[EM] Brief Comment on IRV debate

[robert bristow-johnson]

still haven't found the time to do another treatise on this thing.

On 1/21/12 2:47 PM, David L Wetzell wrote:
>
>
> On Fri, Jan 20, 2012 at 3:59 PM, Jameson Quinn 
> <jameson.quinn at gmail.com <mailto:jameson.quinn at gmail.com>> wrote:
>
>
>         Like I wrote, the only way a non-CW can win w. IRV is if the
>         two biggest parties do not center themselves around the
>         center.  This possibility is what will goad them to recenter
>         themselves more often.  That is what would have happened in
>         Burlington if the anti-IRV campaign had not succeeded.
>
>
>     1. This sounds to me like: "Sorry about your house. And your
>     neighbors'. But actually, it's kind of good this happened. It will
>     teach people in the future to leave enough space between their
>     houses for us to fire these test missiles. And that way, they'll
>     have room for bigger gardens." That is, a disaster is excused as
>     good because it will encourage actions that will prevent later
>     disasters and also have other benefits. To me, a disaster is a
>     disaster.
>

that's a tautology, so it's pretty hard to argue with.  what's not a 
tautology is equating electing the non-CW IRV winner to office to a 
missile disaster or burning houses.

the election of the non-CW in the IRV election in Burlington in 2009 is 
an "anomaly" and is indicative of a "pathology".  "disaster" is too 
strong of a description.  because of some local political issues (this 
Prog mayor had propped up a city-owned telecommunications company with 
$17 million from the city's cash pool and there is an additional $33 
million owed to Citibank or Citigroup) some had been calling it a 
"disaster" and was one of the bullshit reasons for opposing IRV.

> dlw: There is no good reason to call the election of the Prog mayor in 
> Burlington with IRV a disaster.  It is an outcome whose effect would 
> have been to move the status quo to relocate the two biggest parties 
> around the true political center for Burlington.  If it was a disaster 
> then you'd think it would have been shot-down by more than just a 
> little over 50% after a very deceptive campaign funded by the 
> defenders of the status quo.
>
> Why are you giving oxygen to the tripe peddled by the lovers of FPTP 
> or Democracy In Name Only(DINO)?

because the "IRV happy talk" was lacking of credibility.  anyone who 
denied that there was any problem with the 2009 election in Burlington 
quickly lost credibility in the argument.  that is why i am mad at that 
pro-IRV side in the 2010 repeal election even though i voted with them 
on the repeal question.

>
>     2. Exactly who is supposed to move the center? First you have D vs
>     R. Then the center moves left. D is quite happy with their
>     permanent victories, so they won't move left; and so R has no room
>     to. Then discontented leftists start P and pass R. The rational
>     strategic response would be for D to move so far right that it
>     crushes R and makes it into the final round, so that D and P are
>     the new major parties. But center squeeze makes that hard; and
>     also human nature makes it hard to respond to a new interloper
>     from your left by moving right. It might also work R to move left,
>     ignoring D, and hope that when D and R really are clones then D
>     will be circumstantially a bit behind and so be eliminated first,
>     so that R and P will be the new major parties. But that means
>     embracing a 50/50 gamble. It's almost impossible for all of this
>     to work itself out without at least one, and probably more,
>     spoiled elections; and then we're back to point 1 above.
>
>
> dlw: Yes, it's indeterminant who will be the new two major parties for 
> the local area, but either way they'll be centered around the true 
> center.  And, the "spoiled" election will be spoiled in the right (or 
> center-ward) direction rather than the wrong direction with FPTP.  And 
> with learning, it'll become more customary for the two biggest parties 
> to adjust to the moving center so that these sorts of "spoiled" 
> elections are not likely.
>
>
>         But it's inane to think that a particular third party is going
>         to be able to get as strong as the Prog party of VT in
>         Burlington at the Nat'l level w.o. getting coopted by the
>         major parties, jealous to keep their duopoly positions.  So
>         the chances of this happening are nil in prez elections.
>
>
>     Watching the Republicans repeatedly shoot themselves in both feet,
>     throwing away what could have been a winnable election for them,
>     doesn't give me confidence that major US parties will always
>     embrace rationality.
>
>
> dlw: The wages of FPTP are the game "Bloody the leader for a bloody 
> long time"... but it's only because of the lack of intra-party 
> discipline that their primaries are going on so long.  That can be 
> remedied easily...
>
>
>         The problem with IRV in such elections wd be vote-counting and
>         that is fixed by IRV3/AV3....
>
>
>     Well... OK, I guess, although it's still only summable O(Nˆ3),
>     which is worse than any other polynomial-summable system I know of
>     (except some summable RRV-like pseudo-PR systems I've invented and
>     never published).
>
>
> dlw: Dude, in real life, the difficulties of winning a single-winner 
> election(time/money commitments vs relative chance of winning) keeps N 
> down and so it's not useful to extrapolate....

with computers counting, i am not terribly worried about the 
computational burden of either IRV or Condorcet.  but the precinct 
summability remains important.  even IRV can be precinct summable, but 
the number of piles (it's the subtotals for each pile that is 
transmitted up the line to the central election authority and also 
promulgated to the media, campaigns, and other interested parties) when 
the number of candidates grows is much higher than either FPTP and 
Condorcet.  of course, precinct summability is important for 
transparency which is important for election integrity.

>
>         dlw: What about IRV3/AV3?  Like I said above, it's not easy
>         for a third party to get as strong as VT Prog was in
>         Burlington at the state or nat'l level... and the dynamics wd
>         be towards a change in the nature of the two major parties
>         more so than the continuation of a competitive 3-way election....
>
>
>      IRV3/AV3 doesn't fix spoilers. And as I said above... (see args 1
>     and 2).
>

it can, but does not always.  i would say IRV fixes the hard-core 
spoiler problem (strictly speaking, a "spoiler" is a candidate with NO 
chance of being elected whose presence in the race changes the outcome 
of the election) but not the "spoiler-lite" problem.  the GOP candidate 
was the spoiler-lite in 2009.  nonetheless, IIA applies whether the 
spoiler is hard-core (like Nader) or lite (like Kurt Wright in 2009 and 
who is the GOP candidate again this year).

IRV works like shit when the election is between three or more 
candidates with roughly equal support.  that is how i imagine when we 
finally arrive at a viable and lasting multiparty environment.  if only 
2 parties hog nearly all of the support, it's essentially still a 
two-party system.

> It matters for the world who "spoils" for who.  The threat of spoiling 
> instills a center-ward dynamism that is abstracted from in most 
> rational choice theories of electoral rules because it's hard to model 
> deterministically, as you demonstrate above.

i do not understand what you mean.  *any* spoiler, when discovered, will 
result in a tactical pressure for the group that supported that spoiler 
and hated who was actually elected.  here in Burlington, it was the GOP 
Prog-haters.  if IRV survived to this year, these folks, if they're 
savvy, would have to consider whether marking their favorite candidate 
as #1 will *again* cause the election of their least favorite 
candidate.  that is precisely the tactical voting (the tactic is called 
"compromising") that we were trying to avoid with the adoption of IRV.  
rather than eliminate that burden of tactical voting, it was reduced by 
transferring it from the shoulders of the liberal majority (that didn't 
have a painful choice between Dem and Prog) to the GOP Prog-haters, a 
smaller group who found out that simply by marking their favorite as #1, 
they *caused* the election of their least favorite.  that's gotta make 
some people mad.

-- 

r b-j                  rbj at audioimagination.com

"Imagination is more important than knowledge."