[EM] Dave: Emphatic preferences in Approval

[Dave Ketchum]

On Wed, 19 Oct 2005 05:00:52 +0000 MIKE OSSIPOFF wrote:

> 
> I"d said:
> 
>> In a posting some days ago I pointed out that, with pairwise-count, if 
>> voting is sincere and more prefer X to Y than Y to X, then X will beat 
>> Y pairwise.  And that, in Approval, if more people emphatically prefer 
>> X to Y than Y to X, then Y can't win, because X will outpoll Y.
>>
>> I pointed out that, in that way, by emphatic preference, Approval 
>> matches what Condorcet does.
> 
> 
> You replied:
> 
> That is no more useful than measuring hair color, when the task depends on
> muscle.
> 
> I reply:
> 
> Come again? That depends on what the task is. If the task involves 
> ordinary preferences, then emphatic preferences don't help the task. In 
> particular, if the task is to minimize strategy need, then the ordinary 
> preferences counted by MDDA will do a better job than the emphatic 
> preferences counted by Approval.


The task is allowing the voters to express their preferences in a manner 
that best lets their preferred candidate get elected.

In sorting this out we do need to get into exotica.

Perhaps I overstated my case, but your words implied unreasonable 
preference for Approval.

> 
> But that doesn't mean that emphatic preferences aren't valid or 
> meaningful. In a way they're more meaningful. In a strategy-free 
> sincere-voting society, where people don't care about optimizing the 
> outcome for themselves, but only care about the greater good, and want 
> to vote sincerely, and have the election maximize social utility, then 
> the emphatic preferences of Approval are more important than ordinary 
> bare preferences.
> 
> Because, if the task is to count the strongest preferences, and not 
> count a weak preference equal to a strong one, then Approval is what 
> accomplishes the task.


Agreed that it is undesirable to count a weak preference the same as a 
strong one.

Still, in Tom, Dick, and Harry, after ranking Tom and Harry, and having a 
weak preference for Dick over Harry, and a weak preference for Tom over 
Dick, Approval forces me to vote a STRONG preference for Tom over Dick, or 
else for Dick over Harry.

Agreed that ranking is not perfect for indicating strength of preferences, 
and that rating allows more precise statements, rating gets in trouble 
when the ballot counters try to understand what the voters have said.

> 
> You continued:
> 
> For that matter and your measurements, does not Plurality land in the same
> ballpark?
> 
> I reply:
> 
> Go for it. Define a type of pairwise prefernce based on that.  But I 
> don't think it would mean as much, since Plurality doesn't allow a voter 
> to rate all the candidates.


Plurality and Approval are alike in a way that makes me dislike both - 
they each let the voter rate all the candidates, but into ONLY TWO groups 
- acceptable and rejectable.  Only difference is that, with Plurality, the 
acceptable group is limited to one member.

> 
> Mike Ossipoff

-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.