[EM] Re: IRV letter

[Forest Simmons]

On Thu, 29 Apr 2004, wclark at xoom.org wrote:

> Adam Tarr wrote:
>
> > People don't like being told that they've put a ton of time and effort
> > into the wrong reform, and they need to step back from many of the
> > advances they've worked hard to get and start over.
>
> That's why I think it needs to be presented as an improvement, a step
> forward, and not a step back.

In one of the FairVote OR meetings I suggested that they add just one
simple rule to the IRV rules in their initiative (after all, their rules
deviated from standard IRV rules already by only allowing voters to rank
their top three choices): I suggested that they not eliminate any
candidate until they verified that at least one of the other candidates
was ranked ahead of the one about to be eliminated on more ballots than he
was ranked behind him.

Oregon, always known for progressive democracy (and as "Little Beirut" to
Bush), could then be a leader in election reform by having the best
version of IRV in the world.

Everybody looked to their leader for his opinion.  He expressed extreme
indignation at the suggestion, and said, "No."  And that was that.  Nobody
expressed any reasons for rejecting the idea.  I'm sure that nobody had
any reasons for rejecting the idea except that it didn't come from their
leaders or from the CVD, the only orthodox authority.

But if they did have any reasons they were discouraged from voicing them
by the antidemocratic climate of the meeting.

I appreciate your ecumenical efforts.  You are a good diplomat.  I haven't
given up entirely on the IRV folks either, but all of my efforts with them
are towards conversion; I won't volunteer to get signatures on IRV
initiatives.

I hate to offend anybody, especially anybody on this list, but in a sense
no method based on rankings or ratings is uniformly better than Plurality,
for the simple reason that Plurality uses a simpler ballot.

It is easy to argue that Approval is uniformly better than Plurality: if
everyone were to approve whomever they would have voted for under
Plurality as well as everybody they liked better, the result would be no
worse than under Plurality, and would often be better.  Any improvement on
this rudimentary strategy is just frosting on the cake.

Similarly, Candidate Proxy is arguably uniformly better than Plurality.

Both of these methods use ordinary Plurality ballots (with slightly
different instructions in the case of Approval).

I believe that the best Condorcet methods may just barely justify the use
of ranked Ballots in public elections.  IRV certainly doesn't.

For these reasons I have more success arguing against IRV on the basis of
its complicated nature (including lack of transparency and lack of
summability, but mostly just for its use of plain old ranked ballots)
relative to the two simplest methods that improve on Plurality.

I prefer Approval over Candidate Proxy, but most IRV supporters seem to
prefer Candidate Proxy over IRV because they prefer an approximation to
the strong FBC over complete compliance with the weak FBC. They think that
Candidate Proxy solves the spoiler / lesser evil problem in a more
satisfactory way than Approval does.

For those IRV supporters who think that ranked ballots are great and worth
the extra cost and complication, only for those do I introduce the
Condorcet alternative.

Best of Luck to you in your diplomatic efforts!

Forest