[EM] IRV mailing party adventures

Tom Ruen tomruen at itascacg.com
Tue Apr 3 20:14:04 PDT 2001


I went to a mailing party for FairVote Minnesota, which supports IRV.

Some opinions that came out of discussion:
1. Thinking ranking is much easier than thinking approval since approval
requires a cut-off strategy.
2. Tied ranks in IRV by split-votes may be fair but may complicates the
system more than warranted considering most voters would rather rank fully.
Making a special case for a small minority of voters who want to be lazy and
not decide is not important.

These are practical opinions to consider. Generally it would see to me that
approval won't be supported if more than 90% voters never want to vote for 2
or more choices, even if it is a practical reform. IRV has a chance for
being implemented because it makes no new rules to runoff elections, only
simplifies the task for voters to only need one ballot. Any other reform is
going to get more resistance.

I think the reason IRV supporters get so defensive is because they're
fighting both sides, those that don't want any change at all and are content
with plurality, and those like election-method people that think runoffs and
IRV as inferior. Systems, especially systems allocating power, are always
very conservative against any change. Your best bet is to use the rules of
the existing system and make it work better. Anything else among those
interested in actual reform will more likely as not lead to insanity!

There was also a question about Condorcet which I couldn't well answer.
(Because I'm dense!)

If a voter leaves some candidates unranked, what does that mean in terms of
pair elections? My assumption is that a ranked candidate automatically beats
an unranked candidate on a ballot. A man I talked to argued that there
should be no tally in that pair election for that ballot. I first thought
that it should make no difference, that it wouldn't change the results, but
then I was unsure and confused.

For evaluating each ballot with election between A and B:

Approach 1: (My approach)
1. A and B ranked, A has a higher rank. Give 1 point to A.
2. A and B ranked, B has a higher rank. Give 1 point to B.
3. A ranked, B unranked. Give 1 point to A.
4. B ranked, A unranked. Give 1 point to B.
5. A and B unranked. No points.

Approach 2: (alternative)
1. A and B ranked, A has a higher rank. Give 1 point to A.
2. A and B ranked, B has a higher rank. Give 1 point to B.
3. A or B not ranked, or both. No points.

I would assume a voter ALWAYS wants a ranked candidate to get a vote over an
unranked candidate, but I suppose this may not be the case. He claimed the
system shouldn't make assumptions about what it means to leave candidates

Now I see this MUST make a difference, that we must give a win to a ranked
candidate over an unranked candidate! We can't say Gore supporters don't
have an opinion between Gore and Bush just because Bush is left unranked!
That is too crazy to contemplate! The unranked default must always mean last

I wish I could think better on my feet! Damn philosopher tricked me!

Tom Ruen

More information about the Election-Methods mailing list