[EM] piling on against IRV

[robert bristow-johnson]

On May 11, 2010, at 10:28 AM, Jameson Quinn wrote:

> Personally, I can say that I respected both of you before this  
> exchange, and that I respect each of you a little less after it.

well, i'm sorry you see such an equivalence.  i just will not tolerate  
it when my own words and position are repeatedly misrepresented.  it  
takes too much work to have to repeatedly set the record straight.

Jameson, if you *want* to, please look at the timeline a little to see  
when the "lie" word was first bantered about.  and also when it was  
first used by me.

> You got so caught up in name-calling, you couldn't see how close you  
> both were to common ground.
>
> There are three reasons a person might rank less-than-fully in IRV.  
> It might be a choice, akin to the choice to not vote in a second  
> round. In that case, the last round majority is in any common usage  
> a majority, even if it is less than 50% of the first-round votes. It  
> might be forced by the system, because of a limited number of voting  
> slots. In that case, anything under 50% of first round is NOT a  
> majority. Or it could be due to misunderstanding of the voting  
> system, or mismarking the ballot (say, equal rankings). In that  
> case, the usage is debatable, but I probably wouldn't call it a  
> majority.
>
> There is no way to tell the exact combination of these three  
> reasons. But if there is no limit on how many rankings, then I'd say  
> that using "majority" is arguable either way - not an out-and-out  
> lie, but arguably deceptive - and if there is a limit, then it's an  
> undefensible lie.
>
> There's room for honest disagreement here without name-calling.

there certainly is.

> Also, TTR theoretically gives almost always the same result as IRV,  
> but in reality the "comes from behind" scenario happens in a non- 
> negligible fraction of cases.
>
> One way to avoid getting caught in a back-and-forth gotcha-fest is  
> to restate your position succinctly, without a line-by-line response.

their position is that with IRV, some voters were involuntarily  
excluded (the term "disenfranchised" was used repeatedly in the  
appalling debate in Burlington).  they say this because because either

    1.  some people (it was less than 7% of those who voted in  
Burlington) did not participate in the final runoff and therefore did  
not participate in a "yeah" or "nay" regarding the candidate who was  
actually elected.  my response is that there is no evidence that this  
was involuntary.  there was no evidence that anyone, even little old  
ladies, did not understand that any candidate not ranked was not voted  
for.  some people didn't *like* the idea that they had to consider  
their contingency preferences on a single Election Day (which is, i  
believe where the real problem lied).  but this is directly comparable  
to not returning to the polls for the runoff in TTR, for which our  
experience in Burlington is that at least 45% fail to return, so the  
candidate is elected with an even smaller electorate.  to use their  
language, TTR "disenfranchises" 6+ times more voters than IRV did.   
but nobody was disenfranchised.  people who choose not to vote,  
whether it's at a different location on their ballot or at a different  
time, are not disenfranchised.
        about TTR, where IRV (or any other method that has no delayed  
runoff) is consistently superior to TTR is that the burden of  
returning to the polls for the runoff is removed, and election policy  
that reduces voter convenience reduces voter participation.  electing  
someone with reduced voter participation is less indicative of the  
will of the people and less democratic (the standard pro-IRV  
propaganda).  the problem with the pure FPTP is that, although there  
is no runoff, it can elect the wrong candidate (one without majority  
support if independent alternatives were removed), but the turnout is  
better than in the runoff.
    2.  their other claim for "disenfranchisement" is quite curious  
because it's about their 2nd-choice not counting in IRV because their  
1st-choice was not eliminated.  if their 1st-choice was left standing  
to the final round, they actually *did* participate in the "yeah" or  
"nay" choice regarding the candidate who would take office.  it's no  
less than what they could hope for with a traditional election.  their  
vote counted all the way.  what are the complaining about?
        but also, i recognize that, with what i call the "IRV kabuki  
dance of transferred votes", the failure to consider candidate support  
in the 2nd-choices was at the root for why the Condorcet winner (who  
is, in my opinion, the only candidate deserving of the title *THE*  
majority winner) did not advance to the final runoff (where he would  
beat anyone he meets there).  so that's where i agree that IRV sucks.   
but it's a failure in the method to properly discern candidate support  
from the electorate, not "disenfranchisement" or any violation of  
voters' rights.
        there is also the goofy situation of possibly *some* that had  
a 1st-choice vote that was eventually eliminated, and before it was  
eliminated, their 2nd-choice votes still didn't get to have any effect  
(maybe even before the final round, but that would be only that 7%).   
but that is again simply that the IRV kabuki dance of the transferred  
votes sucks.  nothing really new, but that has been hyped up by the  
IRV opponents with frothing mouths. it's not "disenfranchisement" or  
any violation of voters' rights.  it's pretty dishonest to frame it  
that way and they do it solely to hype up the rhetoric.

there are other complaints that Kathy has that i agree are flaws with  
IRV:
    3.  the lack of precinct-summability (this is Kathy's big deal).   
this is less a problem for a small venue like a small city, but really  
is unworkable for a statewide or nationwide election.  for a small  
venue, election integrity at a precinct level can be attained by each  
ward clerk handing a thumb drive (that has every ballot of that ward  
anonymously recorded) to every legally interested party and to the  
media.
    4.  non-monotonicity (this was also demonstrated, first by Warren,  
for the Burlington 2009 election).  i don't find non-monotonicity  
happening to the non-CW such a problem because the non-CW shouldn't be  
winning the election in the first place.

i can even state some of them better than she:
    5.  the transference of the burden of strategic voting from the  
majority (in Burlington, these were the liberals who didn't have to  
make a painful choice between the Dem and Prog candidate, which was  
the main reason IRV was adopted in the first place) to a minority (the  
GOP Prog-haters who found out that marking their favorite candidate as  
1st-choice kept their 2nd-choice, who happened to be the CW, out of  
the IRV final round resulting in their worst choice, the Prog, getting  
elected).  so when IRV proponents repeated the claim "you can vote for  
the candidate you really want without worrying about helping to elect  
the candidate you hate the most", that claim only applied to some  
folks, not to everyone.  i might call this "LNH-lite" because some  
people found out that they *did* harm their political interest by  
voting sincerely.
    6.  IIA.  the Burlington 2009 election *did* have a "spoiler- 
lite".  the "spoiler" was the Republican candidate who was also the  
Plurality winner and would come in at 3rd place, from a Condorcet  
POV.  we put "spoiler" in quotes because he was not a candidate with  
no chance of winning (like Nader in 2000).  but he was a loser whose  
presence in the race changed who the winner was.
    7.  lastly, but most importantly, both IRV or TTR risks thwarting  
the majority of the electorate.  this is the simple flaw in that it  
simply elects the wrong candidate.  this happens when there is a CW  
and the method elects someone else.  of course, with IRV, if the CW is  
not eliminated before the final round (i had thought that this would  
be nearly always the case, but we're 1 for 2 in Burlington), IRV will  
successfully elect the CW.  with the traditional ballot (FPTP and/or  
TTR), the CW can be missed (and the wrong candidate elected) and we  
wouldn't even know because not enough information was collected from  
the voters.  if there is no CW, it's really questionable if there even  
*is* a clear majority, and then it's a sorta crap shoot: Schulze is  
probably the best but Ranked Pairs is simpler to understand, and they  
elect the same candidate in the case of 3 left standing in the Smith  
set.  the point is that the return to the "traditional ballot" (in  
Burlington we have returned to a combination of FPTP and TTR with a  
40% threshold) made it worse, not better.
        that is the salient place where Kathy is just wrong.  now we  
risk electing the 3rd-most supported candidate (from a Condorcet POV)  
either by plurality (if he gets between 40% and 50%) or by TTR, with  
greatly reduced voter turnout at the runoff which is what our  
experience is in Burlington, and would be common at any other venue  
with TTR.  at least the worst risk with IRV was that it may elect the  
2nd most supported candidate when it fails to elect the CW.

> Then, even more succinctly, you can charitably restate the  
> couterarguments. The goal when restating the counterarguments is to  
> show your understanding of their strengths, not to undermine them by  
> showing their weaknesses. Last and briefest of all, you can give  
> counter-counter-arguments; these should essentially just point back  
> at the relevant bits of your initial statement.

naw, that's it.  it all falls under points 1 to 7.  whatever Kathy or  
Abd have to say, one can just point to one or more of the points 1 to  
7 to answer them.  there is nothing else to say about it.

all of these anomalies happened in the Burlington 2009 election which  
is why it is such a *perfect* case study about the failure of IRV to  
accomplish the very goals for which it was adopted.  there are  
important lessons for FairVote to learn here and it's frustrating for  
me to be standing between the obstinate denial of this failure at  
FairVote and the frothing mouths of the rabid IRV opponents that emit  
(besides spit) a lot of heat, but little light.  FairVote knows what  
is going on (they study it, and i consider Terry Bouricius to be a  
legit scholar of this topic, something that Kathy Dopp would never  
concede) but they're in denial and the rabid IRV opponents really  
don't have a clue.  like the Tea Baggers.  all they can tell is that  
something is wrong, but they don't know who to blame but that stop  
them from shooting anyway.  but FairVote is in total denial when they  
claim that the 2009 election in Burlington VT went off "without a  
hitch".  this dichotomy is appalling.

> If you're going to send someone to the killfile, don't tell them  
> when you've done it.

it's too late for that.

> Threatening it as a last resort might sometimes be productive,

the only product i wanted to create was to remove the temptation by me  
to peek at their response and then to feel compelled to reply.

> though you should avoid it if your dander is up; but saying "you're  
> in my killfile now" is about equivalent to "la la la la la I can't  
> hear you".

it is literally what i mean.  Kathy or Abd (who was kill-filed 4  
months ago) can send me email directly and i can hear that (but they  
better not abuse it or i'll become fully deaf to them).  but any post  
they put here to the EM list, i cannot hear.

> My opinion,

which is valued by me, even if you grant a moral equivalence between  
Kathy and me.

--

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."