[EM] IRV vs Condorcet

Alex Small asmall at physics.ucsb.edu
Thu Jul 24 11:14:51 PDT 2003

Eric Gorr said:
> Perhaps it isn't an argument for or against any particular method,  but
> I would consider it an important point to make about any
> particular method.
> It is because IRV fails to count votes in a better way that
> contributes to (if not causes) its failing of more fundamental
> criteria important to every voting method.

But the problem is this:  What is the better way to count votes?  That is
a large part of the disagreement between IRV and Condorcet fans.  The
description of a method isn't an argument for or against the method,
because each camp can look at the descriptions and put their own spin on
it.  When somebody says "Condorcet looks at all pairs of candidates, while
IRV looks primarily at first choices and then at lower choices as needed",
there are two possible responses:

An IRV fan will say "Exactly!  IRV gives greater weight to a voter's top
priorities.  Condorcet gives too much weight to a voter's lesser
priorities, which runs against the voter's wishes."

A Condorcet fan will say "Exactly!  Condorcet makes maximum use of
information provided by voters.  IRV doesn't use all of the information
that a voter provides, so it isn't really heeding the voter's wishes."

So where do we go?  These subjective arguments about "giving greater
priority to first choices" vs. "use as much information as possible" can't
really be resolved, because neither one really refutes the other.  The
only thing to do is get into the nitty-gritty about what can happen in
each method and why each phenomenon is either desirable or undesirable.


More information about the Election-Methods mailing list