[EM] Why the concept of "sincere" votes in Range is flawed.

[Abd ul-Rahman Lomax]

At 11:40 AM 12/2/2008, Jonathan Lundell wrote:
That's not really what an approval cutoff is. An approval cutoff is
>>used by some methods to denote "the candidates above are those I can
>>accept; those below, I really don't like". At least that's what I
>>understand, though some methods may reward strategic placement of
>>the cutoff as well.
>
>Abd's point, and mine, is that such interpretations of some "approval
>cutoff" isn't really justified, except perhaps as a shorthand way of
>describing how a voter *might* behave. The only instructions a voter
>is bound by the rules to follow are "vote for as many as you choose;
>the candidate with the most vote wins".

Thanks, Jonathan. Indeed, Approval minimally binds the voter. The 
voter may vote any pattern at all, and, with decent rules, can even 
vote "None of the above," by writing it in, thus contributing to 
majority failure explicitly, causing a runoff (again, with good 
rules. We often think only of public elections, as well, so, I'll 
note, Robert's Rules of Order, Newly Revised, requires an election to 
be repeated if there is majority failure, even if the voting is done 
with preferential voting. They are not taken in by the phony "last 
round majority" arguments. The last round of an IRV election is *not* 
very much like a real runoff, and results show that.)

Indeed, we wouldn't even need to provide the instruction described 
except that voters may otherwise assume that they are restricted to 
one vote. This is, purely and simply, voting with an arbitrary 
restriction removed.

We don't think of it that way -- and I think Brams didn't think of it 
that way -- because we are so accustomed to vote-for-one Plurality. 
Plurality makes very good sense in a direct democratic assembly, 
because the assembly doesn't automatically accept any election 
result, it votes on that as a separate question. (Sometimes the rules 
set this aside and consider an election automatically accepted if a 
majority was found. Obviously, in a Plurality election, if a majority 
has been found, and if people voted sincerely for their favorite, 
they have a Condorcet winner, pretty good for such a simple method! 
Majority requirements, unless the majority is forced, *must* find a 
true Condorcet winner with sincere voting.

But, of course, voters in Plurality don't vote with "full sincerity." 
In determining their favorite, many of them will discount 
possibilities that they consider considered unlikely. That's easy to 
fix, and, as I've pointed out, when voting is by show of hands, it's 
already fixed. When the name of your favorite is called out, voters 
would generally raise their hands. There is nothing stopping them 
from raising their hands for someone else. I have no idea how often 
this actually happens, but I'm sure it happens. It would be detected 
if the sum of votes exceeded the number of voters present, but, in 
fact, that often wouldn't show it, because of abstentions, and it 
might be thought that there was a counting error -- and usually it 
would be moot.

There is no rule in RRONR against voting for more than one. There is 
only advice to the clerk not to count such votes "because the 
intention of the voter is not clear." That's an assumption, not a rule.



>Assuming that the voter a preference ranking, the decision as to where
>to place the cutoff is inherently a strategic decision. Obviously I
>should vote for my favorite candidate. It's also obvious that if, for
>whatever reason, I vote for candidate X, I should vote for all the
>candidates that I prefer to X. What's not obvious is where to place
>the cutoff. Making that calculation optimally, especially in the light
>of imprecise polling, is difficult to impossible.
>----
>Election-Methods mailing list - see http://electorama.com/em for list info