The fact that equal ranks are valid seems to be a notable, easily explained advantage of Condorcet over IRV. As Australia jailed Albert Langer for the crime of advocating assigning equal ranks to candidates, perhaps voters there would be especially receptive to switching to Condorcet.
"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."
If there were a close race between two candidates A and B, with other candidates trailing far behind, then the best strategy would be to allocate all votes for whichever of those two is preferred, regardless of how much the voter might like C, D, etc. In other words, the 'wasted vote' problem does not go away; rather, it applies just as forcefully to each of the several votes that the voter now has to spend.
"But if you were to try to use it in a situation like our Australian Senate elections (with dozens of candidates on each ballot), the number of choices to make would place a burden on the voter. You could use Condorcet in the presidential ballot and preferential (IRV) in bigger ballots..."
Condorcet and IRV both use the same style of ranked ballot, so the 'number of choices' would be identical. The difference is in how the votes are tabulated, and in how the winner is determined.
"That said, more recent statistical analysis (which I can't find right now, but there are some papers on the web) that takes into account the current structure of political power in the United States shows that, in fact, the net effect of the electoral college is pretty close to zero."
A working paper by Katz, Gelman, and King comes to mind. (Also, another analysis of voting power under the Electoral College system by Gelman and Katz, as well as a critique of the applicability of the Banzhaf index in this case.)
'[I]f we want to be civilized humans, we have to go against these basic, animalistic instincts.'
'Is anyone else here thinking "Prisoner's dilemma"?'
In the iterated prisoner's dilemma, the potential for revenge is an incentive to cooperate. In light of this, your assumption -- that revenge is inherently inimical to civilization -- does not obviously hold.
While hierarchical structure is not always appropriate, it can be very useful for certain things. For example, GCC's info pages contain a tremendous amount of information, ranging from machine-specific flags to tips on porting GCC to new architectures. perldoc's subject-segregated organization achieves a somewhat similar effect.
While it is arguable that not all of this information needs to be available through a man-page-like interface, it is more or less congruent with the traditional notion of a manual, and can offer added structure where traditional UNIX man pages typically do not.
'Bottom line, if we have one year to live, it would be better for everyone if that last year were not spent in anarchy.'
I am reminded of the efforts of embalmers to enhance the appearances of corpses.
Certain individual reactions to situations of crisis may exhibit aspects that are undesirable when considered from a standpoint of non-crisis, just as a human individual's immune response to disease may make someone less effective at driving a car, or hauling water, or posting to Slashdot.
However, it is not clear that reactions to crisis should be judged from the standpoint of non-crisis.
Slashdot Mirror
The fact that equal ranks are valid seems to be a notable, easily explained advantage of Condorcet over IRV. As Australia jailed Albert Langer for the crime of advocating assigning equal ranks to candidates, perhaps voters there would be especially receptive to switching to Condorcet.
Given that IRV is non-monotonic (i.e., voting a candidate higher can cause them to lose the election), how is it better than what we have now?
If there were a close race between two candidates A and B, with other candidates trailing far behind, then the best strategy would be to allocate all votes for whichever of those two is preferred, regardless of how much the voter might like C, D, etc. In other words, the 'wasted vote' problem does not go away; rather, it applies just as forcefully to each of the several votes that the voter now has to spend.
Condorcet and IRV both use the same style of ranked ballot, so the 'number of choices' would be identical. The difference is in how the votes are tabulated, and in how the winner is determined.
Or am I missing something?
"That said, more recent statistical analysis (which I can't find right now, but there are some papers on the web) that takes into account the current structure of political power in the United States shows that, in fact, the net effect of the electoral college is pretty close to zero."
A working paper by Katz, Gelman, and King comes to mind. (Also, another analysis of voting power under the Electoral College system by Gelman and Katz, as well as a critique of the applicability of the Banzhaf index in this case.)
'[I]f we want to be civilized humans, we have to go against these basic, animalistic instincts.'
'Is anyone else here thinking "Prisoner's dilemma"?'
In the iterated prisoner's dilemma, the potential for revenge is an incentive to cooperate. In light of this, your assumption -- that revenge is inherently inimical to civilization -- does not obviously hold.
Um, who are you quoting?
While hierarchical structure is not always appropriate, it can be very useful for certain things. For example, GCC's info pages contain a tremendous amount of information, ranging from machine-specific flags to tips on porting GCC to new architectures. perldoc's subject-segregated organization achieves a somewhat similar effect.
While it is arguable that not all of this information needs to be available through a man-page-like interface, it is more or less congruent with the traditional notion of a manual, and can offer added structure where traditional UNIX man pages typically do not.
'Bottom line, if we have one year to live, it would be better for everyone if that last year were not spent in anarchy.'
I am reminded of the efforts of embalmers to enhance the appearances of corpses.
Certain individual reactions to situations of crisis may exhibit aspects that are undesirable when considered from a standpoint of non-crisis, just as a human individual's immune response to disease may make someone less effective at driving a car, or hauling water, or posting to Slashdot.
However, it is not clear that reactions to crisis should be judged from the standpoint of non-crisis.
Is using GCC not an option?