[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: RFD: informal proposal

On Sun, Nov 17, 2002 at 12:01:28PM -0500, Raul Miller wrote:
> So.. how does this sound:

Horrendously complicated was my first thought...

> Group options according to their supermajority requirements -- all
> options with the same supermajority requirement go in the same group.
> Use CpSSD to pick a winning candidate from within each group.
> Use CpSSD with scaled preferences to pick a winner from the set of
> winning candidates.
> Default the election if the picked winner didn't meet quorum.

A needs 2:1, B simple majority, D default option.

	3 ABD
	1 BDA
	1 DAB

group A = {A}, group B = {B,D}; winners are A and B. Scaled preferences
give: A defeats B: 4:(1x2), A wins.

> Does anyone see any significant strategies for inconsistency with this
> kind of mechanism?

It has the problem that it's much easier for a non-supermajority option to
win. If you have:

	A - remove non-free (2:1 supermajority required, say)
	B - handwave about the issue, don't really do anything
	D - further discussion

you might get results something like:

	60 A B D
	30 B A D
	10 D B A

with the groups being {A} and {B,D}. B defeats D by 90:10, so next we
work with {A,B} scaling A down, ending up with B defeats A 40:30, in
spite of a majority of developers wanting to remove non-free entirely.

Of course, that result's probably _stable_ -- the next vote would probably
result in D defeats A 40:30, but I don't know that it's entirely fair
to have B work as a "spoiler" option -- if B hadn't been an option,
A would have passed it's supermajority requirement easily.

Alternatively you could possibly end up with something like:

	60 S T D 
	50 T D S

where S and T both have a 2:1 supermajority requirement, and D
doesn't. The result is S beats T, 60:50, and D beats S 50:30, and
D wins. Given T was unanimously preferred to D, that seems like a
significant loss.


Anthony Towns <aj@humbug.org.au> <http://azure.humbug.org.au/~aj/>
I don't speak for anyone save myself. GPG signed mail preferred.

 ``If you don't do it now, you'll be one year older when you do.''

Attachment: pgpLznY09Pz8Y.pgp
Description: PGP signature

Reply to: