[EM] Re: EM] There's nothing wrong with Average Rating.

[wclark at xoom.org]

Mike Ossipoff wrote:

> CR is strategically equiovalent to Approval: In CR, you maximize your
expectation by giving maximum points to those candidates for whom you'd
vote in Approval, and giving minimum points to the rest.

That's not entirely accurate.  It's true that in CR, you maximize your
expectation by giving either maximum or minimum points to various
candidates -- but *which* candidates you give maximum (minimum) scores
might differ from those you'd approve (disapprove) in Approval.

The reason is that the actions of sincere voters may differ, depending on
whether CR or Approval is used.  These sincere voters can end up changing
the conditions relevant to strategy decisions.  I posted some data
demonstrating this (as a link to an excel spreadsheet) in an earlier
thread, and Richard Moore provided a more general example of such
situations.

Here is a link to the thread:

http://lists.electorama.com/pipermail/election-methods-electorama.com/2004-January/011896.html

As Richard also pointed out, in practice this type of information wouldn't
be available, so strategic voters would likely vote the same in either
case.  But I think it's important to note that in principle, strategic
equivalence doesn't imply equivalence of probabilities (or of results.)

> CR & Approval are excellent, but they don't get rid of defensive
strategy need to the degree that Condorcet wv does.

I like CR & Approval as well, in particular because Arrow's Impossibility
Theorem doesn't apply to them.  That means that some variant of these
cardinal methods might actually be "perfect."

-Bill

-- 
Ralph Nader for US President in 2004
http://votenader.org/