[EM] Voting Systems Study of the League of Women Voters of Minnesota

Markus Schulze markus.schulze at alumni.tu-berlin.de
Mon Oct 11 03:41:31 PDT 2004

Dear Ralph,

they don't explicitly write that they recommend IRV, but
the study is written in such a manner that a naive reader
will necessarily get to the conclusion that IRV was the
best method.


Example 1: In this study, they compare FPP, Approval Voting,
Borda, Condorcet, and IRV. In the summary, they criticize
FPP, Approval Voting, Borda, and Condorcet for being
vulnerable to manipulation. But they don't criticize IRV
for being vulnerable to manipulation. A naive reader will
get to the conclusion that IRV was immune to manipulation.
Otherwise, this criticism doesn't make any sense.


Example 2: Usually, a "spoiler" is a candidate A who takes
away the win from candidate B _without being elected_. But
in this study, they define a "spoiler" as a "third candidate"
who takes away the win from one of the other two candidates.

For example, suppose candidate A is the Democratic candidate
and candidate B is the Republican candidate. Suppose
candidate C is an independent candidate. Suppose furthermore
that a Condorcet-consistent single-winner election method
is being used and that candidate C is the Condorcet winner.
Then candidate C is _not_ a spoiler according to the common
definition of "spoilers". But according to the definition
in this study, candidate C is a spoiler.

For example, in the glossary they write: "Spoiler effect:
Occurs when a third candidate takes enough votes away
from a candidate that it causes the candidate to lose."
In the summary, they write that IRV "eliminates problems
of spoiler candidates knocking off major candidates".


Example 3: In this study, they always stress that Approval
Voting, Borda, and Condorcet were good methods only according
to "mathematical" considerations. They always stress that
IRV's violation of monotonicity was only a "mathematical"
problem. They always stress that only mathematicians
consider e.g. the Condorcet criterion and the monotonicity
criterion important.


Therefore, a naive reader will get the impression that they
have done a good job in analyzing the different election
methods and that IRV was the best method to "ensure majority
rule" and to get rid of the spoiler problem.


You wrote (10 Oct 2004):
> If you are an advocate of IRV, you are going to have to
> support your position with better evidence than this.

I disagree with you. In my opinion, this study clearly is
in favor of IRV. This contradicts your claim that you need
better evidence.

In my opinion, this study rather raises other problems:

Problem 1: In the summary the fact that there is not
a single Condorcet-consistent method, but rather a very
large number of Condorcet-consistent methods is used as
an argument against Condorcet-consistent methods.

They write: "The Condorcet system does not always produce
a winner. As a result, election officials must decide before
the election on a method to break a tie." "Condorcet may
result in a tie that requires pre-election decision on how
to break tie."

Problem 2: We need an evidence _from a professional law
scientist_ that Condorcet-consistent methods are not

Markus Schulze

More information about the Election-Methods mailing list