Re: Fw: counting preferential ballots

From: Hubert Bray <bray_at_math_dot_duke_dot_edu>
Date: Fri Oct 29 2004 - 14:53:33 CDT

Dear Arnie,

Since, as you say, the goal here is to have an open source solution which is
voting method neutral, then the bottom line (for now) is to make sure the
"counting" of the votes keeps tracks of the proper statistics. Hence, in
addition to keeping track of how many 1st choice, 2nd choice, and 3rd choice
votes (etc.) a candidate gets, the software should also keep track of what I
was calling the "Margin of Victory Matrix," or what Arnie was calling (I
think) the Condorcet scores of each binary relationship. In other words, we
need to know by how many votes Candidate A would beat Candidate B by in a
head-to-head matchup (which can be determined by looking at each voter's
preferences), for all possible combinations of two candidates.

Hubert Bray

On Oct 28,2004 12:25:38 -0400, Arnold Urken <> wrote :
>Dear Alan and Hubert:
>I am glad to see a revived discussion of voting methods. The implications
>for vote counting and auditing have not been properly addressed to the best
>of my knowledge. But let me suggest a couple of things to remember:
>1. There may not be a tie for the Condorcet winner, so it may be
>unreasonable to expect a unique universal winner. Generally, unique winners
>are more likely under Copeland scoring, which compares the Condorcet scores
>of each binary relationship.
>2. All voting methods are manipulable, including the least-worst metric. An
>important distinguishing characteristic is the type and amount of knowledge
>required to carry out a manipulation.
>Open source solutions should be voting method neutral, but should be used
>intelligently so that adopters of new voting methods are not surprised by
>problems such as frequently-tied elections, something that can happen with
>approval voting.

Hubert Bray
Mathematics Department
Duke University

= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
Received on Mon Nov 1 15:28:55 2004

This archive was generated by hypermail 2.1.8 : Mon Nov 01 2004 - 15:28:59 CST