I completely agree that this is better than nothing and some sort or ranking would be good, and that IRV is not great. But
Borda count is vulnerable to very bad tactical voting. People will put candidates who "can't win" above people who they like better but don't want to win. And when this goes wrong you'll have people elected who no one wants. See this chart, showing that fully tactical Borda is abysmal :http://electionscience.github.io/vse-sim/vse.html (from http://electionscience.github.io/vse-sim/VSE/)
If we're going to do rankings we should use something that meets the Condorcet criterion, but overall I think Approval is probably the best for its simplicity.
2nded for Approval. While it may be suboptimal at capturing preferences, it has the advantages of being easy to count, and easy to explain: whoever gets the most votes wins, end of story. Most importantly, effectively nothing changes for those who want to vote a straight major-party ticket.
I used to think this, and then I tried describing it to a room full of educated people -- some with advanced degrees -- and persuading them it was better than plurality for the system we were setting up for a nonprofit board.
It's certainly possible I was the point of failure, but everyone kept getting concerned about each person having multiple votes
"Like, what if someone cheats and gives all their votes to one person?"
"You can't do that, it's approval voting, you can cast one approval vote for each candidate"
"But that means each person can vote for multiple candidates!"
"Yes. Every person uses their ballot to indicate which candidates they find acceptable."
"But what if I like one candidate more than all the others?"
"Then you can either vote for them and everyone else you think would be OK, or just vote for them."
"But what if someone else votes for all the people I don't like."
"How is that different from people voting for a candidate you don't like now?"
"Well, they can only pick one candidate I don't like now."
I understand no voting system is perfect, but I was not expecting this one to be such a hard sell.
Interesting, I wouldn't have expected that either. I wonder if the "some people get more votes" concern disappears if reframed from "vote for you who approve", to "vote yes or no for each candidate". In a 7-candidate race, it's not that one person casts 1 vote, and another casts 4; both cast 7 votes (1-6 vs 4-3).
I suspect we're so conditioned to treating votes as exclusionary/zero-sum, the mere idea of one person ticking multiple boxes can easily trigger instincts for fairness / cheater detection.
I like the frame! Would love to have thought of that then, my intuition is also that it has a lot of potential to carry the idea, likely better than my various diagrams of tally marks.
This makes me wonder if the relative complexity of IRV makes it paradoxically easier to sell. People tuning out the details and deciding based on stories of past elections that could've been different, plus their impressions of the advocates.
I think Range Voting is easier to pitch. You vote each candidate on a 5-star basis similar to your ride sharing and delivery apps. As a bonus it collects more fine-gained data per voter.
The solution for having minority representation is to use some sort of proportional representation with a multiple-winner system. But, in elections that can only have one winner (i.e. for president, governor, mayor, etc...) there's no way around the risk that a majority of people have bad opinions. That's not a problem with the voting system, but with the electorate.
Or you do what parliamentary systems do and have the executive position be chosen by the elected representatives, and try to set it up so that minority representation still has to be taken into account to "form a government" as the Brits say.
This is a problem with democracy in general; while constitutional constraints theoretically limit the damage the majority can do, in practice "the fox guards the henhouse" with regard to reifying those constraints.
One can easily envision scenarios where the milquetoast centrists defeat stronger positions that would otherwise have won in a two-party, lesser-evil scenario; on the other hand, it could just as easily break the other direction.
I'd argue that incentives towards consensus building are net wins for minority voters (in all senses of the word), because it puts the maximum number of voters in play for each candidate. Right now, there is a very real sense in which Texas Democrats and California Republicans have very poor representation; 49% may as well round to zero. (Proportional representation is also worthy of consideration here.) Putting those votes in play for third parties and independents makes candidates more responsive to the whole electorate rather than chasing a small number of swing voters.
Even if I manage to gain 70% support of voters, it will all be for naught if my opponent opens an even bigger tent at 75%. It shifts us away from thinking of politics as a zero-sum game.
Approval voting is too similar to FPTP. Enough voters will continue to mark a single box that you're not better off in practical terms. Major candidates will aim to muddy the waters because they won't want "leakage" to third parties.
Even if 70% of people Bullet vote Approval still has a very high Voter Satisfaction Efficiency. Its simplicity makes it very robust to tactical voting.
VSE is based on simulation and gee whiz, seems to favour those awfully enthusiastic range voter folks who assume we're all going to do some calculus in the voting booth.
IRV has actually been used. Thousands of times. At multi-mega-votes scale.
For what it's worth the people behind VSE primarily advocate for Approval Voting: its very simple while performing almost as well as much more complex systems
I agree, IRV actually provides psychological reinforcement for ranking all the candidates - I'd rather voters over vote than under vote... and we can assume a lot of poor voters since voter education (and general education) is a serious issue in the US.
The proximate problem in US elections is the spoiler effect, and Approval solves that as effectively as any other system. We won't have to have endless debates about whether Nader voters "would have" voted for Gore. We'll know for sure.
People's votes aren't spoiled because they will always approve of their preferred candidate and the singular mainstream candidate. Because of this, one of the two mainstream ones will always win. So there's basically zero chance of the non mainstream candidate to win, even if a majority of people would pick them as their #1 choice.
> People's votes aren't spoiled because they will always approve of their preferred candidate and the singular mainstream candidate. Because of this, one of the two mainstream ones will always win.
This is demonstrably untrue.
Minor parties are represented in every legislature in Australia, including in single-member lower house electorates.
If we're going to do rankings we should use something that meets the Condorcet criterion, but overall I think Approval is probably the best for its simplicity.
A resource I like: https://www.lesswrong.com/posts/D6trAzh6DApKPhbv4/a-voting-t...