[EM] remember Toby Nixon?

[fsimmons at pcc.edu]

Jameson,

I like the idea of candidates submitting binding public rankings before the election.  Besides resolving 
the "kingmaker" problem, it could be useful information to the voters in other ways as well..



----- Original Message -----
From: Jameson Quinn 
Date: Tuesday, May 24, 2011 4:41 pm
Subject: Re: [EM] remember Toby Nixon?
To: fsimmons at pcc.edu
Cc: election-methods at lists.electorama.com

> I think DYN is my new favorite practical proposal. It's simple 
> and it would
> work beautifully.
> 
> The one downside of that system would be the possibility of 
> granting too
> much power to a minority kingmaker. For instance, a 4% candidate 
> could have
> the power to swing the election to either one of two 48% 
> candidates. They
> might well be able to negotiate concessions for their party (or 
> worse, for
> themselves personally) which amounted to, say, a 20% share of 
> the power, far
> in excess of their actual support. The only way to minimize this 
> risk is to
> minimize the enforceability of any promises made between the 
> voting rounds -
> for instance, by ensuring that all cabinet positions can be 
> dismissed at
> will.
> 
> Hmm... another way to address this would be to have candidates 
> pre-decide
> their full preference order. After the first round, they would 
> only be free
> to set their threshold. This would halve the chances that they'd 
> end up as
> kingmakers, which is fair, because the winning 51% coalition gets
> essentially twice that much power.
> 
> Anyway, this issue is actually a pretty good problem to have. 
> Giving a
> slightly-larger minority of power to a minority in some 
> circumstances is not
> the end of the world.
> 
> I like it.
> 
> Jameson
> 
> 2011/5/24 
> 
> >
> > About six years ago Toby Nixon asked the members of this EM 
> list for a
> > advice on what election method
> > to try propose in the Washington State Legislature. He finally 
> settled on
> > CSSD beatpath. As near as I
> > know nothing came of it. What would we propose if we had another
> > opportunity like that?
> > It seems to me that people have rejected IRV, Bucklin, and 
> other methods
> > based on ranked ballots
> > because they don’t want to rank the candidates. Charles 
> Dodgson (aka Lewis
> > Carroll) anticipated this
> > difficulty in 1884, and he suggested what we now call Asset 
> Voting as a
> > solution.
> > Asset voting is the simplest solution to the spoiler problem. 
> Approval is
> > the next simplest. IMHO
> > anything much more complicated than Approval or Asset voting 
> doesn’t stand
> > a chance with the general
> > public here in America. For this reason most IRV proposals 
> have actually
> > truncated IRV to rank only
> > three candidates. This destroys IRV’s clone independence.
> > Asset Voting in its simplest form tends to squeeze out the CW, 
> because when
> > flanked closely on both
> > sides by other candidates, the CW tends to have too few first place
> > preferences (assets or bargaining
> > chips) to survive.
> > On the other hand Approval requires reliable polling 
> information for
> > informed strategy. This fact makes
> > Approval vulnerable to manipulation by disinformation.
> > That brings us to Delegable Yes/No (DYN) voting, which is a 
> hybrid between
> > Asset Voting and Approval
> > that overcomes the weaknesses of those methods without 
> increasing the
> > complexity to the level of IRV:
> > In DYN you circle the name of your favorite candidate and then 
> optionally> mark “Yes” next to the
> > candidates that you are sure you want to approve of, and “No” 
> next to those
> > that you are sure that you
> > want to disapprove of. You automatically delegate the rest of 
> the Yes/No
> > decisions to the candidate that
> > you circled as “favorite.”
> > Those delegated decisions are made by the candidates after the 
> partial> results have been made public,
> > so that no false polls can manipulate the strategy.
> > What do you think?
> > ----
> > Election-Methods mailing list - see http://electorama.com/em 
> for list info
> >
>