Comment by cylinder714 | Hacker News Reader

Comment by cylinder714 | Hacker News Reader

Parent Full thread

Here's a more detailed article about the situation in Alaska:

https://electionscience.org/commentary-analysis/rcv-fools-pa...

Note that's it's from a group promoting approval voting as an alternative to ranked-choice voting, so it may be biased, but the details look correct at first glance.

My impression of RCV is not good, beginning with Jean Quan's poor turn as mayor of Oakland some years back. I'm sympathetic to those who'd like an alternative to first-past-the-post, but any system that allows no one's first choice to win seems broken.

I'm interested in approval voting because it's simpler—no algorithmic folderol—and it seems even more fair to third parties than RCV: vote for whichever candidates you find acceptable, regardless of party, and whoever gets the greatest number of votes wins, period.

A Republican who doesn't like Trump could vote for, say, Ron DeSantis, Marco Rubio and Tom Cotton, while a progressive Democrat might mark down Bernie Sanders, Elizabeth Warren and Jill Stein (assuming an open primary). I see this as a way for independents and moderates, too, to make a bigger impact, as they could pick moderates from multiple parties.

I don't see any obvious downsides. Thoughts?

EDIT: A link to "The Center for Election Science," advocating for approval voting:

https://electionscience.org/

I agree that Approval Voting seems better in almost every respect to RCV/IRV, and (therefore) much better than FPTP. For that reason I would encourage more states to experiment with Approval Voting so we can gather some more data about what sort of results and UX it produces in practice.

The reason I mention UX, though, is that I think there is one respect in which Approval might be lacking, namely how voters interpret the instruction of "vote for whichever candidates you find acceptable". Have you ever stopped to think what "acceptable" means in that sentence? More importantly, have you asked a wide range of different people what they think it means, to check they all interpret it the same way?

That's a rhetorical question, but my point is that some people (based on their personality, or the advice from their party, or misinformation that they are exposed to) might think that "acceptable" means "I can tolerate this candidate winning, because they're not quite as bad as the worst outcome", while other people might interpret it as "this candidate is practically as good as my first choice candidate (but I don't like his haircut as much)".

This is somewhat similar to the "disenfranchisement" issue I mentioned above, where some voters are able to cast a more-powerful ballot by carefully calibrating the threshold at which they "approve" of a candidate. There's also the potential for people regretting the reform, if they discover that, for example, millions of anxious Democrats "approved" of DeSantis out of fear that Trump might win, when actually Bernie could have won if they had just stayed loyal and believed in Bernie's chances of success.

>Have you ever stopped to think what "acceptable" means in that sentence?

No, but I don't know how that would affect outcomes given a mass of voters. My sense is that it would all "even out" because no one would be forced to choose a single candidate, or to rank their favorites.

It would only "even out" if the different understandings of "acceptable" didn't have any correlation with people's political preferences. Given that we know there are personality traits that are strongly indicative of party membership[0] (not to mention education level), I think the default assumption should be that different parties will be affected differently by voters interpreting the question differently.

Also, all aspects of voting systems have to be considered in the context of an adversarial environment, and you have to imagine that some parties will be more unscrupulous than others at exploiting voters' confusion, or just better at using polling data or other strategies to instruct their voters how to make their votes more powerful.

I suppose you could argue that we already accept that tactical voting is a reality under FPTP, and maybe in practice the situation under Approval wouldn't be any worse over all (although it might be better for some parties and worse for others), but even if it is better, you have the problem that any change adds to the "attack surface" of things that election losers can point at to cast the validity of the results into doubt.

That's why any reform needs to ideally involve voter instructions and election outcomes that are more intuitive than FPTP, to counteract the potential for additional misinformation to be introduced.

[0] https://practicalpie.com/personality-and-politics/

>when actually Bernie could have won if they had just stayed loyal and believed in Bernie's chances of success.

They still can. I think approval voting eliminates the need for tactical voting because each persons' vote has equal weight. A vote for DeSantis only improves his chance of winning in your case, it doesn't harm Trump, because in the end, the candidate with the most votes wins. There's no algorithm that throws away votes in a particular round, as we see in RCV.

> A vote for DeSantis only improves his chance of winning in your case, it doesn't harm Trump

But imagine that 3 Bernie supporters all feared that most Republicans would only approve Trump, meaning Trump would beat DeSantis, so "just to be safe" the Bernie supporters decided to approve DeSantis (as well as Bernie), leading to DeSantis beating Bernie by 1 vote (whereas Bernie would have won if they didn't do that).

In this example, the 3 extra approvals for DeSantis might have harmed Trump, if he had been 1 approval ahead of DeSantis (and 2 approvals ahead of Bernie) before these Bernie supporters cast their votes.

Reply on news.ycombinator.com