[EM] Why I think IRV isn't a serious alternative

[Abd ul-Rahman Lomax]

At 12:42 AM 11/26/2008, Kevin Venzke wrote:
>Hello,
>
>--- En date de : Mar 25.11.08, Abd ul-Rahman 
>Lomax <abd at lomaxdesign.com> a écrit :
> > If we must have a
> > single ballot, and a single winner, period, Range Voting is
> > actually a trick: it is the only relatively objective method
> > of assessing the expected voter satisfaction with an
> > outcome, turned into an election method. It's ideal
> > because it's designed that way. (The only fly in the
> > ointment is the charges about strategic voting, but I've
> > been arguing that this is based on a total misconception of
> > what we are doing when we vote.)
>
>I don't understand how you reconcile the two ideas here. Range is
>"objective" and "ideal because it's designed that way" based on the
>idea that voters have internal utilities and, if they vote them exactly,
>under Range voting, the best candidate according to overall utility
>will be elected every time.

The "only relatively objective method" is not 
Range Voting, itself, it is the study of Bayesian 
regret using reasonably simulated internal, 
absolute utilities (To be most accurate, it would 
use the "HH" scale, "Heaven-Hell," which does 
make the assumption that the absolute worst 
possible outcome and the absolute best possible 
outcome, with an effectively unlimited universe 
of candidate, have equal values for all voters.) 
in simulated elections. From these utilities, the 
best candidate in a simulated election. The value 
for each candidate is determined by some 
reasonable algorithm, either by relatively simple 
assumptions, or by the positions of the 
candidates in an n-dimensional issue space, or 
some other means. To take this further, 
simulations would be guided by actual voter 
preferences and intensities from polls.

That is, the performance of an election method 
may depend on which kind of preference profiles 
exist among the voters, not only upon the method 
or how the voters decide to vote.

 From the voter utility profiles, a preference 
list is easily constructed. Actual voter 
behavior, then, can be simulated, using various 
kinds of voting strategy, from "fully sincere" in 
Range, to maximally strategic, designed to 
maximize personal voter outcome. That's tricker, 
to be sure. However, within the simulation, 
"maximally strategic" votes take place when the 
voter "has accurate knowledge of the preferences 
of others," and thus their likely votes.

Once this is done, with a large number of 
elections, the deviation of the actual method 
results, under various voter strategy profiles, 
from the ideal result is determined, as a 
"regret" value. The best method, when comparing 
two methods, is considered to be the one with the lowest average regret.

"Ideal" assumes maximizing overall utility, i.e., 
the sum of the individual HH utilities, which are 
both absolute and, we assume, commensurable (we 
would want to minimize the number of voters 
tossed into Hell, or close to it, and maximize 
the number who experience a Heavenly result; 
however, so far, only the sum is considered, to 
my knowledge, in the work that has been done. In 
real elections, it's possible that the ideal 
winner would minimize the Hell results until the 
worst utilities are above some level, this is 
what true, functional organizations do, because 
they value group unity and every time someone is 
totally shut out of a decision, their position 
neglected, the functional group becomes smaller. 
Over many elections, this can have a major impact 
on the success of the organization. Tyranny of 
the majority is highly dysfunctional. However, 
such is the state of the art, to my knowledge, of election simulation.)

(Note: this is *my idea* of Warren did or should 
have done. What he *actually* did may differ in 
some respects. Perhaps someone will point us to a 
more accurate description. The approach is as I 
described, *relatively* objective. If I'm 
correct, Warren settled on Range Voting as a 
result of his utility studies, not the reverse. 
But, again, he or someone else could correct that.)

Prior to the use of this method, voting systems 
were compared using voting criteria, and systems 
either pass a given criterion or they don't. And 
no method satisfies all proposed election 
criteria, and, I've elsewhere argued, some 
criteria commonly considered reasonable and 
important *must* be violated in order for the 
method to produce results that we would 
rationally -- and following actual practice in 
small democratic organizations with access, 
because of the scale, to much more sophisticated 
decision-making methods than are possible with a single ballot.

This is the "relatively objective method of 
assessing" election outcome. When it's easy to 
determine, in a real situation, the absolute 
individual voter utilities, "fully sincere Range 
Voting" implements it as a method. That is, if 
the voters are honest, or if their "votes" are 
determined for them by some objective method -- 
such as a measurement of tax impact based on, 
say, the previous year's income tax return -- 
this obviously would produce an objective result 
that could be considered ideal. In real 
elections, of course, determining absolute, 
commensurable utilities may not be possible. 
(There are voting systems involving lotteries and 
real bets made by voters that should encourage 
the voting of absolute utilities, but these aren't being considered here.)

However, I stated "average voter satisfaction." 
This represents normalization. It assumes that 
the full satisfaction of all voters is equated, 
and likewise the minimal satisfaction, given a 
particular candidate set. This produces 
normalized utilities. Because of the simulations, 
which assume a full absolute utility profile, we 
can then determine reasonable "fully sincere 
votes." We can then sum these to determine a 
somewhat different optimal winner, and assess 
regret from actual outcomes based on that.

In real elections, voter behavior will deviate 
from those "fully sincere" votes. (Fully sincere 
means disclosing true preference strength, within 
the resolution of the method.) It deviates from 
it for two reasons: (1) voters don't care to put 
that much effort into rating, it's easier to 
rank, generally, because it only involves 
pairwise comparisons; however, voters usually 
only have meaningful preferences between a few of 
the candidate pairs involved. And (2) voters are 
accustomed to elections being a choice, or a set 
of choices, and choices are normally made within 
a context that considers outcome probabilities 
where choice power is used to choose between 
realistic possibilities, not merely to compare the value of each outcome.

So: How does Range, with realistic voting 
patterns, compare with other methods. Range does 
*not* produce zero regret. It produces relatively 
low regret. If fully sincere voting could be 
somehow guaranteed -- probably impossible -- it 
would always choose the ideal winner (within 
certain restrictions, basically normalization). 
So there are two deviations from the ideal. The 
first is from normalization, and the second is from strategic voting.

Range *with strategic voting* is better with 
respect to regret than any other method that has 
been simulated, to my knowledge. There is an 
exception: Top Two Runoff Range Voting beats 
Range. That's not surprising. It would detect and 
fix some of the deviations due to normalization and strategic voting.

Now, to prevent the advantage that knowledgeable 
voters would have from being able to vote 
accurately, should we damage the outcome averaged over all voters?

This is the implication of the argument that we 
should prevent "strategic voting" as it applies to Range.

To be continued.