[EM] Binomial STV counters Irish strategic voting.

[Richard Lung]

To STV voting and to Election methods groups.


The Electoral Reform Society survey of the 2016 Irish general election 
describes a strategic voting practice, by which allied candidates seek 
to bolster the first preferences of their least prefered colleagues, to 
prevent their early exclusion.

My invention of (abstentions-inclusive keep-value averaged) Binomial STV 
would do away with the need for this insincere voting. It does this by a 
minimum of two complementary counts: an election count and an exclusion 
count. The latter is a rational count, in its own right, conducted on 
the voters preferences in reverse order, instead of an arbitrary 
exclusion, when the transferable surpluses run out, in the election count.

To ensure that the exclusion count is not given undue importance, 
compared to the election count, all preferences are counted, including 
abstentions, which are generally at the end of the ballot papers, when 
voters cease to express a preference.

Hence it is possible for the abstentions to reach a quota, in which 
case, a seat remains unfilled.

It is possible for consistently rational counts, both for election and 
for exclusion, by extending the Meek method use of the re-adjustable 
keep value, to candidates in deficit of a quota, as well as in surplus 
of a quota.

Each candidates election keep value and exclusion keep value, inverted 
to provide a back-up election keep value, are averaged to arrive at 
deciding keep values.

This describes only the simplest first order Binomial STV, corresponding 
to the first order of the binomial theorem, consisting of just two 
terms: one election count of preferences and one exclusion count of 
unpreferences.

But it is also possible to have a second-order Binomial STV, based on 
the four combinations of the second-order binomial theorem. And so on.

An example of how first and second order Binomial STV work is given here:

http://www.voting.ukscientists.com/Binomial_STV.html

The example is drastically over-simplified, for a system requiring 
computer programming, as does Meek method. Unlike Meek method, all the 
abstentions are counted, so there is no requirement to reduce the Droop 
quota, as the preferences run out. Otherwise, the way to code Binomial 
STV is to start from the Meek method program and adapt to the modified 
rules, extending the use of the keep value, and so forth.

I am looking for some organisation that might take up this work of 
implementing Binomial STV and running preliminary trials.

-- 
Richard Lung.

E-books (mostly available free or reader-sets-price)
http://www.voting.ukscientists.com/colverse.html
Includes the series of books on:
Democracy Science (starting with electoral reform and research);
Commentaries (literature and liberty; science and democracy);
Collected verse (in five books).

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20160405/4bddb688/attachment.htm>