[EM] IRV in Burlington VT

[robert bristow-johnson]

On Jan 9, 2010, at 4:28 PM, Warren Smith wrote:

> See this report on the Burlington 2009 IRV pathologies:
>
> http://rangevoting.org/Burlington.html
>
> one of the co-authors, Anthony Gierzynski,
> is a UVM professor who lives in Burlington.
>

Prof. Gierzynski *works* in Burlington, but lives in Montpelier.  i  
am not sure where he was living last March 3, but i doubt that his  
stake in this election is direct.  his stake is professional and  
scholarly.  but, as a non-resident, he doesn't have the same stake  
that either i or Terry have.

> The report refutes a lot of lies commonly told about IRV,
> and concludes that this election
> "singled out IRV & plurality as nearly-uniquely bad performers  
> [among all
> commonly-proposed election methods]."

i might even agree with that


> Given this experience, I would suggest that Burlington switch to a
> method different from plurality and different from IRV.   Some obvious
> contenders are approval and range voting.

Warren, why is Condorcet not included?

the problem with Approval voting is that it requires too little  
information from the voters.
and the problem with Range voting is that it requires too much.

if you support a particular candidate (let's say the Dem) but you  
approve of two (say both the Dem and the Prog), you have a problem  
with voting strategy with Approval.  maybe, depending on how other  
folks vote, you can help your favorite against the candidate you  
tepidly support, so you punch "Approve" for your fav only.  or, maybe  
your fav is actually not in really in the running and the election  
becomes a contest between your fallback candidate and the guy you  
think is a piece of crap.  then "what to do, what to do?? oh me oh  
my!" (the savvy voter must thing strategically.)  most people will  
just approve the single candidate they support and, if nearly all do  
that, there is no difference between Approval and Plurality.

with Range voting, again if you support a particular candidate but  
you sorta approve of two, you have a problem with voting strategy  
again.  depending on what other people are doing, it may be best (for  
your interests) to plop all of your points onto your fav.  but what  
if it becomes a contest between your fallback and the piece of crap  
candidate?  again "what to do, what to do?? oh me oh my!"

the Ranked-Order Ballot extracts the correct amount of information  
from the voter.  if the voter rankes A>B>C>D, the ballot makes no  
assumption regarding how much more the voter likes A over B compared  
to how much the voter likes B>C.  maybe the voter likes A and B  
almost equally (but A just a little more) but thinks C is a piece of  
crap.  or maybe both B and C are crappy, but B is a little more  
tolerable than C.  the problem with Borda is that it assumes the  
difference in preference are the same between adjacently ranked  
candidates (an assumption not supported or refuted by the voter or  
his ballot).

the ranked-order ballot only asks the voter: if the election were  
solely between A and B (or any other pair), who would the voter vote  
for?  the problem with IRV is that it doesn't decide the election by  
viewing the ballot information as such. instead IRV and the single  
transferrable vote method looks at a voter's vote as a commodity to  
be transferred from one preference to another based on the arbitrary  
rule that only 1st-choice votes count when determining who the  
biggest loser is and eliminating that "loser".  2nd-choice votes are  
no better than last place in that evaluation.  i am a DSP algorithm  
engineer; one of the things we hate the most (and are sometime  
unavoidable) are arbitrarily-set thresholds, we know that bad shit  
happens when the numbers just don't quite get to the threshold.   
immediately, in 2005, when i read the IRV law i was signing, i  
recognized that as a problem, but hoped it would elect the Condorcet  
winner anyway (and it did in 2006).

even though there is the *possibility* of a cycle (and thus a  
*possible* strategy of some to try to throw a Condorcet election into  
a cycle) the likelyhood is soooo low, because of political alignment  
along the major axis of the political spectrum.  Nader voters in 2000  
were not likely to choose Bush for their 2nd-choice over Gore.  even  
if voting strategists succeed in throwing a Condorcet election into a  
cycle, they have no solid control in advance about how that cycle  
will be resolved.  sounds like a dangerous "strategy" and i doubt  
anyone will try it or be successful doing so in reality.

> The difficulty is, that due to the fact that the world in
> general and Burlington in particular, is inhabited by morons,

some of us are morons, that is true.

> the ballot issue probably is going to consist of exactly of the two
> worst choices "(a) IRV or (b) go back to plurality?"
> with no third choice being offered.

no shit.  that's what i've been crying in the wilderness since last  
March.  lately, where i've been crying is:
http://7d.blogs.com/blurt/2009/12/burlington-residents-seek-repeal-of- 
instant-runoff-voting/comments/page/2/#comments .

> Furthermore, without a decent voting system,
> even HAVING an election with 3 or more choices, is rendered dubious  
> and risky.

no shit.  but, at least with IRV, we are not risking electing the  
slightly unpopular GOP candidate in this town of liberals.  with the  
old law (which requires a 40%+ plurality or it goes to runoff), there  
is a greater risk of electing the unpopular GOP candidate, which is  
why the GOPers are the biggest supporters of the repeal question.

> Ludicrous, isn't it?

about as ludi as the Electoral College.

--

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."