[Foundation-l] elections

Robert Horning robert_horning at netzero.net
Sat Apr 7 18:02:47 UTC 2007

Christopher Hagar wrote:
> Instant run-off voting is a bad idea and fails some basic requirements for a
> good voting system. For example, you can have situations where: if you rank
> a candidate _higher_ on your ballot, he will actually do _worse_ in the
> election than if you had ranked him lower.
> Ranked ballots are good, but instant run-off is a bad method for evaluating
> the ballots; a better way to do it is with the Condorcet method, which is
> what Debian uses. See also http://en.wikipedia.org/wiki/Condorcet_method and
> http://www.debian.org/devel/constitution#A (section A.6)

As I pointed out, ever single voting system has its flaws, although in 
the particular case I would tend to disagree with you in the 
particulars.  If you rank a candidate lower, they are more likely to be 
eliminated in the algorithms that cull out the weaker candidates.  It 
really is to your advantage to rank candidates according to your 
preference.  Of course there are also multiple algorithms that can be 
considered even in an instant run-off election.

I don't want this to turn into a holy war over various voting methods, 
but I would like to call attention that systems other than what has been 
done in the past for the board may not be the ideal method of choosing 
members.... particularly when we are now in a situation where there will 
likely be many very strong candidates running for multiple positions.  I 
expressed my preference, but certainly other suggestions are not only 
going to happen, but I openly encourage them to be made here.  I would 
hope that a reasoned discussion of how the actual voting process has 
worked out in the past and what changes could be made might occur.

-- Robert Horning

More information about the foundation-l mailing list