[EM] IRV inconsistency

[Forest Simmons]

On Wed, 16 May 2001, Markus Schulze wrote:

> Dear Forest,
> 

<snip>

> 
> You wrote (15 May 2001):
> > As Richard says, IRV ignores and trashes valuable information
> > willy nilly and still pretends to come up with a top notch
> > winner. All of IRV's shortcomings can be traced back to this
> > wasting of information.
> 
> It is not feasible to say that the one election method uses
> more information than the other. All election methods use the
> same information. They only interpret it differently. Election
> methods only differ in how important they consider which part
> of this information.
> 

Although many methods have the same available information they obviously
don't all use it to the same degree, for example random candidate doesn't
use any of the available information (except who the candidates are), and
random ballot doesn't use as much as most other methods. 

IRV looks only at first choices before making an irrevocable decision of
whom to eliminate first.  After the elimination is accomplished all of the
comparisons relative to the eliminated candidate are irretrievably lost. 

This would be of no consequence if IRV both satisfied the IIAC and made a
careful consideration of whom to eliminate in the first round. But it does
neither of those.

And yet IRV proponents claim that IRV picks "THE CANDIDATE PREFERRED BY A
MAJORITY." 

Here's the context for this last remark:

I recently attended a meeting of FAVOR, an Oregon group in their final
stages of putting an IRV initiative on the ballot.

The first 45 minutes of their meeting was devoted to agonizing over the
wording that would be placed in the voter pamphlet and the ballot summary.

The two main issues that they wanted to put across to the voters were:

(1) IRV is a "one man, one vote" system because it merely takes your one
vote and shuffles it around until it puts it into a position where it will
count.

(2) IRV is a system where "the majority" decides the winner, while
our current plurality system is not.

Here are exact quotes (including upper case letters) from the proposed
language:

ESTABLISHES AN INSTANT RUNOFF VOTING PROCEDURE IN WHICH CANDIDATE WITH A
MAJORITY OF VOTES WINS

ESTABLISHES AN INSTANT RUNOFF VOTING PROCEDURE TO ELECT THE CANDIDATE WITH
A MAJORITY OF VOTES

ESTABLISHES INSTANT RUNOFF VOTING PROCEDURE TO ELECT THE CANDIDATE
PREFERRED BY A MAJORITY OF VOTERS

After much (sickening) deliberation they decided to go with the last of
these. They didn't like my suggestion to make it a little more honest by
either changing "the" to "a" or else appending the phrase "in the last
round."

Another section of the ballot describes the change by contrasting the
status quo with the proposed change. In this section part of the language
went thusly: 

[Under IRV] "the candidate with a majority of votes wins", [while under
the current system] "the candidate with the highest number of votes wins." 

There was much discussion about how voters would react to this because in
the minds of most voters if a candidate got a majority of votes, that same
candidate would be the one with the highest number of votes.

I held my tongue and watched them agonize and squirm over this. This was a
big dilemma, because in order to clarify this in the minds of the voters
they would have to spell out the kind of "majority" that was being
referred to.

They finally decided to leave the wording as it was rather than confuse
the voters with the facts.

Later (via email), I suggested to the leaders of the group that they
replace "majority of votes" with "majority of transferred votes" to
clarify the issue for the voters.  I said, "After all, we're not trying to
trick the voters into supporting IRV, are we?" 

(also via email) I remarked that Condorcet method supporters reserved the
phrase "majority winner" for a candidate with more than 50% of the first
place votes, even though a Condorcet winner would qualify for that title
under IRV usage, if fact more so than the IRV winner since a CW has to be
preferred over every candidate (including the IRV winner) by a majority,
not just over one other candidate.

I suggested that since they were so anxious to give the voters the
impression that the main advantage of IRV over the status quo was that it
relied on majority preference, why not supercharge IRV by attaching a
majority check on each elimination to remove all doubt about
(Supercharged) IRV's majoritarian credentials. 

I explained that this could be done at no extra cost by doing a pairwise
comparison of the candidate about to be eliminated with the one next in
line. After all, doesn't each candidate deserve a majority hearing before
being summarily dismissed on superficial, circumstantial evidence?

I gave an example showing the advantage of the majority check:

1  ABCDE
2  BACDE
4  CABDE
8  DABCE
14 EABCD

Without majority check, D wins (by a pretty half hearted majority) against
E, with whom he shares all of the last place votes. 

With majority check, A wins.

I pointed out that this majority safety check would never allow IRV to
inadvertently eliminate the Condorcet winner. 

All to no avail. In spite of all the "majority" rhetoric, they were
unwilling to consider letting a majority check each elimination. 

I explained that I knew they were under the wire for deadlines, so that it
was too late to make the change this time around, but to consider it as a
suggestion for the future.

The leader emailed me back to tell me that he found my suggestion to be
highly insulting.

He went on to say that since all voting methods are flawed, it is a
subjective matter anyway, and that he preferred the IRV winner over the
Condorcet winner.

I thought of explaining to him that just because no perpetual motion
machine exists doesn't mean that any car is as good as another, and
that any choice between cars is purely a subjective matter, so a Nova is
as good as a BMW.

That is where the comparison between Arrow's Impossibility theorem and the
Second Law of Thermodynamics came to mind, and IRV wasting information
like a junky engine wasting energy.

To paraphrase Markus we could say, "All gasoline engines with the same
size gas tank use the same amount of energy."  But it would be more
accurate to say, "All such engines have the same available energy, but
some make actual use of more of it, while others send unburned fuel out
the exhaust pipe."

Even engines that oxidize all of the fuel in the combustion chamber don't
all convert it to mechanical energy with the same efficiency. 

Even methods that consider all of the available information before
discarding it don't find the same quality winner. 

But an engine that sends unburned fuel out the tail pipe has no chance of
approaching the natural limits of efficiency imposed by the second law of
thermodynamics.

So IRV has no chance of approaching the limits imposed by Arrow's (and
others') impossibility theorems. 

Peace and Understanding to All,

Forest