[Election-Methods] RE : Re: RE : Re: Simple two candidate election

[Abd ul-Rahman Lomax]

At 11:18 PM 12/23/2007, rob brown wrote:
>On Dec 23, 2007 7:28 PM, Kevin Venzke 
><<mailto:stepjak at yahoo.fr>stepjak at yahoo.fr> wrote:
>For Range to work as well as hoped, it is important that voters do *not*
>realize how to vote effectively. Or, that they choose not to.
>
>
>You've summed up, in 25 words or less, everything that is wrong about Range.
>
>Very nice job. :)

Would have been, if it were correct. Range, voted with full strategic 
effect, reduces to Approval Voting, which may reduce to bullet 
voting. It *still* is not Plurality, because it only takes a few 
percent of voters adding multiple votes to eliminate the spoiler effect.

There is *nothing wrong* with voters, in Range, voting effectively. 
What I've argued is that there is no reason to suppose that these are 
"insincere votes." If you care, vote full strength!

But Range does not allow you to express maximal caring with more than 
a span of one vote. I.e., if you want to vote A>B with full strength, 
you cannot, at the same time, vote B>C full strength. There is no way 
for strategic voters to get around this. They are given one full vote 
to express. If you prefer A>B with equal strength to B>C, then the 
rational -- and strategically quite powerful -- vote is A:100, B:50, C: 0.

The meaning of this is obvious when specific outcome utilities can be 
assigned. That such utilities are difficult or impossible to come by 
in real public elections does not make the principle inapplicable. If 
you would pay $1000 to insure the election of A over B, and $1000 to 
insure the election of B over C, it does not make sense that you 
would only pay $1000 to insure the election of A over C. -- unless 
you ran into resource limits. Range, as it were, hands voters N votes 
to cast in an Approval election for each candidate. If voters want to 
put all their eggs in one basket, they can. But they might end up 
regretting it. If they prefer A>B>>C, and they have zero knowledge of 
which candidate is more likely to win, the vote of A:100, B:0, C:0 is 
a poor bet, it's easy to show. If A and B are the frontrunners, yes, 
that vote is fine and strategically most effective.

Just as with Plurality and Approval, how one votes depends on the 
context, and, in particular, on the *real* choice being presented. In 
2000, the real choice was not three-way, it was two-way (except for 
minor effects such as federal campaign matching funds). That won't 
change with Approval or Range.

But it might give third parties a leg up, they will be able to show 
vote numbers that are not distorted by strategic necessity. Warren 
Smith makes the point that Range will measure the support for minor 
candidates far more accurately than other methods..... he calls it 
the "incubator effect," and he is likely right, more or less, though 
I think he tends to exaggerate it.