive will never exist withApproval or Range Voting.

That wasn't what I was referring to.  Where do I draw the Approval
line?  

The strategy I've always heard for Approval is "vote for your favorite
frontrunner, and everyone you like better", with the assumption always
being that there's at most a couple frontrunners.  That becomes a
problem with a close three-way race.  If I like A, I hate B, but I
really hate C, do I risk approving B to make sure C doesn't get elected?
Sounds like I might end up back in the position of approving "the lesser
of two evils".

> It's a question of which you'd rather the LO2E progressives do: Shall they 
> vote Dean equal to Nader, or shall they vote Dean, and all the other 
> Democrats, over Nader, when preferring Nader to the Democrats?
> 
> Sure. Some will, some won't, and we can't predict the exact percentages. 
> Maybe only a few will. Maybe my report about the two voters I can speak for 
> is just a very unlikely coincidence. But, based on what is known to be 
> optimal for voters who perceive an acceptable/unacceptable situation, 
> there's a good chance that lots of LO2E progressives will do that with 
> Condorcet, in public political elections.

Over the long haul, with feedback from actual elections and learning
from the consequences of their decisions, we can hope that they would do
what is in their best interests.  Voting sincerely is going to be the
right strategy in the vast majority of cases.  If we assume that voters
are going to do stupid and counterintuitive things, then we should
probably rethink the whole idea of democracy.  Garbage in, garbage out.

> Please note that I myself like wv Condorcet, and that I advocate 
> BeatpathWinner or CSSD for committees and organizations, where there's no 
> really scary lesser-evil.

I'm assuming that year five of the Bush administration has worn down
your resolve to push anything less-than-safe for public elections  ;-)
I can't blame you, but I still think Schulze(wv) is the best way to go.

Let's take the worst case scenario with Schulze(wv).  You and other
Nader supporters decide to take the "safe" road, and rank lesser-evil
candidates (e.g. Kerry, Gore) above your favorite.  That strategy is
every bit as safe, if not safer, than Approval, isn't it?  If there
happens to be be Republicans less scary than Bush on the ticket (e.g. a
Republican who would make a competent administrator and who at least
reads newspapers), you can rank them higher, just to be sure.

If your goal is to avoid favorite betrayal at all costs, you're right,
Approval is best.  We'll have to agree to disagree, because that's not
my goal.

My main goal is to correctly find the candidate who would beat every
other candidate in a head-to-head election, because I believe that
candidate is likely the best candidate for the position.  Also, I'd like
to make sure that scary extremists of an opposite political bent than
myself don't get into power.  The price I'm willing to pay is that I'm
willing to live without seeing a "scary extremist" that I support get
elected without majority support.

Given the scenarios I've seen where sincere voting in a Schulze(wv)
election could lead to regret, it seems like a really small risk.  Not a
risk worthy of scrapping the whole thing.

Rob