[EM] More non-altruistic attacks on IRV usage.

[Dave Ketchum]

Thanks for worthy comments, but I disagree a bit:
      We need single-member districts, for we have offices that fit,  
such as mayor and governor.
           We need to ban plurality.   While plurality is enough on a  
good day, most any election can have bad days.  I will promote  
Condorcet (see B2 below) - among its advantages are that voting here  
is no more effort than plurality on good days (think of a community's  
treasurer - simply reelect via ranking only such on good days; want to  
demand replacements on bad days).
     We need an agreed method for doing PR for such as legislatures.   
They can be done single-member, but those managing elections can  
choose PR.  While STV exists, I suggest having the voters use  
something more like Condorcet for PR.

On Dec 2, 2011, at 1:35 AM, Jameson Quinn wrote:

> This thread now has 50 messages, back-and-forth. I'll try to make  
> this my last word on the subject.
>
> Basically, the bottom line for me is that I trust real evidence more  
> than I trust theory, but I need to find room to take hopeful action.  
> That's not a matter of building an elaborate model of reality in my  
> head and then repeatedly claiming that I'm a pragmatist; it's a  
> matter of trying to make my questions as simple as possible,  
> answering them with evidence, and then finding the shortest path of  
> least resistance to hope.
>
> What does the evidence tell us?
>
> A. Evidence about the status quo says:
> 1. Plurality is a theoretically-horrible system, with no redeeming  
> features.
> 2. Single-member districts have certain advantages, but also serious  
> problems; I'd say that on the whole the problems dominate. (?)
> 3. In practice, the problems with both plurality and single-member  
> districts seem to culminate in two-party domination.
> 4. It takes a lot of money to get elected in the current system.
> 5. Status quo politics are badly broken.
> 6. It's likely that 3 is one main cause of 4, and that 3 and 4  
> together are the main causes of 5. Thus there is a need to change  
> either plurality, single-member districts, or both. (?)
>
> B. Evidence about IRV says:
> 1. There's been a well-organized and decently-funded national  
> campaign for IRV. I'm speakin of course about Fairvote, whose  
> spending on IRV over its history has probably totalled millions of  
> dollars.
      Fair Vote offers a valuable service to voters - better than just  
approving candidates, as in Approval, voters use ranking to indicate  
whether they like A or B better - but are not required to indicate  
amount that A is better than B.
      Fair Vote also gives a simple task to vote counters -  
recognizing that small groups of voters can like best different groups  
of candidates, discard such top groups until the winner has a majority  
of what remain top.

> 1a. It's had real successes
> 1b. It's still fallen widely short of the progress that is needed.
       It can happen that one of the top groups discarded, per above,  
was only part of the votes for the truly best liked candidate - who  
thus fails to win.

> 2. Even in places that were initially favorable to IRV, and have  
> tried it, opposition is persistent. (This includes Australia, where  
> reputable polls have found majorities favoring changing the system.)
> 3. IRV pathologies can happen in real life.
      Burlington proves what simulations tell us to expect.

> 4. Especially when pathologies happen, IRV is subject to repeal.
> 5. IRV does not seem to end two-party domination; certainly it does  
> not do so reliably. (?)
> 6. In a hard-fought national referendum in the UK, where both sides  
> had significant funding and organization, IRV lost resoundingly.

B2. Condorcet has had less use than IRV.
1. It offers the same service to voters, except also permitting equal  
ranking.
2. Counting is (as if) into an x*x matrix showing which of each  
possible pair of candidates would win in a race between those two.
3. There is value in humans reading x*x - it tells how third parties  
are doing even when they do not win - clues as to whether they are  
worth joining; clues as to where the center of gravity is moving.
4. The Condorcet Winner (CW) is recognized as proper winner even when  
discussing other methods.  It means winning when racing against each  
other candidate with Condorcet counting.
5. Only by having at least 3 strong candidates and them being voted in  
a cycle such as A>B>C>A, is there no CW.
6. I and a few others argue strongly that only candidates the voter  
could approve getting elected should be ranked - and against ranking  
others such as enemies.
>
> C. Evidence about other single-winner systems says:
> 1. Non-IRV voting activists are, as a whole, fractious and  
> disorganized.
> 2. It is very difficult to get all voting reform advocates to agree  
> on a single best system.
> 2a. It's especially difficult to get theorists to support IRV in  
> spite of its theoretical flaws. (?)
      Not surprising, since the flaws are real.

> 3. It is less difficult to get reform advocates and theorists to  
> agree that a set of systems are all better than plurality.
> 4. Other single-winner reforms haven't been implemented much.
> 5. Therefore, there is little evidence of what would happen after  
> they were implemented, although we can theorize. (?)
>
> D. Evidence about PR says:
      It cannot do its thing without electing multiple officers, as  
for a legislature.  There it competes with single-member.

> 1. PR can end two-party domination.
> 2. With PR, there can still be fewer competitive elections and more  
> safe seats than voters would like to see. (?)
> 3. When combined with a parliamentary system, PR can lead to  
> instability.
> 3a. But there are reasons to believe that those problems would not  
> generalize to a presidential system. (?)
> 4. PR is a more-radical change than single-winner reform.
> 4a. It may be harder to promote to an American audience.
> 4b. It may be harder to sell to politicians who have won in the  
> status quo.
> 5. PR systems can be tuned to optimize various advantages, but it's  
> hard to find a system which is perfect in all ways (simple, local,  
> voter-centric, doesn't require ranking dozens of candidates) (?)
>
> There's plenty of reasons for pessimism in the above. David seems to  
> find his optimism by emphasizing points B1a, C1, C4, D1, and D5, and  
> giving (plausible) counterarguments for points B1b, B2, B3, B4, B5,  
> B6, C2a (though he backed off from a bet), D4a, and D4b. That's 9  
> points he's trying to overcome (though since B4 is little more than  
> B2+B3, I guess it may be more like 8 than 9).
>
> I on the other hand think that the path of least resistance is to  
> emphasize C3 as a way to overcome C1, C2, and C4. I think that it's  
> better to fight reality on 2-3 points than on 8-9, no matter how  
> plausible the arguments that the 8 or 9 battles are winnable.
>
> One specific response:
> JQ:
> 3. Some other organization pushes some other system(s), and reaches  
> a tipping point.
>             dlw:IOW, they need to reinvent what FairVote's been  
> working hard to build up for some time...
> Yep. It's a lot of work. If voting reform were an easy task, we (and  
> I include Fairvote in that "we") would have won already.
>
> JQ
Dave Ketchum
> ----