[EM] FW: IRV Challenge - Press Announcement

[Raph Frank]

Also, there is an example given in the pdf.

42: C>B
39: A>B
10: B>A
4: B
5: B>C

(incidentally, B is the condorcet winner)

Anyway, the argument is based on the assumption that each rank should
be given an equal weight, i.e. everyone's first choice should be given
a weight and everyone's second choice should be given a weight.

This is how Borda counting works, with set values given for each rank.

However, STV is based on moving ballots between piles.  The ballot
always has a value of one and it applies that value to the pile that
it is part of.

In round one, the ballots are divided into 'piles' based on the first choice

'A' pile -> 39 ballots
39: A>B

'B' pile -> 19 ballots
10: B>A
4: B
5: B>C

'C' pile -> 42 ballots
42: C>B

B is eliminated, and in round 2, there are only two piles

'A' pile -> 49 ballots
39: A>B
10: B>A

'B' pile -> 4 votes remain/exhausted
4: B
Note: B now only gets 4 votes, as the other ballots have been transferred away.

'C' pile -> 47 ballots
42: C>B
5: B>C

C is eliminated and A wins.

The name single transferable vote is pretty descriptive.  Under the
system ballots are 'transferred' between the piles for each candidate
based on the ordering given on the ballot.

In PR-STV, ballots are can also be divided into parts such that the
sum of the weights of all the parts adds to one vote and again they
are transferred according to the ordering given on the ballot.