[EM] RCV Challenge

[robert bristow-johnson]

Hay guys, my paper titled

    "The Failure of Instant Runoff Voting to accomplish 
        the very purpose for which it was adopted:
         An object lesson in Burlington Vermont"

has gotten revised, I think, for the last time.  I added a little color to it with 6 figures that illustrate the Center Squeeze using specifically the Burlington 2009 numbers (they ain't just hypothetical numbers).  It's 15 pages now.

Anyway, if you are so inclined, please take a look at it: https://tinyurl.com/2tety9tj .

There are at least two conventions that the scholarship, debaters, and psephologists practice that I kinda reject.  (I do not reserve "STV" to mean multi-winner elections, to me "STV" means a method.  And I really hate the square pairwise-defeat matrix that people use.  That is really stupid because the number in row R and column C has meaning only in comparison to the number in row C and column R.  You should just put those two numbers together as I do in Table 3.)  And I use more pedestrian labels for important principles and properties.  (Like instead of "IIA", I call it "Avoid the Spoiler effect" or instead of "Condorcet criterion", I call it "Majority rule". And, I know that the root of the phrase "One person, one vote" has to do with redistricting and representation, but the term well applies for the principle of the equality of the votes of the enfranchised.  And I consider "strategic voting" to be a little different than "tactical voting".)  The original intended audience are voters, media, activists, and legislators in Vermont.  I want to educate them that FairVote does not get to define what "RCV" means and what really went wrong in 2009.

But otherwise, I am happy to hear or read critique. But if you're gonna critique my use of "STV" or "One person, one vote" or "Majority rule", I have already had that argument.

I just want Ranked-Choice Voting to be what it is meant to be.

> On 12/28/2021 9:55 PM Forest Simmons <forest.simmons21 at gmail.com> wrote:
> 
> 
> I agree with Kristofer .... as he said, "... we only get one shot ..."
> 

And I do, too.  Even though, in reality, we in Burlington Vermont are getting a second shot at this, which is why I am so dismayed at FairVote's and RCV Resources' refusal to learn anything and just repeat the same mistake in the same jurisdiction where the Hare RCV screwup was the worst and the method promptly repealed.  They are depending on people having a short memory or just a poor understanding of the problem.

> So why take any unnecessary chances?
> 

Exactly.  What would happen in Burlington if we re-adopted the same Hare RCV, then eventually had another failure like we had in 2009?  Would Rob Richie and Terry Bouricius just write it off again and call it another "... Highly Successful IRV Election."  That kind of arrogance and naked denial makes me angry (it's so very Trump-like) and just more committed to opposing them.

> And truly the risk is unnecessary, because there are many perfectly good manipulation resistant methods that are uniformly better and much simpler than the 2nd rate kluges that, out of ignorance, are routinely proposed.
> 

But, the only risk with Condorcet is what to do with a cycle.  Even implicit methods that do not need an explicit "Condorcet completion" procedure can be expressed as a two-method system: First see by the simplest method if there is a Condorcet winner, then if not, do something else (the "Condorcet completion" procedure).

The "do something else" else will only happen under very rare circumstances.  Of the 440 RCV elections analyzed by FairVote, *every* *one* had a Condorcet winner and all but one elected that Condorcet winner.

In my opinion, what is more important regarding "do something else" is that the exception handling be easily understood *and* *accepted* by the public, so that it does not appear that the method is sneakily trying to game or throw or steal an election.

> We need to educate the teachable, while advocating only for the best possible methods.
> 

What's "best"?  Is simplicity and transparency "best"?  Or is mathematical resistance to strategy and tactic what is best?

Now, regarding the latter, this is not best, but regarding the former, I believe that the "best" contingency procedure in case of a cycle is to simply elect the plurality winner.  People will understand that there was no "Consistent Majority Candidate" (a neologism of my creation, it's better than "Beats-All Candidate") and then we ask them "In that case, which candidate appears to have the greatest voter support?"  Ostensibly, it would be the plurality winner.

> Let the reactionary pooh-bahs defend their pet recycled 2nd rate methods ... leave to them the opportunity to reveal their own ignorance and arrogant disregard for what we has been learned in the last 25 years (thanks to Rob Lanphier's cultivation of nitty-gritty election science) ... about election methods in general and Condorcet methods in particular.
> 

Yea!  I'm on that side.  Better include the real-life example of Burlington 2009 for what happens when IRV fails to elect the Condorcet winner.  That's an object lesson.  Too bad FairVote and some Progressives in Burlington refuse to learn the object lesson.

L8r,

r b-j

> 
> El mar., 28 de dic. de 2021 1:55 a. m., Kristofer Munsterhjelm <km_elmet at t-online.de> escribió:
> > On 28.12.2021 06:56, Forest Simmons wrote:
> >  > Robert,
> >  > 
> >  > You wrote ...
> >  > 
> >  > "I'm just a Condorcet guy. How cycles get resolved is less motivating
> >  > to me than insuring that the Condorcet winner is always elected"
> >  > 
> >  > But "how cycles get resolved" has a big influence on whether or not they
> >  > come into existence. And when these cycles are created, the sincere
> >  > Condorcet Winner goes out the window ... no longer showing up as a
> >  > "beats-all" candidate on the ballot set. Can the cycle resolution method
> >  > reconstruct the lost CW? All bets are off.
> >  > 
> >  > The easier it is to game the cycle resolution method, the more incentive
> >  > for the gamers to create the cycles.
> >  > 
> >  > It seems that most Condorcet cycles are artificially created.
> >  > 
> >  > For example, one way to create a cycle is by an insincere rank reversal
> >  > technique called "burial". 
> >  > 
> >  > If the cycle resolution method has no built in disincentive/negative
> >  > feedback for this (or any other) kind of cycle creation, over time the
> >  > gamers find out that they can manipulate elections to their advantage
> >  > with impunity, so that artificially created cycles become more and more
> >  > common, giving the false impression that cycles are a normal fact of nature.
> >  
> >  There's still a question of just how much coordinated strategy will
> >  happen. If the voters are mostly honest, then insisting on strategic
> >  resistance will only be giving up honest performance for nothing
> >  important in return. On the other hand, something like Borda obviously
> >  collapses because it can't handle *any* kind of strategy.
> >  
> >  I suspect that the initial amount of strategy depends on the voters -
> >  that some voters are more inclined to strategize than others. Probably
> >  voters who are used to FPTP would be more strategically inclined,
> >  although I don't have any proof of this (it just seems intuitive). But
> >  Debian used Schulze however many years without having any trouble with
> >  burial (that they could detect, at least), and most places that use STV
> >  nowadays don't seem to have much of a problem with vote management even
> >  though it was pretty common in New York and some Canadian elections that
> >  used to use STV.
> >  
> >  In the face of this uncertainty, it's reasonable to want to have a
> >  method that resists strategy well so that it doesn't get repealed after
> >  a disastrous result. But that doesn't mean that it's strictly necessary
> >  - only that we can't tell if it is, and we only get one shot.
> >  
> >  -km
> >

--

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

"Imagination is more important than knowledge."

.
.
.