The "computer program" solution was exactly what was used in the London
Mayoral election last time around.

All the ballot papers were scanned (which took hours longer than it had
been expected to) and then a computer system worked out the combinations
of first and second choices before declaring Boris the winner.

The flaw of the whole system is its closed nature; I was there as an
election observer with the Open Rights Group and our main concern was
that no one had a clue what the computer was doing and whether the
result was correct. The views of the group ranged from the sane "it is
probably accurate, honest and right, but we don't know this for certain"
to the rather over the top "it's clearly a fiddle and a massive
conspiracy"! To be fair, though, the election committee did have three
independent individuals who were given all the raw data and allowed to
create their own algorithms to check the result.

One of the interesting points of the day was speaking to the Election
Commission observer who was an Australian (where they use a similar
transferable vote system, but count by hand). She declared that the
computers were pointless and, had this been an election in Oz, they
would have finished the count hours earlier and been in the bar for five
hours already by the time we get the result!

Ultimately, though, we never found out what the in-built flow charts and
algorithms were - possibly because this information is commercially
sensitive for the private business that ran the computers and scanners.

Daryl

On 14/04/2011 20:29, Ted Harding wrote:
> Greetings All!
> I have been reading (not for the first time) the booklet
> from the Electoral Commission which was dropped through
> my letter-box the other day. The description given of
> how "Alternative" votes would be counted seems to leave
> unresolved some possible scenarios.
> 
> Consider (as in their example) 4 candidates: A, B, C, D.
> 
> In the constituency of Chew Backie, the Number 1 votes
> were roughly equally distributed over A, B and C, all
> getting less than half. Candidate D was everybody's
> second choice, and nobody's first choice. So, on the
> basis of the "Round one" rules (page 5), there will be
> at least one further round of counting.
> 
> So, according to the "Round two" rules (page 6), the
> candidate with the fewest number 1 votes is removed
> from the contest -- in this case Candidate D.
> 
> NOW:
> ** Each ballot paper on Candidate D's pile is looked
>    at again.
> 
> BUT: Candidate D doesn't have a pile -- he got no
> "number 1" votes! So nothing gets added to A or B or C.
> 
> AND: Candidate D has now been removed from the contest.
> 
> BUT:
> ** If the ballot paper does not show a number 2 vote
>    it is no longer used.
> 
> At this point, in the absence of an explicit resolution
> in the booklet, there are (at least) three possibilities.
> 
> [1]: each ballot paper (all had D as number 2) is deemed
> to have a number 2 vote which is for D, in which case
> D is re-instated and gets a huge pile, and will probably
> win, since he now has a vote from every constituent and
> other candidates will progressively drop out by having
> least votes.
> 
> [2]: Candidate D having been eliminated in Round one,
> all votes for candidate D are eliminated. Now there
> is no paper with a number 2 vote, so all ballot papers
> are no longer used and the election is void.
> 
> [3]: Candidate D (in position 2 on all papers) having
> been eliminated, the candidates with "number 3" votes
> are promoted to "number 2" and the candidates with
> "number 4" votes are promoted to "number 3". We now
> have ballot papers with "number 2" votes which are
> added to the A, B and C piles accordingly.
> 
> Possibility [3] is compatible with the "Round three"
> rules (page 7), where (candidate C having been removed):
> ** Each ballot paper on Candidate C's pile is looked
>    at again to see if any of the remaining candidates
>    are ranked.
> ** If so, the ballot paper is moved to the pile of
>    the candidate ranked highest on that ballot paper.
> 
> However, that's not how it was stated for Round two.
> One suspects that this is what they mean for Round two
> as well. But, if so, why can't they damn well say so???
> 
> But even so there is a complication. The "number 2"
> votes (promoted as above from "number 3" votes) are
> now in the A, B and C piles, but (say) a "number 2"
> for B which is in the C ("number 1") pile can't be
> added to B's withnout removing it from the C pile
> (see also below).
> 
> Another interesting (though perhaps unlikely) unresolved
> situation is where the "number 1" votes are exactly equal
> across candidates. In that case all candidates get less
> than half, and *each* candidate has the fewest votes!
> So should they all now be removed from the contest??!!
> 
> Well, presumably that is not what is intended. I guess
> that they would then leave all the candidates in, and
> look at the "number 2" votes. But you can't now put
> a ballot paper with (say) A as number 2 onto A's pile
> without taking it away from (say) B's pile (where B
> was the "number 1" choice on that paper).
> 
> So how do you proceed in that situation? Make a copy
> of that paper and put it on A's pile??? I think things
> would then start to get out of hand ...
> 
> Alternatively (and it might be the wisest things to do
> with a procedure as complex -- and undoubtedly error
> prone -- as this) one might simply enter the lot into
> a computer and leave that to work it out. And that may
> well be the intention -- so bye-bye to the small army
> of vote-counters sitting at tables and sorting ballot
> papers into little piles by FPTP single votes, to be
> gathered up and summated at the end of the day. (And
> probably just as well with a system like this one).
> 
> But if there is a computer program for this, presumably
> it has a strictly defined algorithm, with flow-chart,
> to cover all possibilities. Why can't we be shown this??
> 
> Just a few thoughts! And I'd be grateful for comments
> from anyone who knows how it is really suppoed to work.
> The "information" sent out by the Electoral Commission
> is too noddy by more than half, and seems designed to
> soothe the uncritical who have no stomach for trying
> to pursue implications!
> 
> Best wishes to all,
> Ted.
> 
> --------------------------------------------------------------------
> E-Mail: (Ted Harding) <[log in to unmask]>
> Fax-to-email: +44 (0)870 094 0861
> Date: 14-Apr-11                                       Time: 20:29:13
> ------------------------------ XFMail ------------------------------
> 
> ******************************************************
> Please note that if you press the 'Reply' button your
> message will go only to the sender of this message.
> If you want to reply to the whole list, use your mailer's
> 'Reply-to-All' button to send your message automatically
> to [log in to unmask]
> Disclaimer: The messages sent to this list are the views of the sender and cannot be assumed to be representative of the range of views held by subscribers to the Radical Statistics Group. To find out more about Radical Statistics and its aims and activities and read current and past issues of our newsletter you are invited to visit our web site www.radstats.org.uk.
> *******************************************************

******************************************************
Please note that if you press the 'Reply' button your
message will go only to the sender of this message.
If you want to reply to the whole list, use your mailer's
'Reply-to-All' button to send your message automatically
to [log in to unmask]
Disclaimer: The messages sent to this list are the views of the sender and cannot be assumed to be representative of the range of views held by subscribers to the Radical Statistics Group. To find out more about Radical Statistics and its aims and activities and read current and past issues of our newsletter you are invited to visit our web site www.radstats.org.uk.
*******************************************************