Surely you have thought about using a condorcet method.
(Condorcet = if any candidate would will all 1 on 1's, that candidate must win)
During last year's NYC mayor's race the three strongest candidate were Adams, Garcia and Wiley.
It looked quite possible to me that Garcia would be eliminated in the penultimate round and hand final victory to Adams even though Garcia could be the condorcet choice.
Surely the whole idea of ranked choice is to produce a condorcet winner ...
P.S. If Garcia had been knocked out by Wiley, and Adams went on to win, we would have been drowned in editorials using the word "condorcet" - maybe such an event is what is needed?
I would pose the question in the following quantitative manner: Let’s define a “true positive“ as an election that successfully elects the Condorcet winner. Then we can calculate the rate of true positives. The closer it is to 100%, the less need there is for implementing true Condorcet voting.
Next we can ask what affects the true positive rate. Is it higher for federal elections which we would expect to be more polarized and therefore have variation along a single dimension of ideology? is it affected by public awareness of minor candidates?
Finally, there is the question of false positives: ways in which bad actors can manipulate the system. My the standing is that Trump told Palin supporters to not list Begich as a second choice. How much does that distort the process?
Finally, one can get back towards your question and ask what system would maximize true positives and minimize false positives? Is it practical to implement?
Sorry, some dictation typos which I will fix when at my desk. I meant to say one can evaluate different voting systems to see how they perform. Then ask what is practical to implement.
And, keep in mind that we need to ask: is this multi-round knock-out process actually easier to implement than just condorcet?
I think you make an excellent observation that in D vs R federal elections the knockout approach is almost always gonna get you to condorcet.*
But in single party democracies, like NYC we shouldn't expect that.
*and, I have a vague feeling that you over-focus on federal D vs R when a lot of the daily experience in places like NYC is in the hands of state and local office holders - consider the current congestion charge battle ...
Edited to add: Actually, I don't think I believe that condorcet thing about federal elections either. I think if we had a ranked choice general election then a moderate candidate could easily win by being everyone's second choice. Consider, for example, a Trump / Biden / Liz Cheney election.
Thanks for the excellent analysis.
A question on ranked choice voting:
Surely you have thought about using a condorcet method.
(Condorcet = if any candidate would will all 1 on 1's, that candidate must win)
During last year's NYC mayor's race the three strongest candidate were Adams, Garcia and Wiley.
It looked quite possible to me that Garcia would be eliminated in the penultimate round and hand final victory to Adams even though Garcia could be the condorcet choice.
See https://en.wikipedia.org/wiki/2021_New_York_City_mayoral_election#Results_by_round
So, it is far from a theoretical concern.
Would the public have trouble understanding?
Would it be too hard to do the IT work?
Surely the whole idea of ranked choice is to produce a condorcet winner ...
P.S. If Garcia had been knocked out by Wiley, and Adams went on to win, we would have been drowned in editorials using the word "condorcet" - maybe such an event is what is needed?
These are very interesting questions.
I would pose the question in the following quantitative manner: Let’s define a “true positive“ as an election that successfully elects the Condorcet winner. Then we can calculate the rate of true positives. The closer it is to 100%, the less need there is for implementing true Condorcet voting.
Next we can ask what affects the true positive rate. Is it higher for federal elections which we would expect to be more polarized and therefore have variation along a single dimension of ideology? is it affected by public awareness of minor candidates?
Finally, there is the question of false positives: ways in which bad actors can manipulate the system. My the standing is that Trump told Palin supporters to not list Begich as a second choice. How much does that distort the process?
Finally, one can get back towards your question and ask what system would maximize true positives and minimize false positives? Is it practical to implement?
I am working on all of this.
Sorry, some dictation typos which I will fix when at my desk. I meant to say one can evaluate different voting systems to see how they perform. Then ask what is practical to implement.
And, keep in mind that we need to ask: is this multi-round knock-out process actually easier to implement than just condorcet?
I think you make an excellent observation that in D vs R federal elections the knockout approach is almost always gonna get you to condorcet.*
But in single party democracies, like NYC we shouldn't expect that.
*and, I have a vague feeling that you over-focus on federal D vs R when a lot of the daily experience in places like NYC is in the hands of state and local office holders - consider the current congestion charge battle ...
Edited to add: Actually, I don't think I believe that condorcet thing about federal elections either. I think if we had a ranked choice general election then a moderate candidate could easily win by being everyone's second choice. Consider, for example, a Trump / Biden / Liz Cheney election.