[Noisebridge-discuss] approval voting

[Leif Ryge]

For our upcoming Board of Directors election, I propose that we use
approval voting[1] instead of the Schulze[2] condorcet[3] method which
we used previously.

Condorcet methods are designed to select the candidate most preferred
by a majority of voters. Approval voting selects the candidate who is
acceptable to the largest number of voters. I think the latter is more
appropriate for Noisebridge.

Paraphrasing wikipedia:
> The Schulze method is a voting system that selects a single winner
> using votes that express preferences. The method can also be used to
> create a sorted list of winners. It is a Condorcet method, which
> means the following: if there is a candidate who is preferred over
> every other candidate in pairwise comparisons, then this candidate
> will be the winner when the Schulze method is applied.

> Approval voting is a single-winner voting system. The method can
> also be used to create a sorted list of winners. Each voter may vote
> for (or 'approve' of) as many of the candidates as the voter wishes.
> The winner is the candidate receiving the most votes. Each voter may
> vote for any combination of candidates and may give each candidate at
> most one vote.

It is worth noting that neither method is really meant for electing
multiple winners, but either can be modified for that purpose. With
approval voting, the multi-winner method is fairly obvious since the
candidates can simply be ranked by how many votes they received. With
condorcet methods such as Schulze, after a single winner is selected,
to select each additional winner the counting process must be repeated
ignoring votes for the previous winner(s).

What are the downsides to approval voting? As far as I can tell (please
correct me if I'm missing something!), they boil down to these two:
1) Sincere voters (those who honestly vote for every candidate who is
acceptable to them) can help their more-preferred candidate lose to a
less-preferred-but-acceptable candidate. (I think, in our case at least, 
that this is actually a feature.)
2) When strategic (insincere) voters' predictions of the outcome are
incorrect, their strategy can fail. For example, imagine the USA were
using approval voting in the 2004 presidential election. A voter who
would prefer Dennis Kucinich over John Kerry (but would prefer either
of them over George W. Bush) would logically cast votes of approval for
both Kucinich and Kerry if they expected that only Bush and Kerry were
"viable" candidates. However, if they mistakenly believed that Kerry
and Kucinich were both viable and Bush was not, they might
strategically withhold their vote for Kerry.

What is the downside to condorcet voting? Simply put, using a condorcet
method, a candidate who is unacceptable to 49% of voters yet preferred
by 51% will beat a candidate who is acceptable to 100% of the voters.

I hope that using approval voting (and encouraging sincere voting) we
will have enough candidates with 100% approval to fill the board, and
then we will have effectively consensed on an acceptable set of board
members! If we have more than enough winning candidates, in lieu of a
runoff election I suggest we could use a lottery and/or expand the size
of the board.

I'll be bringing this up for discussion at this week's meeting, and
hopefully we can get consensus about it at the following meeting.

~leif

1: http://en.wikipedia.org/wiki/Approval_voting
2: http://en.wikipedia.org/wiki/Schulze_method
3: http://en.wikipedia.org/wiki/Condorcet_method