Would Instant Runoff Voting Be Better For Voters?

[thebrokenladder]

Score Voting is simpler and vastly superior. Approval Voting is almost as good as Score Voting, and even simpler.

It is simply untrue to say that people can safely vote sincerely (non-strategically) with IRV. If your favorite candidate isn't a frontrunner, then sincere voting has a negative expected value. E.g. in the 2009 mayoral runoff in Burlington, a group of Republicans got a worse result for voting Republican. They could have gotten their 2nd choice instead of 3rd, by ranking 2nd as 1st.

[Mutex50]

IRV is better than plurality, but it still isn't a very good system. It gives you very unintuitive results and candidates can still be spoilers. Also, IRV is a pain in the ass to implement.

The best solution is to have an open primary with approval voting and allow the top two to runoff in the general election. Approval voting can use the same voting machines we use now. It is the most pragmatic solution.

[jeff4justice]

@Mutex50 What is approval voting?

Also how is any candidate ever a spoiler if that's who people cote for?

Next isn't what you're talking about with top two in a general the same as Prop 14 in CA?

How will a top two be better when it limits choice?

[Mutex50]

@jeff4justice Approval Voting is where you can vote for any number of candidates. It stops the problem of vote splitting. The runoffs in Prop 14 still has vote splitting and the two worst candidates can go off in the general election.

In IRV, candidates can still be spoilers because in early rounds a candidate that is generally popular can be knocked off for someone who is popular only within his own in-group. This can result in a bad candidate(not the worst) getting elected.

[jeff4justice]

@Mutex50 Sounds like top two meets IRV. Has it been done?

[thebrokenladder]

@Mutex50 Thanks for explaining this. I actually had coffee with Richard last year and explained this to him, but apparently it didn't stick. That's pretty much par for the course with IRV advocates. Which kind of makes sense, since in order to be an IRV advocate, you must have already ignored the massive amount of evidence favoring other systems.

[Mutex50]

@thebrokenladder I used to be an IRV supporter for a long time. I had to reevaluate that after someone showed me the Yee graphs and I looked more closely at the inter-workings of IRV.

Overall, I think the ideal solution is the Condorcet method, but I think we'd need more data on how real voters tactically vote to conclude that for sure. Right now, approval voting is the most pragmatic solution.

[Mutex50]

@Mutex50 For just online voting, I think the best solution is to rank the candidates in order of preference and also score them all. The Condorcet winner would runoff with the score-voting winner. That would probably be a good defense against tactical voting.

[thebrokenladder]

@Mutex50 I started as a Condorcet supporter, but then I found out about Bayesian Regret, the "one right measure" of voting method performance. Score Voting is even better, and Approval is usually better. Also there is a theorem that, given plausible models of voter strategy, they will be better Condorcet methods than ACTUAL Condorcet methods.

ScoreVotingDOTnetSLASH BayRegsFig

ScoreVotingDOTnetSLASH AppCW

[zippy1avion]

Richard Winger is the best!!! He has a voluminous mind of knowledge regarding ballot access rules and its history in the US.

[jeff4justice]

@zippy1avion Amen. He was an awesome person to chat with.

[DuckSmuggler1]

@zippy1avion dude omg your the bassist from nirvana dude you my hero oh my god oh my god like i love you soo much your like a legend i have poster of you all around my bed room thank you for help make my life ok with your music your music saved my life thank you soo much thank you soo much krist <3