[EM] Methods

[MIKE OSSIPOFF]

>> Also, Approval is like a solid, reliable and simple hand-tool. It
>> isn't as labor-saving as a good rank method.
>> The rank-methods are labor-saving machines. But machines can have
>> their problems &/or idiosyncracies.

> The ranking methods are like the saw, labor
> intensive and expensive to use whereas the approval method is like the
> axe, rough and crude but fast and efficient and does exactly what needs
> to be done and no more. 

Approval definitely does what needs to be done. And, as you pointed out,
Approval-polls before the election will give a whole different kind and amount
of information, as compared to our present Plurality polls. Our polls always ask
people how they'd vote if the election were held today. So of course, they vote for
their "lesser-evil", whom they don't like, and no one ever find out, from the poll, 
what people's actual preferences are.
 
That trick wouldn't work with Approval polling. Asked how they'd vote if the election
were today, people would give Approval votes to everyone they like more than their
perceived necessary compromise. There's no way to word an Approval poll that
could avoid the voting public having that information.
 
So yes, with Approval polling, as you said, the first Approval election might be enough
to bring full improvement.
 
(By the way, let's do, in all our communities around the country, a Condorcet poll on the 
2012 candidates, to find out who's CW. Then we aggregate the results, and share that CW info with various progressive
political leaders, small-party leaders, and progressive media.)
 
Relevant to something else said in the thread, in a different message, one of the
good strategies in Approval is to vote for every candidate who's better than your 
expectation in the election. In other words, vote for every candidate whom (if you had
that power) you'd rather put in office instead of holding an election.
 
The result would be: We'd get a result that would be a pleasant surprise to the most 
people, and an unpleasant surprise to the fewest people. 
 
Only a very few of the very best rank methods are as good as Approval.
 
But the best rank methods are quite adequate too. I'd be glad to have them, if
that's what we eventually get. Either would be fine.
 
Of course Approval meets the Favorite Betrayal Criterion (FBC). Condorcet(wv) doesn't.
I've been present when a friend voted in an Internet presidential poll, by rank balloting.
I don't remember what the count method was. Though she prefers Nader's policies to 
those of Kerry and other Democrats (and of course, of the Republicans who have the same
policies), she voted all of the Democrats over Nader. With Condorcet(wv), I couldn't assure
her that she can never benefit from that. That's when I began to feel that FBC is
absolutely essential in U.S. elections.
 
There are some good rank methods that meet FBC, along with other helpful criteria such as
SDSC, SFC.
 
The deciding factor, in the choice between Approval and a briefly-defined, FBC-complying
rank method should be "Which will the voters sign an initiative petition for? Which will they
vote yes for the enactment of? That's what matters. I'd be delighted if voters enacted
Approval or a good rank method.
 
I don't know which people would be more likely to support. I've noticed that rank methods
seem to elicit more interest. More like "Yeah!" than "Yeah".
 
On the other hand, as I said, we'd have the big issue and battle about which rank count. The
IRVists might have already wined, dined, and power-lunched the national leaders of small parties we
approach. Choosing a good method is one thing, but telling the man-in-the-street why it's a
good method is a difficult task.
 
Polling is the only way to find out for sure what the voters will be most likely to enact.
 
> However I think in many elections nuance
> is wasted effort and allowing it is actually harmful to the process,
> especially since ranking and range can be used strategically (I guess
> you guys call it burial?).
 
In some rank methods. In Condorcet(wv). But, there, the order reversers must have
the biggest candidate (in terms of favoriteness), and a quite large fraction of them
must reverse. And only a small fraction of the intended victims need to truncate, in
order for the reversers to get an outcome worse, in their view, than the one that they buried.
 
But there's no offensive order-reversal in Bucklin. Maybe it can be tried in MDDA, but most likely
defensive truncation thwarts it as in wv. I haven't thorougly checked that out yet. The version of
Bucklin that I like allows equal ranking, giving whole votes to those equal ranked, with all rankings
giving to next choices simultaneously if no one yet has a vote total exceeding half the number of
voters. I call that version "Stepwise Approval" (SA), though I don't claim to be its first proponent.
 
There's fascinatingly large array of rank methods that meet FBC, and many that meet SFC & SDSC
too. I've barely begun to check out the reversal-deterrence of the pairwise-count ones. But I've
read of the impressive criterion-compliances that they have.
 
SA has strategy resembling Approval. Though I don' t know it as specifically as Approval's strategy,
SA can be said to have three levels of protection for majorities, as opposed to Approval's two: Equal-ranked
at top; ranked; not ranked. That means that voters could fully protect a majority-supported Democrat from
the Republican, while still fully protecting majority-protected progressives from the Democrat.
 
But you're right: Rank methods, as good as some are, are gadgets, and the solid, sturdy and
reliable hand tool, Approval is quite sufficient.
 
Let's find out what the public will most support, because that's the deciding factor.
 
And, here, let's keep systematically discussing and listing the criterion-compliances of the proposed
methods, and also discuss the desirability and need for the various criteria, and why one or another
method will better suit the needs of voters.
 
Mike Ossipoff
 
 
 


  		 	   		  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20111018/5de14a4c/attachment-0002.htm>