[EM] A simple thought experiment.

[Abd ul-Rahman Lomax]

At 02:52 PM 5/29/2013, David L Wetzell wrote:
>Also, the bottom line is that when you're advocating for a change in 
>which single-winner election rule alternative ought to be used, it's 
>not right to dump the burden of proof on IRV advocates.  The amount 
>of time spent marketing IRV already is a sunk cost and so the burden 
>of proof for switching ought to lie on the challengers not the 
>defenders of the status quo progressive electoral alternative to fptp.

Sunk cost for you, David. The rest of us are singularly unimpressed. 
We didn't ask you to spend that time and money. Voting systems 
scientists have been advising strongly against the method you adopted 
since the 19th century.

The voting system community, including *many* former IRV supporters 
and even FairVote activists, settled on a first voting system reform 
propoosal, not as the "ideal voting system," but as a do-no-harm 
improvement, Count All the Votes. I.e,. Approval Voting.

It will not fix all problems. But it costs almost nothing.

It has an obvious problem, but that problem only arises because, with 
it, voters who support a minor party will be able to express a vote 
for their favorite party, and all analysts agree that they will do 
this, it is strategically sound. Approval always allows voting for 
your favorite.

However, once voters can do this, they will *also* want to be able to 
express a preference for their favorite, which they cannot do in 
Approval where they choose to support, say, their minor party 
favorite and to cast a vote in the major election.

This is the problem that IRV solves. However, the problem was solved 
long ago, with a voting system that does not have IRV's serious 
malfunctions: Bucklin. It's ranked approval voting. It actually uses 
a truncated Range ballot, this has often been missed by analysts. A 
voter who has a strong preference can skip ranks to express it, 
causing the second preference vote to show up in a later round of 
canvassing. I call that "Limited Later-no-Harm protection." Voters 
will use this -- or bullet vote -- depending on preference strength, 
which is precisely how the system performs well in utility evaluations.

Bucklin was oversold, as was IRV recently, as a way to guarantee 
majorities. No voting system can do that except by restricting the 
freedom of the voter, in which case the majority is coerced or 
artificial. However, in contested public elections, Bucklin *did* 
find majorities even with many candidates on the ballot. Later, in 
party primary elections, with many candidates and bullet voting rates 
approaching 90%, it didn't find majorities. In that context, runoff 
voting makes *much more sense,* because what voters need is 
*information.* It's not about Later-no-Harm failure, an old 
speculation that FairVote enshrined as being The Reason why Bucklin 
didn't find majorities. And, given that, what would really have made 
sense would have been a Bucklin primary, with intelligent choice of 
runoff candidates if needed. And maybe a Bucklin runoff; with an 
advanced voting system, finding a optimal winner with three 
candidates should be possible.

Bucklin is *vastly* easier to canvass than IRV, it is just sums of votes.

So, David, sunk cost is also water under the bridge. What you have 
left is an organization with some established reputation. How you use 
that will determine if all the cost is truly sunk, or there is 
something that can be salvaged and used to build a brighter future.