The Davison Run-off (Coombs) #2 letter

[New Democracy]

Dear members of this list,

On Thu, 26 Dec 1996, Dr. Marsha Pripstein wrote:

please correct me if I['m wrong, but I don't see how this could work.
IRO does not require that every voter rank every candidate, so so might
only go 1-2-3, while others would go 1...5, etc.  so how do you decide on
who gets the most "last place/last selection" votes?

MPP
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - December 27 1996
Dear Dr. Marsha Pripstein,

     You are correct to be concerned about the vacant selections in the
last set but I am on solid ground. The mathematical calculations to
determine which candidate to drop are done on the side. If a voter did not
make a last selection I am safe in saying to this voter the following: "You
did not select one candidate so I am going to give that candidate one
demerit towards being dropped." And in a case in which a voter did not
select two candidates I will say: "You did not select two candidates so I
am going to give those two candidates one half demerit each towards being
dropped."
     These demerits from the vacant selections are only going to be used to
help determine which candidate is to be dropped - these demerits will not
be carried back to the election results. These demerits will be added to
the demerits we already have created from the last selections.

Allow me to give an example - one of Steve's many examples:

   ELECTION               SELECTION SETS
    RESULTS           ONE      TWO      THREE
     46 AB            46 A     46 B     (    )
     20 B             20 B    (    )    (    )
     34 CB            34 C     34 B     (    )

In order to decide which candidate to drop I will calculate on the side.

     Side Calculations:     A     B     C
                                       46
                           10          10
                           34
                          ----  ----  ----
         Total Demerits    44     0    56

The 46 AB voters did not make a third and last selection so I am giving 46
demerits to candidate C towards being dropped.

The 20 B voters did not make two selections so I am giving ten demerits
each to candidate A and C towards being dropped.

The 34 CB voters did not make a third and last selection so I am giving 34
demerits to candidate A towards being dropped.

I total all the demerits and I find that candidate C has the most with 56
demerits - so candidate C is dropped.

     In the event we had the extreme condition - an election in which all
the voters only made one selection each - my method of constructing
demerits for the missing last selections would end up dropping the
candidate with the lowest votes on the first and only set of selections. I
mention this to show that my method will fit in with Instant Run-off under
any condition.

Ties: In the event there was a tie between candidates in the side
calculations I would look to the first selection numbers in order to break
the tie. The candidate of the tied candidates with the lowest number in the
first selection set would be dropped. If two of three or more of the tied
candidates have the same lowest numbers in the first set of selections they
are both dropped.

     In the case of more than three candidates running in the race the
winner of the last set of selections is dropped first as I did in the above
example. This would produce some changes in the other selections - causing
new selection sets. The winner of the last set of selections of these new
selections would now be dropped - etc - until we have a winner with a
majority.

     Condorcet people have often said that it is not proper for Instant
Run-off to drop the last candidate. The solution is simple - we do not drop
the last - we drop the candidate that is the lead candidate in the last set
of selections. The reason this candidate has the most demerits is because
he received the least favor from the voters.

     In a single seat election we are only going to elect one candidate.
The field must be reduced down to one. I have always believed in dropping
candidates and I still believe in dropping candidates. I was willing to be
arbitrary and drop the lowest candidate even if this candidate had support
in the second selection set - who else to drop - I must drop someone - I
will drop somone.

     I often would think if I could somehow reach in and grab the least
supported candidate I would drop him - well now I can - the least supported
candidate is the candidate with the largest number of demerits from the
sidebar calculations.

     Thank you, Dr Marsha Pripstein for writing - I gained the most. I must
now think how this change can be used in multi seat elections.

Sincerely yours,

Donald Eric Davison of New Democracy at http://www.mich.com/~donald