[EM] 01/17/03 - No Problem, Olli:

[James Gilmour]

Donald replied to Olli:
> Olli, you wrote: "I can see nothing that could be excluded."
>
> Donald here:  Then I shall explain for you that which `could be excluded'.
> In your example of three seats and three candidates with one-third each,
> that which is excluded are the votes between the Droop quota and the
> one-third, so called surplus held by each of the three candidates.

This would NOT be done in a Droop STV-PR election count.

> The
> Droop method demands that surplus votes be transferred, but none of the
> three are able to accept any votes because they are already over the limit
> of the Droop quota.

This statement is incorrect.

The Droop method does NOT "demand" that surplus votes are transferred.  Votes are
transferred only if such transfers are necessary to complete the election, ie to
determine the required number of winners.

In the case presented by Olli, no votes would be transferred.  The three one-third
candidates each has a Droop quota of votes at the first stage of the count, so
they are the three winners and the count would be terminated at that point.


>  We cannot say the votes become exhausted because the
> voters may have made enough lower choices.  The problem is that there is
> not enough room in the quotas of the three candidates to accept additional
> votes, therefore the votes are said to be excluded.

There is no problem.  No votes would be transferred.  No transfer is necessary.
Three candidates have achieved a quota of votes  on first preferences alone, so
these three are the three winners the election was held to identify.


> Olli: "It's unnecessary to do the math because A, B and C reach the quota.
> You can declare the result and have a coffee or whatever."

Olli is absolutely correct.


> Donald:  It is necessary to do the math because, like I've said to James,
> `The job is not over until the paperwork is done.'

This statement is meaningless as the election has achieved its purpose, ie
identified the three winners.  All the paperwork required for that purpose has
been done.


> Doing the math is
> normal for every election method, so it should also be normal for the Droop
> method of STV.

This is "normal" only by Donald's definition of "normal".  It is never done in any
public election I have encountered, no matter what voting and counting system is
used.  Votes are sorted and counted only so far as is required to secure the
election of the required number of candidates.


>  Besides, the final math will tell us how representative
> these candidates will end up being.

This may be true, but it is irrelevant to the purpose of the election, ie to
identify the three winners.

> They started at one-third each,
> perfect representation, but the Droop method reduced them down to only
> one-quarter each, something wrong about that Droop guy.

This statement is, once again, wrong.  No matter how many times Donald repeats it,
it will never become correct.  In the example with  three candidates each of whom
has one-third each of the first preferences, there will be no reduction of the
votes of these three candidates.  Each has more than the relevant Droop quota, so
these are three winners and the counting procedure stops.


> Olli: "What do you propose to do if you use the Hare quota and there are
> four rather equal candidates vying for three seats: 25.1% A,  25.!% B,
> 25.1% C,  24.7% D.[?]
>
> Donald:  No problem!  In your example, candidate D is the lowest, so I
> would eliminate candidate D and transfer his votes.  Depending on the lower
> choices we could end up with the following results:  33% A   32% B  31% C
> which would be very good results, very representative of the public, the
> Droop method would never be able to have results as good as that.

The Droop quota would give exactly the same result in this election, ie it would
identify the same three candidates (A, B and C) as the three winners.  The rest is
irrelevant to the purpose of the election.

James

----
For more information about this list (subscribe, unsubscribe, FAQ, etc), 
please see http://www.eskimo.com/~robla/em