[EM] PR-STV and vote management

[James Gilmour]

> raphfrk at netscape.net > Sent: 20 April 2007 19:17
> Optimal strategy in PR-STV is to rank all the candidates

This is optimal only if you genuinely have preferences among all the
candidates.  Voters should always be encouraged to mark as many
preferences as they have, but they should never be advised (or forced!!)
to mark preferences they do not have.

When a voter leaves some candidates blank (truncates), what that voter
is saying to the returning officer is: "If it comes to a choice among
the candidates I've left blank, I'm out.  I am happy to leave that
decision to those voters who do have preferences among those
candidates".

Of course, a voter may have some high order preferences, and a real
dislike for one or more candidates, but no preference at all AMONG the
others, but they are obviously below the high order and above the
disliked.   In that situation that voter should, of course, mark the
high order preferences, put the disliked ones last and put the
indifferent ones in any order just to fill the spaces and ensure a
continuous sequence of preferences.  I've done that myself in real
STV-PR elections!


> (actually all bar 1).

This makes no difference to the result, i.e. the winners.  But it can
make a difference to the appearance of the result sheet if the counting
rules specify that all votes must be transferred, i.e. even AFTER the
required number of winners has been identified.  Then the effect of a
'last preference' and 'last blank' is different  -  the 'last blank'
becomes non-transferable.  This is what will happen under the STV
counting rules we shall use for the local government elections here in
Scotland next month when the votes are counted electronically.  If a
manual count is employed (a contingency provision!), the transfers are
stopped when the all the places have been filled.  Personally, I think
transferring all possible votes after the winners have identified in
real public elections is completely pointless, but I didn't write the
election rules.  It was a case of "Because we can, we shall", which is
rarely a sound basis for good policy or good practice.
James Gilmour