Why 3 and not 4?

Forest Simmons fsimmons at pcc.edu
Fri Sep 14 16:02:57 PDT 2001

The following is not intended as a major distraction from the Oregon IRV
initiative campaign, so only those with time on their hands who are also
curious enough to read another response to the "Why 3 not 4 ?" query
should read further :-) 

On Sat, 11 Aug 2001, Fillard Rhyne wrote in reply to the query:

> >SM> Does anyone know why we are being offered to rank 3 instead of
> >SM> four candidates?
> Ideally, an instant runoff ballot would allow you to mark an
> unlimited number of choices. For example, Cambridge, Massachusetts
> uses a system closely related to instant runoff, and if there are
> 20 candidates printed on the ballot, they get to rank 20 choices.
> However, most of the ballot-scanning machines currently used in
> Oregon can't read more than three ovals in any one row. Since we
> feel ballots are most intuitive when all the ovals for a single
> candidate can be laid out in a single row, we decided to allow
> elections officials to limit the number of choices to as few as 3..
> The phrasing we use in the initiative is "at least 3", thereby
> leaving the door open for elections officials to allow more than 3
> choices in the future when ballot-scanning machines are more
> capable.

My alternate response:

Another possible use of the "three ovals per candidate" ballot is as

Each voter rates each candidate on a scale of one to three, with three
considered the best rating. An unrated candidate on a ballot gets a
default rating of zero from that ballot. 

The candidate with the highest average rating wins the election.

Alternately, the candidate with the greatest number of two's and three's
wins the election. 

Both of these simple methods have several practical technical advantages
over the IRV use of the same ballots, namely Summability, Consistency,
Monotonicity, Reverse Symmetry, Independence of Irrelevant Alternatives,
and satisfaction of the FBC.

This last property means absolute immunity from strategic incentives to
rank a favorite below a compromise. IRV is not absolutely immune to this
problem even though it does solve the "lesser evil" problem in certain
common cases.

For more information about these technical criteria or about even better
uses for the three oval style ballot, contact me at the above email

I'm not proposing switching horses in midstream. This alternative response
to the query about three oval ballots is only for those who are curious
about the variety of available election methods for future improvements.

Peace to All,


More information about the Election-Methods mailing list