[EM] British Colombia considering change to STV

[Raph Frank]

On Wed, May 6, 2009 at 12:44 PM, Kristofer Munsterhjelm
<km-elmet at broadpark.no> wrote:
> There's still some information leakage. For instance, if there had been just
> one vote-seller, the elided information:
>  100: A > C > D > B > E
>    1: A > C > ***

You could require that each row has at least 50 votes.

If there is only 1 vote seller, then it isn't really that big an issue.

The row with the fewest votes would be combined with another row.  I
am not sure how to do that fairly.

Would it be acceptable with the above to just convert it to

101: A>C>D>B>E

In effect, your vote is added to the row which matches the N highest
ranks, where N is the highest possible.

Alternatively, the rule could be that there will be a guaranteed slot
for bullet voting for each candidate, all other rows require at least
100 votes to be activated.

Thus the above would be converted to

100: A>C>D>B>E
1: A


> Another approach, which would be party-based, would be where the voters vote
> for local candidates, and the system completes the ballot by ranking all
> other candidates of that party ahead of the rest.

Right, this is party list with override, but restricted to local candidates.

> If you do something like this, it would be relatively easy to have "fluid
> districts", where everybody that's within a certain distance of the polling
> place gets on the local ballot, unless that would clutter it unduly.

Right, assuming the problem of recording so many ballots is resolved.

In principle, each candidate could be given a code, but there is also
a method to get ballot access on a polling station by polling station
basis.