[EM] Real IRV Election, Disputable Result

Jan Kok jan.kok.5y at gmail.com
Fri Mar 10 20:31:03 PST 2006

> >Thanks for doing this analysis!  This is BIG news in the small world
> >of voting methods! :-)
> How so? It's well known that IRV/AV/STV doesn't necessarily find the
> Condorcet winner. It shouldn't be too surprising that there are
> real-world examples.
> --
> /Jonathan Lundell.

Yes, Jonathan, of course it is well known -- to people on the EM list
and a few other lists that aren't under the spell of the Center for
Voting and Democracy's propaganda.  I suspect that most people who
promote IRV have no clue about the numerous flaws of IRV, or if they
have heard about the problems from some person or some web site, the
significance of those problems hasn't "sunk in".  (I was among those
clueless for about a year.)

I have kept my eye on the EM list for several years, and I have not
seen, until today, an example of Condorcet failure in a real-life
public IRV election.  I believe the reasons for not seeing any such
examples are:

1. The raw ballot data which is needed to do the analysis has been
unavailable except for a very few, recent elections.  Australia has
been doing IRV elections for many decades, but the way the IRV results
are summarized and reported, it is impossible to determine if the IRV
winner is not the same as the Condorcet winner.  (Of course, if some
candidate gets a majority of first-choice votes (which is often the
case), then that candidate is always the IRV winner and Condorcet

2. Despite many decades of using IRV and proportional representation,
Australia still has two dominant parties.  (It is debatable whether
IRV helps keep that two-party system in place, the way Plurality
voting does in the US.  There might be other factors that keep the
third parties small in Australia.)  Because third parties seldom get a
small fraction (typically less than 20%) of the first-choice votes,
there are few opportunities for IRV and Condorcet to get different

So, it's great to have this example of Condorcet failure in an IRV
election -- right here in the US! -- so soon after IRV was implemented
in a few cities around the US.

I don't want to slow down voting reform in general.  But I would love
to slow down the spread of IRV, so we non-IRV reformers can implement
some other, hopefully better voting methods.  I would welcome the help
of any IRV enthusiasts who decide, for whatever reason, that IRV is
not the best method to promote.

And yes, _any_ non-Condorcet method can exhibit Condorcet failure, by
definition!  But most of us who promote those other methods will admit
that fact.  In my experience, it's hard to get IRV supporters to
understand and acknowledge the problems with IRV.

- Jan

More information about the Election-Methods mailing list