Overstating RCV's benefits?

[spencerwilson]

Hey, thanks for the tool! I had two questions about statements in the README: https://github.com/artoonie/rcvis/blob/7fafd134f8ed895ab48602d4c1e744f8574da161/README.md

Lower rate of spoiler effect

In an RCV election, you can't spoil votes. Third-party candidates don't waste votes. Similar candidates help each other instead of hurting each other.

Are the results from Alaska's August 16, 2022 election a disproof of these statements? According to analysis of the ballots, Palin acted as a spoiler: if you "delete" her from ballots, Begich (a candidate ideologically closer to Palin) would have won.

Maybe the statements should be tempered, from absolutes to something statistical, like:

In an RCV election, the spoiler effect occurs less frequently.

I personally don't have any studies to cite to back that up, though.

From a theory point of view, the relevant criterion might be favorite betrayal. RCV/IRV doesn't satisfy it, meaning we can't say "It never hurts you to rank your favorite candidate at the top". Other analysis:

If at least 5,200 (Palin > Begich) voters instead ranked their second favorite (Begich) as first, then they would have gotten their second favorite candidate rather than their least favorite. That’s because Palin would have been eliminated first causing Begich to beat Peltola in the next round.

Instead, Palin voters got their worst outcome because they honestly ranked Palin first.

Higher rate of consensus-driven candidates

Another statement in the README:

[RCV elections are] less polarized and more fair.

The Alaska election also indicates that Peltola (the progressive, winner) would have been hurt if she'd taken some top-rank positions from Palin. If that'd happened sufficiently such that Palin was eliminated in round 1, then Peltola would have lost in her match-up with Begich. I believe this is a demonstration of RCV/IRV's non-monotonicity. It feels relevant here since future candidates might look to this election and think "If Peltola had tried to be more of a consensus-driven candidate, she'd have lost. I'll learn from her near-miss and stay near my end of the ideological spectrum, away from the middle."

What I'm not sure, though, is to what degree this Alaska result is representative or an anomaly. Maybe something like "It can result in less polarized candidates" would be more accurate? Again, I don't have any studies here to cite.

[artoonie]

Thanks for the feedback! Fully agreed, the current text overstates the benefits. I'll look into toning down the language and adding some qualifiers to these statements.

Appreciate your time in making this Issue!