Slashdot Mirror
An Analysis of Various Election Methods
An anonymous reader writes "David Cobb talked about Instant Runoff Voting (IRV) as the best choice in electoral methods in his interview here, but is it really? The folks over at electionmethods.org seem to think it isn't. They favor Condorcet voting, which is another ranking style method using simulated one on one elections. Here is an evaluation of various methods, including IRV and Condorcet."
Much as we need a better system, it won't catch on if it can't be explained in one simple sentence.
If you must moderate, please moderate as irrelevent, not something bad, because I'm sure someone will find this interest
One mechanism I've not seen discussed is one I'll call a "voter economy". It probably has a real name, but it's not on that site and it seems like a reasonable system to me.
In this system, you get a certain number of votes (say 5x the number of candidates) and you can "spend" those votes however you like. So if you really like candidate A, you spend all your votes on A. If you like A a little, hate B, and would prefer C, you can spend 75% of your votes on C, 25% on A, and none on be.
This, to me, seems much better than ranking systems, since you can specifiy how much you prefer one candidate over another. It should be easy to explain, since people are used to the idea of spending.
Mathematicians, tell me whether or not this is a workable system.
The biggest problem that I can see with systems such as approval voting is that it is not non-repudiable. In other words, it would be impossible to verify that election results were not changed. A recount would not be able to detect changes made after a voter made his/her marks.
With a one voter, one vote system, it is easy to count the number of voters and the number of votes and ensure that the results were not modified.
I believe that this is a pretty important characteristic and I am a bit skeptical about who is pushing approval voting.
The reason that it can be true that 1+1 > 2 is that very peculiar nonzero value of the + operator
Actually the only thing I can't decide on is, which is the sillier idea:
I totally agree.
Anybody who advocates one system as more platonically better needs to read Arrow, but anyways, in my analysis, I prefer IRV/STV voting over Condercet voting, especially in multi-seat elections. Why? If all seats were chosen by condercet voting, all seats would be the kinda middle of approval. It doesn't provide for proportional representation, _at all_.
The multi-seat form of IRV, called Choice Voting (generally called Single-Transferable Voting (STV)), is preferable to Condercet if you aren't doing a straight party vote for bringing forth a diversity of representation. STV allows any minority group that can reach the election threshold (VotesTotal / (NumSeats + 1)) at least one seat of representation.
Further, in a representative system where there are multiple seats and they are all elected singularly, IRV would be preferable for the same reason (more likely to provide minority representation to increase the dialectic, because it heavily penalizes the person who can't get first place votes (if you got second place votes on all ballots, you may not win), giving third parties much more representation. In a single seat non-aggregate position (such as the Presidency), Condercet voting would probably be the best system.
However, we should all look back to Arrow's Theorem and remember that all voting systems are merely ways to reduce the input from direct democracy to a "managable" level for the elites, and thus they are flawed because OF COURSE they are losing data by "downsampling". Thus, if you want to really be heard, be active, get out, vote, be involved, write letters, run for office yourself and work to integrate real democracy, not just temporary dictatorships.
The two major parties haven't "stacked" anything. The current voting system was around from the beginning of the country, before either party existed.
Slashdot gets worse every day... Pipedot: News for nerds, without the corporate slant
In my experience, the only thing that electoral bias in favour of rural voters does is to artificially inflate farmers property values by turning them into into welfare recipients (in all but name), while indulging their worst tendancies to blame people who aren't WASPs for the world's problems and tell everybody else what they can and can't do in their own bedrooms.
The subsidy for American farmers works out to about $20,000 per rural job - yep, those salt of the earth folks you love so much have a huge proportion of their income paid by those city pagans. That's what the electoral college, and 2 senators per state regardless of population, does.
Any sufficiently advanced technology is indistinguishable from a rigged demo
--Andy Finkel (J. Klass?)
Arrow's theorem and its relevance to these voting systems is a much more complicated matter than it seems at first. For instance, one of Arrow's "reasonable requirements" is the Independence of Irrelevant Alternatives Criterian (IIAC), and it's been shown in many scenarios that failing the IIAC is actually what you want.
Condorcet fails Arrow's Theorem as do all other methods, but only when there isn't a Condorcet Winner. When there is, Condorcet is perfect. When there isn't a Condorcet Winner (like when there's a defeat loop, A over B, B over C, and C over A), then there are plenty of tiebreaker methods people can use that are "almost perfect". But in large elections, it's actually pretty rare that there isn't a Condorcet Winner.
So the Arrow argument isn't the smackdown that people take it to be.
skkkoooonnnggggkkk ptui