[EM] Clarity of principle.
Richard Lung
voting at ukscientists.com
Sun Oct 15 09:06:17 PDT 2023
Clarity on principles
Dear EM list,
If a thing needs doing, there is no point in complaining about it being
hard. This was a point made by Winston Churchill.
The US Constitution could not ban an election system, not invented till
the next century.
At least, those who introduced a quota-preferential bill to Congress
apparently didn’t think so. And what the cities could do, surely the
federal government also could.
State-level suppression of local autonomy in introducing proportional
representation, notably Massachusetts state banning other cities, in the
state, than Cambridge from using STV/PR, was not a constitutional
decision, but only so far as brute force could go, by way of arbitrary
self-will without principles.
Therefore, the way forward is a principled one. My advice, such as it
is, to any electoral reform proposal, as that made by Forest, is that it
needs to be clear on the principles on which it operates, so that
politicians can be clear to the public, on the kind of election they are
being offered.
The incorporation of the Borda method in a recommendation involves the
first recognition (by Laplace) that lesser preferences count less.
This was in direct contradiction to the systematic elimination count
offered by Condorcet. The advantage of this pioneering debate was that
it offered a clear choice between all-inclusive weighted preference
count and exclusive unweighted preference count. Only the former offers
the Laplacerequirement to count preference order of importance.
One problem I had was that weighted Condorcet pairing, counting each
margin of victory, was about as accurate as Borda method, as one could
tell by inspection, even of a very narrow contest.
I have now come to the conclusion that weighted Condorcet pairing is
effectively (on average) a weighted preference order count, first
required by Laplace.
Each voters preference is not suitably weighted by order of importance.
Instead, the lack of weighted orders of preference for each voter is
made up for, firstly by the fact that all the votes (at any order of
choice) are equally or indiscriminately treated in this manner, but, on
average, a collective order of preference emerges, in the margins of
victory for each candidate pairing.
However, what might be called Condorcet implicit preference weighting
has its limitations. It does not reflect personal orders of preference
for personal representation. And it changes with who votes in the
electorate.
Never the less, weighted Condorcet pairing, to some extent vindicates
(collectively) weighted order of preference. But in doing so, it
reinforces the case against unweighted elimination counts. Hence
weighted Condorcet pairing may be approximately as accurate as the Borda
count, both in principle and in practise for the single member case.
An advantage of a Binomial count, a rational count for the exclusion,
equally as well as the election, of candidates, is that it can apply to
single-member as well as multi-member constituencies. (I.e. a binomial
count offers the consistency of a general theory of elections.) The
necessary practical consideration is thatall the preferences (including
abstentions) are counted, so it is known how much the voters wish to
elect or exclude the candidates. The consideration of principle is to
introduce and uphold the law of conservation of (preferential) information.
Regards,
Richard Lung.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20231015/9e183e75/attachment.htm>
More information about the Election-Methods
mailing list