[EM] draws in IRV

[David Catchpole]

Well, it stands to reason that you should be able to eliminate as many low
ranking candidates as you can given that the aggregate of their vote
doesn't exceed that of the candidate with the next lowest number of votes.
So, in an election with first preferences...

A
35
B
34
C
15
D
10
E
3
F
3

C,D,E and F could all be eliminated in the first round without any change
in the result from a strict elimination. Have to say, I haven't come
across any 90% rule, though in Queensland, booth officers conduct a
"notional two-party-preferred" tally, where preferences between the two
notional leaders in a seat are counted. This is usually a useful
indication to candidates, the media and the central office that conducts
the final tally, though occasionally an upset occurs when an Independent
becomes a major contender.

If I can find the Queensland rules at a spare moment, I'll send them over.

By the way, my party (Labor) won about 65 seats in the (I think...)
89-member Assembly (we don't have an upper house). I was a bit delicate on
Sunday!

People might find the results interesting in analysis of IRV, particularly
the results from the west of the state, where a large number of sitting
Nationals (rural conservatives) have been replaced by Independents.

On Mon, 19 Feb 2001, LAYTON Craig wrote:

> There is usually random selection in the first or last round (I recall a
> coin toss being used in one election).  However, there is a little known
> rule of mass elimination in Australia, that eliminates all candidates except
> the top two in the first round if they poll a certain percentage of the
> valid vote (I think it is 90 per cent).  This is intended to save counting
> time, as single winner IRV elections are still cast and counted manually
> (the old fashioned way, writing numbers in boxes).
>
> In all ties except for the first and last round, you conduct a count back,
> where you go to the previous round.  The candidate of the two with the least
> votes in the previous round is eliminated (if they are tied in the previous
> round, you continue going back to the rounds before that).
>
> -----Original Message-----
> From: David Catchpole [mailto:s349436 at student.uq.edu.au]
> Sent: Monday, 19 February 2001 8:56
> To: martin at myreddice.co.uk
> Cc: election-methods-list at eskimo.com
> Subject: Re: [EM] draws in IRV
>
> Mostly, the rules call for a random selection.
>
> On Sun, 18 Feb 2001, Martin Harper wrote:
>
> > What is the correct way to deal with draws in IRV - ie, in an election
> > between A,B,C,D, where C and D tie for lowest numbers of first place
> > votes. Should both C and D be eliminated? Or should one be chosen at
> > random for elimination? Or is there some other way to tie-break them?
> >
> > Thanks,
> > Martin
>
>

-------------------------------------------------------------------------------
"You have the right to scream your head off. Should you give up the right
to scream your head off, someone who screams _their_ head off will be
provided for you."
	Grouch cop, "Elmo in Grouchland"