Sat 28 Apr 2018 9:41PM

compare STAR with strategic-primary voting…?

AW Aaron Wolf Public Seen by 23

Okay, I'd like help thinking this through:

I'm looking at the Democratic primary here. Let's say for discussion I prefer someone else to the front-runner incumbent (Kate Brown). If the front-runner is going to win anyway, I can just vote for symbolism. But what if enough support could actually knock out the front-runner? Then, I start worrying about which candidate is most likely to beat the Republican who I probably would not support… of course, I can ignore that concern if the Republican is sure to lose to any of the Dems.

How does STAR relate to this?

Kill the primary, nothing's closed by party… so, now is there a scenario where my worries remain?

Say Kate Brown will defeat the Republican but the other candidates might actually lose (say because some of Kate's voters lean toward Republican as their second choice over the other Dems, and Republican base is strong enough to be competitive).

How unlikely is it really that one of the other Dems knocks Kate out of the automatic runoff in STAR but the Republican stays in the runoff and wins, whereas Kate would have won the runoff if she'd made it in?

Seems obvious that this worry scenario requires a close election, but still, that could happen, right?


Emily Dempsey Sat 28 Apr 2018 10:15PM

I think it also requires a Condorcet Cycle.

You are worried about there being too many voters who express this preference:

Kate > Republican > Democrat

While you yourself feel:

Democrat > Kate > Republican

I’m pretty sure that in order for the Democrat to beat Kate into the runoff but still lose to the Republican, there also have to be a large number of voters who express:

Republican > Democrat > Kate.

As far as I’m aware, there’s not much data available about how likely Condorcet Cycles are in real-world elections, so shrug but it does seem like no voting system would really be able to handle it very well - who even is the “right” winner in that case?


Aaron Wolf Sat 28 Apr 2018 10:30PM

Seems on the right track, but I'm not so sure.

What about the more obvious Republican > Kate > Democrat group? If they are near-majority, they only need a small number of Kate > Republican > Democrat group with them to elect the R if Kate is out of the runoff.

Is there a non-cyclic way for Kate to miss the runoff that she would win if she got in? Maybe not really.

Okay, so I guess the concern is this:

Republican > Democrat > Kate may seem unlikely, but what if all or much of the R base switches to that dishonestly because they feel confident that themselves plus the Kate > Republican > Democrat group is a majority.

In other words, they create a Condorcet cycle strategically so that the one candidate who beats their favorite in the cycle is out of the runoff…

I'm not sure I'm thinking this through right though.


Emily Dempsey Sat 28 Apr 2018 10:40PM

Yeah, that's possible. That'd be the Republucans employing a Burying strategy.

By Burying, they risk accidentally electing their least favorite choice if their favorite doesn't make the runoff.

That risk seems even higher than normal in this situation since yeah it seems like all candidates would have to be p neck and neck.

I think that you as a D>K>R voter are better off voting honestly in the hopes that either strategic Republican voters accidentally help the Democrat beat Kate in the runoff, or that enough Republican voters vote honestly to not keep Kate out of the runoff.