[EM] National Popular Vote & Condorcet

[Paul Kislanko]

The idea is a good one, but there's no practical way to make it work.

If all you have is plurality counts, you don't have enough information to
retrieve the voters' Condorcet pairwise preferences. If what you mean is use
States' plurality results to form a condorcet matrix, you get even worse
than what we have now since that results in each state getting a condorcet
vote regardless of the number of voters in each state.

I'm sure that's not what you meant, but this allows me to re-introduce
something I suggested several years ago. If somehow we could define an
"election method" as a combination of "vote collection process" and "vote
counting process" we could select a collection method that would support any
known counting process.

One can infer a plurality ballot from any kind of ranked ballot, but not the
other way around. 

One can infer an approval ballot from any kind of ranked ballot that allows
equal ranks, but not the other way around.

"Any kind of ranked ballot" includes a range ballot, since an ordinal
ranking (allowing equal ranks) can be obtained by sorting the alternatives
on each ballot by their values. One can NOT retrieve the range from an
ordinal ranking derived from the voter's range ballot, but that is only a
problem if the global counting is range. If that's the case all states must
use range ballots and there's no problem.

If every ballot is a ranked ballot, we can choose a COUNTING process that is
plurality, approval, or Condorcet. We can even (and probably should) use a
Condorcet COUNTING process to process range ballots, instead of just summing
values over altertanitives.

But the most general COLLECTION process would be a "Condoret ballot", where
the voter is presented EACH pair of alternatives and asked to pick one. For
instance, in the {A B C} case instead of having to pick ONE from:
 A>B>C
 A>C>B
 B>A>C
 B>C>A
 C>A>B
 C>B>A
But if I got to choose my own pairwise preferences: 
 A > B or B > A
 A > C or C > A
 B > C or C > B
I might have come up with A>B B>C and C>A, which can't be put into a ranked
ballot because you can't have A>B>C>A, but that was my sincere Condorcet
vote. There's no way you can use Condercet COUNTING to find voter
preferences if you don't use a Connoret BALLOT to collect all the votes.

Which means nothing except to count popular vote across all states you have
to use the LEAST representative BALLOT representation across ALL states,
which is plurality (sigh.)

FWIW there's a movement gaining traction for each State to assign its
electors to the winner of the NATIONAL popular vote. Bad as it may be,
Plurality may save us yet.