**INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR

Jack Waugh

I suggest that for evaluating any multi-round voting system, a helpful (although not comprehensive) step is to ask whether, when it eliminates candidates in a non-final round, it eliminates the correct candidates based on equality of political power from one voter to another. And your first round shares with choose-one voting the fact that voters do not get to weigh in on every pair of candidates, but have a limited count of coins to spend. Accordingly, your first round inherits the problems of choose-one plurality voting, with slight mitigation in some cases, since your voters have two coins and voters in a choose-one single-winner election only get one coin. The improvement could show up when there are few enough candidates. With seven candidates, it's not a real improvement. Someone having money support for advertising, or distinction by fame or infamy, will move forward into the final round, because the voters face the prisoner's dilemma.

psp_andrew.s

@cfrank I have done an immense amount of research and have yet to find significant flaws that would make 2CV a worse system than the currently proposed ones such as RCV/STAR/YNA etc. That said, I consider myself HIGHLY objective and I'm willing to admit where I'm mistaken. I'm open to discussion on it, preferably by voice / video chat, since I feel that would be more productive. I will however say that, in EVERY scenario I've tested the model, it DOES NOT allow minority rule by strategic voting. I can show you if you want to look at the model together. Add me on Discord @lamppost

@Jack-Waugh If 2CV were implemented in our own General Elections, it wouldn't matter what method Parties used in their party primaries. Parties are necessary evil, like it or not, since people WILL form groups/coalitions to collectively enact legislative change. Therefore, implementing 2CV would allow multiple Parties to thrive and be competitive SIMPLY by getting 2nd Choice Votes. This is a huge change and significant upgrade on our current voting system while being easy to understand and implement, and works within our existing government and electoral frameworks.

To address your concern - using 2CV in a Ralph Nader-esque scenario would have still allowed Gore, who was more BROADLY approved of, to win, which I believe is the desired result given the polling data. If you want to have a video/screenshare discussion, we can run through a 7-candidate model and you can tell me yourself if it produces the desired result. Add me on Discord @lamppost

And in your scenario of ABCDE candidates, if a A/C 1st and 2nd choice vote would "deny" candidate B the win, then a balanced countervote would be B/C 1st and 2nd choice, which would equivocally restore their win.

Jack Waugh

@psp_andrew-s I'm not suggesting there should be laws to try to prevent people from associating with each other in groups and endorsing, as a group, a candidate. I believe in freedom of association. However, a good voting system will make parties for purposes of nomination so redundant that eventually, even the thickest-headed will see that they are no longer needed for that purpose.

In the Frohnmayer balance test, it's not allowed to assume the voters can know who the frontrunners are so they can tailor their vote. It's necessary that for every vote that can be cast, a single antivote will suffice, regardless of the other votes, to cancel that one.

Jack Waugh

@psp_andrew-s said in **INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR:

it wouldn't matter what method Parties used in their party primaries

If that's the case, why would anyone want to join? If the method used is inadequate, joining would not add to the member's political power.

What would prevent just as many candidates running in the general as would have, had there been no parties? How is it that parties are part of the solution to your system's vulnerabilities? Why do you oppose Approval Voting, which is so much fairer and also simpler (fewer rounds)?

cfrank

@psp_andrew-s I did already explain why this does not occur in the present system. Also I explained why the hypothetical example I gave is fairly reasonable, even though it is idealized.

I think your statement about “promoting sufficient candidate diversity” is also confused: what I mean is, how could, in effect, restricting the number of candidates who can run promote diversity? Even three shades of red has categorically more diversity than just one shade of red. I understand that certain candidates may be “crowded out” by a more homogeneous population of candidates, but what are the odds that a unique candidate gets supported by a large political party as opposed to one of the swath of other potential candidates in a system that suffers from vote splitting? It happens, but it’s rare. Why not let that unique candidate run on his own and argue his merit rather than screening him out from the beginning for the sake of a tactical response to an unnecessary conflict of interest?

Something to understand is Duverger’s law, and the forces that lead to the formation of a small number of large, competitive parties. One major factor, arguably the most important factor, is vote-splitting. Majoritarianism is another. “Patterns of Democracy” by Arend Lijphart lays out the distinction between majoritarian as opposed to consensualistic democracies, in some respects the problem always goes deeper but it’s a start.

I don’t believe that instantiating my example as described will fail to yield the outcome I mention without additional tacit and potentially inaccurate assumptions. One of these tacit assumptions I think I gleaned briefly in a table you had produced attempting to illustrate the computation of my example, which is that virtually everybody in each of the two parties has exactly the same order of preference. If you allow for more diversified preference profiles, I do think you will see the problem.

For the record, STAR voting also fails independence of clones. And if by “RCV” you are indicating “Ranked Choice Voting,” you may actually mean “Instant Runoff Voting” (IRV), which is a very specific and in some respects a particularly pathological example of the more general body of “ranked choice” systems.

Also I think this assumption illustrates correctly that your system is conceived of as operating within a one-dimensional political paradigm. This is the paradigm we currently face, but it is arguably a suboptimal consequence of the current two-party system in the U.S., and of the party-coalitions system in the U.K. (Duverger’s law in action). In other words, with an alternative system of representation in place, the political landscape may very well change to become multi-dimensional and issues-focused rather than one-dimensional and values-focused, if my meaning is clear.

SaraWolk

@psp_andrew-s said in **INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR:

Simple and easy to explain to prospective voters (THIS IS IMPORTANT)

Thank you all for taking the time to comment. I agree with most of the concerns about this proposal that people have already raised. I raised a few of them myself in the Twitter Space where I first heard about it, on Twitter, and also in Open Democracy Discussion.

To summarize a few.

I don't think the proposal is simple to explain, implement, or use, nor are the implications or relative benefits transparent.
A better voting method needs to be able to prevent vote-splitting and the spoiler effect and this does not.
A better method should work in a variety of political landscapes and this does not. That should include crowded fields, primaries, nonpartisan or partisan elections, single and multi-winner races, and many more variables. The stress test elections that are truly competitive are the ones where the system matters the most and failure in that scenario is not acceptable.
A better method should not magnify "electability bias" and by extension magnify the influence of money in politics. This one does that because with the limited choice, there's a strong incentive to only support frontrunners, not underdogs.
Since first hearing about this method last week I've seen many people spend many hours sharing constructive feedback and patiently explaining some of the central issues with it. Rather than hearing that feedback, Andrew, this conversation seems to be going in circles, so my central concern is that this is derailing in nature rather than contributing something novel and beneficial to the conversation.

psp_andrew.s

@sarawolk

We can address these.

Our research has shown that a 2-choice voting system is the 2nd easiest for voters to understand when compared with STAR and RCV. We can do some more research and even conduct live, representatively sampled polls with your assistance if you're interested in finding out which option lay-people feel is easiest to understand.
A better voting method does not need to completely eliminate vote splitting. In fact, any system that forces voters to make any kind of explicit preference decision will incur vote splitting. That's not an issue as long as the systems sufficiently reduces the IMPACT of vote splitting, which 2CV does to negligible levels.
You've given no evidence to support this. Every scenario that we've tested 2CV in, which includes a variety of political landscapes and candidate quantities has produced "desirable*" election outcomes. Please give some examples where this might not be the case.
2CV does not magnify electability bias compared to plurality voting. It may have slightly more bias than unlimited-choice systems, but unlimited choice methods introduce either gamesmanship (in the case of unlimited-choice Borda) or significant complexity (normalization of scores in Score voting). Most voters only have 1 or 2 preferences anyway, as every election results study has shown, and eliminating all data from any system on rankings of 4th choice or lower almost never has a material impact on the election outcome. 2CV asks voters for enough data to create a statistically sufficiently informed single-winner choice (assuming 250k+ voters, representatively sampled).
I've heard all the feedback. I've listened and taken it into account, which led to modifying the runoff election so that no candidate can win without 50% support (either a 1st or 2nd choice from the majority, if none of the top3 get 51%, we drop the 3rd place candidate by points and upgrade all the their voter's 2nd choice votes to 1st choice votes). 2CV now ensures, unlike any other proposed system, that the winner will, under all circumstances, be one who received at least 51% of 1st or 2nd choice votes. Which leads to majority-consensus election outcomes, always.

* We can have a discussion as to what "desirable" means, but for our purposes it means that the winning candidate had a significant amount of 1st choice preference among the voting population, as well as acceptance/preference from among 51% of the entire voting population.

cfrank

@psp_andrew-s said in **INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR:

In fact, any system that forces voters to make any kind of explicit preference decision will incur vote splitting.

Andrew, this doesn't make sense. Ranked pairs operates solely upon explicit preference decisions in the form of a rank ordering of candidates, and it satisfies independence of clones, meaning that there is no vote splitting at all.

We can have a discussion as to what "desirable" means, but for our purposes it means that the winning candidate had a significant amount of 1st choice preference among the voting population, as well as acceptance/preference from among 51% of the entire voting population.

By your criterion of desirability, choose-one is sufficient. You do see that, right?

A better voting method does not need to completely eliminate vote splitting... That's not an issue as long as the systems sufficiently reduces the IMPACT of vote splitting, which 2CV does to negligible levels.

Again, I have given an explicit example where 2CV enables minority rule due to vote splitting. As I repeat, if you instantiate my example as described, including varied preference profiles rather than a monotone one-dimensional spectrum, you will see what is in fact unnecessary to see, since I already proved it mathematically: 2CV enables minority rule due to vote splitting. That certainly is not, as you say, reducing the impact of vote splitting to negligible levels.

Jack Waugh

@psp_andrew-s said in **INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR:

acceptance/preference from among 51% of the entire voting population.

Otherwise, the election is thrown and new candidates sought?

SaraWolk

@psp_andrew-s

2CV now ensures, unlike any other proposed system, that the winner will, under all circumstances, be one who received at least 51% of 1st or 2nd choice votes.

A voters 2nd choice may be as good as their favorite or almost as bad as their last choice. There's no way to know, so ensuring a majority of 1st and 2nd choice votes is meaningless. Also, in any election where you can support multiple candidates there could be multiple majority supported options. The key is to find the one with the most support by looking at strength of support and/or number of voters who prefer them, ideally both.

I think we've already gone in circles about your other responses in previous conversations so I won't repeat that again here.