[EM] Fwd: U/P voting: new name for simple 3-level method.
cbenham at adam.com.au
Fri Sep 9 12:48:41 PDT 2016
Jameson and any interested others,
I remain strongly of the opinion that in any election method the default
rating/ranking should simply be
bottom-most, and I'm opposed to the gimmick that tries to stigmatise
some of the losers at the time of
the following election in the event that they contest it.
But those issues aside, I now rate U/P as not really any worse than MTA
Of those three, only U/P meets the Chicken Dilemma criterion.
And the recently discussed U/P bad example where it fails to elect a
candidate that is both most approved
and the Condorcet winner is shared by MTA.
Of the three methods, I agree that U/P is the most strategy-resistant
but also the one that is the least likely to elect a
voted Condorcet winner and the one that with sincere voting will likely
produce the lowest SU winner. With MCA the
reverse is true, and MTA is intermediate between the two in terms of
In some ways I imagine that U/P would be the hardest to promote to the
wide public, because compared to the others
it looks like the middle-rating doesn't do anything and I doubt that you
want to bring up negative devious strategies
like "chicken" defection.
IBIFA also (along with MTA and MCA and MJ and MAM) fails CD, but, unlike
U/P, MTA, MCA, MJ, meets Irrelevant Ballots
Independence, absolutely dominates those methods in terms of Condorcet
consistency and is happy using ballots with many
more (than 3) ratings slots.
I don't share Jameson's hostility to IRV. Provided the voters are free
to strictly rank from the top however many or few
candidates as they wish (I don't favour allowing above-bottom
equal-ranking in IRV) and the eliminations are (at least in
effect) one-at-a-time, then I rate IRV as a good method, much better
than Top-Two Runoff and in my book the best of the
methods that strictly meets Later-no-Help.
And if we want something that meets Condorcet and CD, then "Benham" (the
name was coined by James Green-Armytage
in an article he wrote about IRV-Condorcet hybrid methods) is a very
good method. It is like IRV, but elects the CW and if there
isn't one checks for one among the remaining candidates before each
elimination and elects the first one to appear.
Of the promoted and other ok Condorcet methods, I think it is the only
one that has in common with IRV that a Condorcet
winner that can make the IRV top-two can't fall victim to a successful
More information about the Election-Methods