[Election-Methods] RE : Is "sincere" voting in Range suboptimal?

[Abd ul-Rahman Lomax]

At 09:51 AM 7/24/2007, Kevin Venzke wrote:
>Warren implemented his own version afterwards; I suggest his results if
>you're really interested. http://rangevoting.org/RVstrat3.html

Yes, I'm familiar with the page. It's worth commenting on, I think. 
First of all some might assume that these are the simulations we 
refer to when, generally, we claim that simulations show Range as one 
of the very best methods. No, those other simulations are looking at 
overall social utility. Somebody should post references to the IEVS pages.

These simulations are looking that the return to the voter from 
various strategies, a direct answer to the issue posed by those who 
claim that voting Approval style is optimal. It turns out that it's 
optimal in some limited cases:

The result of the studies cited: in large elections with three 
candidates, the optimal strategies are Mean-Based Thresholding, which 
is an Approval strategy, and Bisector-based Thresholding, a different 
Approval method which, with the utility distributions in these 
simulations, appears to be identical in result (though individual 
voter votes may vary). However, Scaled Sincerity -- which is what I 
called Normalized Sincere Range -- is close behind.

But with very few voters, Plurality actually beats those methods, and 
so does Scaled Sincerity.

There is a problem, though, one that my simulation tries to answer. 
My work is exhaustive, as far as it goes. By excluding moot votes, it 
is far more accurate. (Warren's simulations effectively exclude moot 
votes, but in large elections the probabilities are so low that there 
is a noise problem. Warren claims sufficient significance, however, 
to confirm the method rankings he found.)

The simulations with results on the cited page, inspired by Mr. 
Venske's work, use a primitive random distribution, quite like the 
"zero knowledge" distribution in my present work. The utility 
distributions are even, random for each voter. This is *not* the case 
with the general purpose simulator Warren has built, rather IEVS has 
utility input options that can use various distributions.

So what is on these pages, rather primitively in some ways, studies 
effecdt of various Range Voting strategies, a question of major 
interest. They do not show how "sincere votes" stack up against 
"approval votes" on the part of the rest of the electorate.

But the question is very important. If we start out with an 
assumption of sincere voters, does a voter improve his utility by 
voting sincerely or by voting Approval style? If not, then the 
alleged motive for voters to vote Approval style, *in general* does 
not exist, and there is little reason to suppose that it would become 
the norm, even if voters are "selfish."

In my own study, I take a quite different approach. My study is an 
exhaustive look at all the possible vote combinations that can affect 
the outcome; thus there may be some improvement in expected utility 
over simply not voting. Given that voters don't just think of 
themselves when they vote, but understand that how they vote, others 
will also vote, because people tend to think alike, it is, I assert, 
quite proper to study and consider significant the effect of a single 
vote by excluding moot votes. If my vote does not affect an outcome, 
(and one kind of outcome is that I cannot cause a particular 
candidate to win, or I cannot cause a particular candidate to lose) 
then I have no cost to sincere voting, nor any penalty for voting any 
other way. Most people, faced with such a situation would either not 
vote at all, or would vote sincerely. Why not?

(In polls, however, people do lie or exaggerate, even when no 
specific consequence is going to come to them, because they know that 
there may be a *general* consequence. The results of a poll may be 
used in some way. If I give a moot vote to a candidate, I may be 
helping that candidate in future elections. All of this, though, 
generally points toward voting sincerely, by some definition, as 
optimal, *aside* from improvements in expected return from voting 
with the best strategy.)

I'll point out that this page is an example of how the Center for 
Range Voting works. It is here reporting information that could be 
considered damaging to Range as far as political prospects. For other 
reasons, I think that this is *not* damaging. But my point is that 
the Center for Range Voting in general is interested in the truth 
about election methods, not in what makes their favorite method look 
good. This distinguishes it from many other advocacy organizations 
and individual experts who will argue until doomsday that their 
favorite is best, and cannot be relied upon to give to us any 
contrary information.

I plan, as I can, to compare my results with what is on the cited 
page and to consider the implications in detail. However, that page 
does make quite understandable the statement that *in large 
elections*, Approval style strategy is optimal. Whether it is true or 
not remains to be seen, I'd want to look at IEVS simulations, which 
are far more accurate in comparison to these.