**INTRODUCING** 2-Choice Voting (2CV) - An Improved Iteration on RCV and STAR
I wanted to share with everyone a new voting system that not too many people are familiar with yet, but that combines some of the best parts of RCV, STAR, and Approval Voting, all into one simple, easy-to-understand voting system!
ELI5 (Explain Like I'm Five)
You have a 1st-Choice Coin (worth 2 Points) and a 2nd-Choice Coin (worth 1 Point). You can give each coin to a separate candidate that you support. We then tally to find out who has the most Points, and if that candidate ALSO has the number of Coins greater than 50.1% of the number of voters, they win! (i.e. the majority of voters gave them a 1st- or 2nd-Choice Coin aka “Majority Support”)
If the candidate with the most points doesn’t have Majority Support, we redo the vote, but with only the Top 3 candidates with the highest number of coins from the first vote (Top 3 Runoff by Support). Everyone votes again with the same method, and if none of the Top 3 received 51% Majority Support, we eliminate the 3rd ranked candidate and upgrade all of that candidate's voters' 2nd choices to 1st choices (doubling their 2nd choice point value, thus making them equivalent with 1st choice value). The winner is the person with the most points!
The Proposal and Model
I've created both a Google Docs Proposal that proposes and explains 2CV in depth, as well as a Google Sheets 2CV Election Model, where you can run up to 7 candidates with a separate 3-candidate runoff in the case of no candidate receiving Majority Support.
Potential Cost / Benefit Attributes
- Does not allow for voting of more than 2 candidates or choices
- most voters don't care to vote down ballot anyway, and those votes are rarely significant in their effect
- Requires a runoff election if no candidate receives Majority Support in the first vote (aka Primary election)
- not a big deal, we already do this, and doing so gives benefits to voters whereby they can uniquely express their preference of those candidates who make it into the runoff.
- Allows for "Majority Support" whereby the winning candidate will have either had 50% support or a
- Simple and easy to explain to prospective voters (THIS IS IMPORTANT)
- Provides for maximal candidate-choice expression by allowing voters to express uniquely in both the Primary and the Runoff elections
- Rewards moderate, populist candidates by allowing those with a fewer number of 1st Choice votes to win by making up the difference by getting a significant number 2nd Choice votes
- Encourages honest, exhaustive voting, since voters only get 2 choices, and their 2nd choice helps prevent the election of their less preferred choices
I'd love to hear everyone's feedback on this one, but please read through the Proposal first and play with the Model to see if it either confirms or satisfies your concerns
Looking forward to seeing if this thing holds up to everyone's much-appreciated scrutiny!
EDIT: Responses to Criticisms
Possibly Requires 2 Rounds of Voting - We already do this, and it's absolutely not a big deal. In fact, I would argue that having a completely separate 2nd round of voting for a Runoff is a BENEFIT and actually INCREASES accurate voter expression. A voter should have the ability to CHANGE their preference once they know who the runoff will be between. E.g. a Voter didn't do a lot of research on a candidate who they ranked low on their STAR ballot, but it turns out many others liked that candidate and they went into a top-2 runoff. Now that voter becomes disenfranchised because they didn't know that candidate would be in the top 2 and may want to do additional research to change their vote, now that they know. You cannot expect voters to adequately research ALL candidates on the ballot, so when there is a runoff, those voters should be able to reevaluate and recast their votes. It actually INCREASES accurate voter expression.
Loss of Expressive Ballot / Only 2 Choices - 2CV gives voters double the expression of our current system, which is already not horribly insufficient. I think many of you think that MAXIMUM voter expression is required from a system; it's not. All that's required is enough expression whereby the system will produce a desirable result given the level of expression, which should be VERY SERIOUSLY CONSIDERED AND WEIGHED against the system's simplicity, as simplicity and ease of use/understanding matters FAR MORE for the average voter if trying to convince them to adopt a new voting system. 2CV provides significant, adequate expression to produce the desired result in the VAST majority (~95%) of electoral simulations. This is a mild tradeoff in favor of simplicity, which is much more important than I think many on here consider. Voter understanding of the system is probably the highest consideration when trying to implement a new voting system for our government.
Voters Forced to Allocate Positive Support - At no time is a voter REQUIRED to cast their 2nd choice vote. However, it is in their benefit to do so, as a 2nd choice vote helps prevent the election of that voters less preferred candidate. It's an optional "insurance" but not a requisite.
Practical Consequence Unclear - Not sure if you read the whole proposal (Which I requested people do before making criticisms), but this is addressed in the Proposal, which I additionally expressed in the "benefits" section. Among them are:
- Easy for voters to participate optimally, without gamification - just pick your favorite and a backup
- Significantly less complicated than other proposed alternatives (STAR, RCV)
- Rewards populist candidates by creating a simple system whereby a candidate can win with fewer "primary preferences" that are made up by a significant number of 2nd choice votes
On Frohmayer Balance - 2CV passes this challenge, to my understanding. E.g. If there are candidates ABCDE and 1 voter selects A/C as 1st/2nd choice, another voter can cast E/C as 1st/2nd choice to create balance between A/C and E/C preferences. The effect of which would be that candidate C would gain additional support from both voters, but that is still EQUAL in weight and thus satisfies Frohmayer, to my understanding. Likewise, if a voter ONLY votes A, another voter can choose to ONLY vote E to create balance, as 2nd choices are optional, not required.
- Does not allow for voting of more than 2 candidates or choices
@psp_andrew-s hi, and welcome! Here is my scrutiny, I hope it’s sensible and please forgive any apparent bluntness, as I am trying to evaluate the system objectively:
One major weakness of this system is that it very possibly has two separate rounds of voting, which in my limited understanding is indeed a big deal, maybe others can weigh in on that.
Second, it seems odd to me that a voter should in some circumstances be forced to allocate positive support to a candidate they may prefer not to endorse. I would say this concern is related to the inability to indicate indifferences between one’s top 3 choices (as well as the inability to indicate preferences among the 3rd ranked and lower-ranked candidates).
Third, while the concept of the system is perhaps fairly simple (which can also be argued against), neither the statement of it nor its motivation is easy to digest immediately, and the practical consequences of using it are not obvious. For this I would recommend providing clearly motivating and illustrating examples—but to this point, and lastly, investigating such examples shows that it suffers from vote splitting, aka it fails independence of clones, which is a more serious violation. Independence of clones I think is a formal necessity, I hesitate to speak for the consensus of forum users here but I think this point in particular has very limited controversy.
For example, say there are 8 candidates A,B,C,D,E,F,G,H, the first five of which are clones from party X and the last three of which are clones from party Y.
If P is the fraction of voters in Y, then what is the expected score of A?
Well, for each voter in X, there is a 4/(5 choose 2)=2/5 chance of A being given a coin, and the coin A may be given has an expected value of 3/2. Therefore the expected score (per voter) of A is (3/5)(1-P).
For F, we have a similar story, but this time the chance of being given a coin is 2/(3 choose 2)=2/3, whereby the expected score becomes P.
But then if P>(3/5)(1-P), meaning that P>0.375, party Y can crowd out the three open slots. Will the candidates in Y have the most coins?
Well they will have 2P/3 coins each, whereas the candidates in X will have 2(1-P)/5 coins. Again, the candidates in Y will have the most coins if P>0.375. Therefore we have minority rule due to vote splitting.
This is just one example and a statistical argument, but this means that parties and voters are keen to become strategic, since there is a non-negligible conflict of interest in how they nominate candidates and cast their ballots, and there are tactics to navigate these conflicts that are easy to implement. Ex: Nominate 3 clone candidates to crowd out the election, then bullet vote. Or bullet vote and then bullet vote again. As far as I can tell, these kinds of tactics can spiral into an adversarial mind-game. So it becomes, unfortunately, an effectively more convoluted FPTP despite its good intentions.
Overall I do not mean to condescend and it’s great that you’re considering alternatives, I think these points are avenues for you and others to explore other thoughtful voting systems and build a strong foundation of comprehension in this area
masiarek last edited by masiarek
My main objection - loss of expressive ballot.
Plurality voting is among the least expressive voting methods there is available.
2CV method seems to force voter to express opinion on two candidates only (lost expressiveness of the voting ballot).
You may very well have opinions about some or all of the candidates on a ballot, and yet you only get a say about two candidates only.
Other voting methods allow you to express yourself in different kinds of ways — for example:
- approval voting allows voters to choose any number of candidates;
- RCV lets you rank candidates by preference;
- score or range voting asks voters to give each candidate a score, with the winner being determined by the candidate with the highest total or average score.
- STAR Voting - multiple candidates - both score and preference !
I suspect that this system fails Frohnmayer balance. In my opinion, that leaves a crack in it into which capital can insert its wedge so as to prevent control of politics by humans.
Here's an alternative that also fails balance, but by way of illustration, I suspect it would provide more human control than your proposal, and maybe be a bit more resistant to attempts to confine voters in prisoner's dilemmas. I'm not seriously proposing this for implementation (because it requires too many rounds of polling and because I believe we must insist on balance).
For each round of polling and tallying, each voter is provided two coins. One is worth -2 points and the other is worth 1 or -1 points (two variants). The voter gives each coin to one or zero candidates. The candidate earning the lowest score is eliminated from the rest of the rounds. When only one candidate hasn't been eliminated, elect that candidate.
@cfrank I've responded to you main criticisms in the edits above! I'll have to dive deeper into the vote splitting hypothetical, but my guess is that you won't often have "Clones from 1 party" since most parties would only allow 1 candidate to have their party's endorsement. So this hypothetical of "clones" seems, at face, extremely unlikely, especially considering no 2 humans are exact clones. There will almost certainly be SOME differences between the "clones," so I feel that the premise is likely flawed. But I'll humor the hypothetical and dive in and reply more in depth
@jack-waugh I've responded to your criticism of not passing the Frohmayer Balance test in my edit to the main post. I think this system does satisfy it, since there is the ability for any voter to bring balance to another voter's ballot cast in an equally weighted and balanced opposing vote, which would many times result in the "common ground" candidate being more highly supported, which is a benefit of the system - rewarding candidates with higher overall support.
@psp_andrew-s even approximate clones will cause vote splitting, such as the Bush-Nader-Gore election. A clone is a formal voting theoretical concept, where two candidates appear adjacently in preference rankings among (very many) voters. Furthermore, parties can and will strategical nominate likely clones if it’s in their interest to do so.
@cfrank I think 2CV satisfies this. Would you be available for a Discord video call to discuss and run it through the Election model that I built? You can add me on discord @lamppost
@psp_andrew-s no it does not. The example I gave demonstrates that it fails independence of clones.
@cfrank I've read through your hypothetical several times and some things don't make sense - not sure I'm understanding the variables correctly. Are you available to discuss via screenshare/video chat so I can better understand?
@psp_andrew-s unfortunately I am not available for a video chat at the moment, but I can elaborate later about the example. The key point is that the coins that get distributed over larger groups of similar candidates will cause those platforms to weaken, and that this enables minority rule.
@cfrank Ok, I think I kind of understand where you're going with that, I'll think hard on it and let you know what I find. I'm also updating the model to support 8 candidates to run it through.
I will say, however, that I think the hypothetical is flawed, as I can't think of a scenario whereby 5 candidates from 1 single party could, or ever have, competed in the same general election as 3 candidates from ANOTHER party. Typically the POINT of parties is to align around 1 candidate, so those parties will have their own Primary before selecting whom to choose as their Party's candidate in a given general election.
Please advise and let me know if my criticisms of your hypothetical are valid or not.
Edit: I think your hypothetical requires the running of MIXED Primaries, which I neither support nor think are a good idea. But we can get into that separately if you genuinely believe that mixed Primaries are preferred to our current unmixed systems.
@psp_andrew-s the reason we don’t see multiple candidates running on similar platforms is precisely because the “vote for one” system we currently employ suffers severely from vote-splitting. So platforms that run with more than one candidate are essentially doomed from the start, which is exactly what happened in the infamous Bush-Nader-Gore election. The primaries really just function to strategically navigate the failure of independence of clones. In a system that satisfies independence of clones, there is no intrinsic reason for many candidates with similar (read as: appealing largely to the same group of voters) platforms not to run for office. What I’m saying is that 2CV will probably converge to 2 large parties nominating 3 clones each, which is only nominally different from what we already have.
EDIT: I may have misunderstood, I think your hypothetical assumes that there are 50/50 weighted voters for liberal vs conservative, right? I'll readjust and look at it.
@cfrank I actually STRONGLY disagree with your premise that having clones in a general election is desirable. I think it's actually highly UNDESIRABLE and that we should want our general elections to offer us REAL CHOICE, meaning true, tangible, distintinct differentiation of candidates and ideas. No healthy election should only offer a choice between 3 shades of Red and 5 shades of Blue.
The party system is good and works, so long as it allows for alternative parties / moderates to win by getting a large amount of general approval (even if they're not 1st choice of the highest number of voters) - I think we can all agree with this premise.
That's the problem with our system - not that it doesn't promote the running of candidate clones (that's actually a GOOD thing, in my opinion), but rather that it doesn't allow for sufficient voter expression whereby they can choose a 2nd or backup candidate to gauge candidate acceptance / support more broadly than Single-Vote can indicate, and allowing for the victory of more populist candidates, which 2CV does elegantly.
I've thought very hard about your hypothetical, and I've come to the conclusion that it's an impossible situation, since it would likely NEVER be the case in an 8-candidate Mixed Primary Election (which is undesirable in and of itself, in my opinion) that no candidate would attempt a strategy of populist appeal and be more "moderate," to gain 2nd Choice Votes, since doing so is a huge competitive advantage. So I don't personally see that as a valid criticism of 2CV, though I'm open to more discussion if you have counter-arguments
@psp_andrew-s I don’t get the impression that you are evaluating your own stance objectively, and I again do not mean to condescend, but it seems like you will need to do more research into this matter.
You are right that a healthy election should offer more than 3 shades of red against 5 shades of blue. What I am telling you is that I am very confident your system, if implemented, would converge almost always to 3 shades of red and 3 shades of blue.
@psp_andrew-s no, my hypothetical assumes that, for example, there are 38% in one party that nominates three clones, and 62% in the other that happens to nominate 5 clones. Which is, actually, roughly realistic, assuming a small degree of naivety among candidates and that a larger and more diversified platform encourages more candidates to run for office. Which it should.
@cfrank I actually STRONGLY disagree with your premise that having clones in a general election is desirable.
I don't think his point was that it was in general desirable or undesirable, but rather, that it should not be available as a strategy whereby a strong-arm party could prevent democracy.
On Frohmayer Balance
If there are candidates ABCDE and 1 voter selects A/C as 1st/2nd choice, another voter can cast E/C as 1st/2nd choice to create balance between A/C and E/C preferences.
These two votes do not balance, because in the context of the other votes (the ones in the election but not stated in the example) the first one could deny candidate B the win, and the second would not restore B's win.
Typically the POINT of parties is to align around 1 candidate, so those parties will have their own Primary before selecting whom to choose as their Party's candidate in a given general election.
Assuming that, first off, the voters, in order to have power, are required to organize themselves into parties, which I object to, and assuming that a given party conducts a primary election to determine its one nominee for the general, which I read you above as recommending, are you suggesting that the party is to use 2CV for the primary? That just pushes the problem from the general to the primary. Are the different factions of voters within the party supposed to organize themselves into subparties, and have a primary in each of those to determine their nomination for the party primary?