[EM] Electronic Voting Bill of Rights?

[Dave Ketchum]

On Mon, 17 Nov 2003 15:37:00 -0000 James Gilmour wrote:

> David GLAUDE asked:
> Once again as this list is dedicated to advanced voting/counting 
method. 
> I would like to know which one are possible to handle with manual 
> counting on large scale.
> 
> Manual sorting and counting works OK for STV-PR in public 
elections.  All electronic voting
> introduced for the first time in the Dáil Éireann election in 2002, 
in three constituencies
> (districts): full ballot preferences have been posted on the web.
> James 

Many of the methods discussed have complexities that make them 
impractical for the original question asked here.

However, a few methods may be worth more thought.  To go for 
large scale, let us look at the California recall, but assume nothing 
is known as to which are sure losers:

Step one is to express your opinion about whoever was responsible 
for setting up such an unbelievable challenge, and move on with 135 
candidates and many hundreds of precincts.

PLURALITY:  Each precinct counts its own ballots, forwards the 
sums, and all that is left is to add the sums together.

APPROVAL:  Same as Plurality, except counting is a bit more effort, 
since you cannot simply pile ballots in stacks as Plurality workers 
may have chosen to do.

IRV:  There are choices:
     Count as in Plurality and hope for a majority winner.  If not you 
have identified at least one loser and can send back for adjusted 
counts.  Repeat until you have a winner.  Let this take many rounds 
and the counters are going to get tired and dizzy and their accuracy 
take a nosedive.
     Forward every ballot for central counting.  The forwarding takes 
an unacceptable amount of time and you still need counters.
     Forward an electronic copy of each paper ballot for central 
counting.  Precinct workers go home early, except that these electronic 
copies could be a problem to prepare, but there is no paper for counters 
to work on.  Variation: Do the forwarding electronically, but FAX so 
that the result is paper as in previous choice.
     Sort ballots by pattern voted, and forward each pattern with its 
count.  Much like previous choice, except there is less data to 
forward.

      Note that with a more believable quantity of candidates, forwarding 
individual ballots would have been LESS likely.

CONDORCET:  Choices here also:
      Count as in Plurality?  Tempting, since a majority winner would end 
the task.
      Forward every ballot, or sort by pattern and forward?  Tempting, as 
with IRV.
      Do the Condorcet matrix for this precinct and forward that.  The 
obvious choice for any reasonable quantity of candidates since the matrix
would be small and contains everything central counting needs to know 
(rejecting any variants of Condorcet that want more detail).

For any reasonable quantity of candidates I see Condorcet ahead for, 
besides liking its choice of winners better than IRV's (whenever they 
disagree), the Condorcet matrix (x by x for x candidates) is easy to 
calculate and less data to forward or to publish as a public record (do it 
once and be done) than the IRV choices.
-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.