[EM] A rant about IRV and it's recent history in Alaska.

[robert bristow-johnson]

> On 12/05/2024 12:38 AM EST Michael Garman <michael.garman at rankthevote.us> wrote:
> 
> Because rb-j is an intellectually dishonest crank who doesn’t want to be confused by facts after he’s made up his mind
> 
> 
> On Thu, Dec 5, 2024 at 12:35 AM Greg Dennis <greg.dennis at voterchoicema.org> wrote:
> > Why leave out the November 2022 election where Peltola was the Condorcet winner against Begich?
> > 

Well, I was waiting the better part of a day to hear from John how my "original statement, is something of an oversimplification."  More often the accusation is that I am unnecessarily complicating matters.

Unraveling a knot that someone else had tied can be complicated.  But the principle is extremely simple:

1. In a modern and righteous democracy, human beings with the same status, that is enfranchised citizens and of voting age, must be treated equally, otherwise we have a form of apartheid that explicitly treats persons of some class worse than others.  Incumbent to that official equality, our votes must count equally.  If our votes don't count equally, then I want my vote to count more than yours.

2. In a single winner election, where there is no proportionality to be had, for our votes to count equally, the only way to make that happen is for the majority to rule in elections.  For that to happen, then if more voters mark their ballots preferring Candidate A over Candidate B than the number of voters marking their ballots to the contrary, then Candidate B is not elected.  If Candidate B were to be elected, that would mean that the fewer voters preferring Candidate B had cast votes that had greater value and counted more than those votes from the greater number of voters preferring Candidate A.

3. A majority must mean more than half of a set.  A "simple majority" is more than half of votes cast, excluding abstentions.  When there are 3 or more candidates, there might not be a majority, but between 2 candidates there is always a simple majority unless they tie.  This simple fact is often misconstrued by IRV advocates that IRV elections "guarantee a majority winner" because they boil the field of candidates in an election down to two candidates in which, between only those two, there is always a simple majority.

Indeed, in Greg's website, https://voterchoicema.org/ , it says right on the front page: "Ranked-Choice Voting - This commonsense change gives Massachusetts voters a stronger voice when we cast our ballots, and guarantee that our elected leaders are supported by a true majority."

That's an objectively false statement.  A technical falsehood.  Right on the outside on the product packaging.  A blatant falsehood.  Yet I am the one who is accused of being "intellectually dishonest".

To be honest, we gotta compare apples to apples.  If RCV is claiming an advantage over FPTP, we have to have the same semantic rule applying to both methods to compare them honestly.  We can't move the goalpost on RCV from its position for FPTP.  And One-Person-One-Vote (or the equality of our vote) must mean that we're counting persons.  Voters.  Bodies.  (Not marks on a bunch of ballots.)  The percentage of the vote that the winner gets is:

(number of voters for the winner)/(total number of non-abstaining voters)

If that ratio is more than half, the winner got a majority of the vote.  If that ratio is not more than 50%, the winner did not get a majority.  Not any kind of majority.

With FPTP, suppose we have an election with 3 candidates and 100 voters.  45 voted for Candidate A, 40 for Candidate B, and 15 for Candidate C.  We don't allow FPTP advocates to claim that A got *any* kind of majority.  Candidate A got a plurality, 45%, of the vote.  Because more voters voted for a loser in the race than the number of voters voting for the winner, the winner did not get a majority.

We don't let Candidate A claim that they got a majority of the vote by limiting "the vote" to be the 85 voters who voted for either A or B and excluding the 15 C voters from the denominator.

To be intellectually honest, you have to accept the same criterion for "majority support" with RCV.  Otherwise you're moving the goalposts from the position you placed for FPTP to an easier position for RCV.

____________________________________________________


To make this crystal clear: you have an election that uses some given election method.  You have a total number of V voters all casting ballots.  There are several different races on the ballot and not every voter will vote in every race.

Now focus on a particular single-winner race.  That set of V voters can be divided into 3 groups:

1. Voters that the method counts as voting for the winner.
2. Voters that voted for a candidate, but the method does not count them as voting for the winner.
3. Voters that abstained from voting for any candidate in that particular race.

Every voter that cast a ballot is in exactly one of those disjoint groups.  Now, the voters in group 1 are considered happy voters.  The voters in group 2 are considered unhappy voters.  And the voters in group 3 don't give a shit.

Now, we will say that it is equivalent that a voter stayed home and didn't vote at all to the case that this voter came in and cast a completely blank ballot.  Let's say that those two cases are to be considered equivalent.

We will also say that what a voter did in some other race has no bearing on what the voter did in the race under consideration.  So that if a voter abstained in a particular race because they cast a completely blank ballot is, within the scope of that race, equivalent to that they abstained in that race but perhaps voted in some other races.  It should not make any difference to how we count the votes for this particular race under consideration.

Then this means "majority" is "simple majority".  More than half of the votes cast and not including abstentions.

The percentage of the vote that a winning candidate got is the ratio of the voters in group 1 divided by the sum of voters in group 1 and group 2.  If that ratio exceeds 50%, then the winner got a simple majority the vote.  If that ration does not exceed 50%, then the "elected leader" was NOT "supported by a true majority."

Now do you see how RCV advocates like FairVote or their satellites (like Better Ballot [your state here]) are not telling the truth when they say that "For a candidate to win in an RCV election, they must get over 50% of the vote."  Do you see how that is a falsehood?  I used Burlington 2009 as a concrete example.

Because in a 3 or more candidate race, "the vote" is the union of the two sets: everyone who voted for the winner and everyone who voted for a loser.  The number of elements of the first set goes into the numerator and the sum of the numbers of the two sets goes into the denominator.  But FairVote and friends want to claim that it is a smaller subset that goes into the denominator.  They exclude the "exhausted ballots" from the count going into the denominator.  But we don't do that for FPTP, why should we do it for RCV or any other method?

If we applied this RCV reasoning to FPTP, then even FPTP can guarantee that the plurality winner got a majority of the vote.  Given the subset of voters that voted for one of the top two vote getters, then the winner got more than 50% of the vote in that subset of voters.  Big fat hairy deal.

If we don't let FPTP get away with that claim, why should we let Hare RCV (or Condorcet RCV or any other method) get away with that claim? 

This leads to the concise result that: If more voters voted for any loser in a race than the number of voters who voted for the winner, that winner did not get a majority of the vote, by any voting method.  Not any kind of "majority".

--

r b-j . _ . _ . _ . _ rbj at audioimagination.com

"Imagination is more important than knowledge."

.
.
.