I would prefer you set the out points as "no support" and "strong support," or something similar, rather than "love" and "hate." Words matter.

Range Voting is a defective voting system whose proponents openly reject majority rule in favor of "social utility efficiency."

While this is an interesting experiment, the results of this poll won't tell us much about politics in practice.

A preferential system - kind of like the second choice polls that Hotline runs - would tell you which candidate had the strongest initial bases of support, and then which candidate had the broadest compromise support.

I would prefer you set the out points as "no support" and "strong support," or something similar, rather than "love" and "hate." Words matter.

Well, sure, but these are just "labels" which have nothing to do with the actual scores.

If we could dispense with the elections and just let me personally pick the next President, it would be Kucinich. My take on some of the rest....

Gore: Didn't vote for him in 2000. His stock rose with me after i finally saw the film.

Obama: Not in love with him yet. When i was getting my first impression of him he was commenting on gay marriage. ROUGHLY paraphrased, his religious convictions just don't allow him to support it yet. WTF! You're running for President, not Pope! What really bothers me, is that i think he's lying. I would much prefer a candidate say candidly that he/she believes marriage shouldn't be a special right, but that pragmatically, it won't be high on the agenda.

Clinton: I still love Bill, damn his eyes. Hillary gives me the "typical politician" vibe, but that may just be the result of the constant media bashing. My mom, a Republican, hated her. And for no good reason, i might add. But the rabid right was so successful in demonizing her. After Dennis, she might be my next pick, just because of the "in your face, republicans!" effect.

Edwards: Didn't impress me in the last campaign. Benefits recently from warm/fuzzy news coverage.

Bottom line: I think it a fabulous field. I WILL vote for the eventual D nominee, which i haven't always in the past. Compare this to the R field, and i feel soooooo good.

The candidates I would vote for - Mike Gravel and Dennis Kucinich and maybe Bill Richardson - don't have a chance of being elected president. So the real question probably is, "Who will the Democratic party put forward as the lesser evil for president?" Because my memory of history extends beyong the average American's two-week time span, I'm not persuaded that Al Gore (Bill Clinton's deputy) is our knight in shining armor; although, I wouldn't rule out the possibility he might have made a Damascene conversion since departing the White House and will give him a 4 or 5. Anyone who voted for the war gets a zero from me. Obama appears to be more of a motivational speaker and probably a front man for some backroom dealers. I'll give him a 2 or 3.

I like this.

Two add on ideas.

Why not try both the Instant Runoff Voting and Range Voting on the same list of candiates with the same voters. Then we could compare the results.

Why not do either or both on an expanded list of possible Oregon Democratic US Senate candidates.

As far as voting effectively, we needn't look farther than France, who just completed their first round of Presidential elections last Sunday. They had twelve candidates participating for votes in this first round, from every aspect of the political spectrum. They had an 85% voter turn out. Since neither of the top two candidates won 50% of the vote, they both go into a second round election in two weeks. Politically speaking, this is what a true democratic system looks like.

LiberalIncarnate makes a good point. That is a politically realistic voting system that privileges majority rule without completely sidelining third parties.

But that French two-round system (TRS) is designed for a political field with two strong parties presumed to always end up in the runoff. When three or four are ideologically similar, it's possible for them to divide support so much that two less preferred ideologies end up in the runoff. A misfire, as happened when Le Pen, with merely 17% of votes in the first round, proceeded to the runoff at the expense of a leftist. Le Pen's vote total did not increase past 18% in the runoff round.

The same may happen when France holds its runoff presently.

A better system is instant runoff voting, which simulates a series of runoffs - not just one runoff - until a majority winner emerges. Eliminates are sequential, not in batch as with the French system. In most cases, that solves the spoiler problem - especially the kind that emerges in two-round systems.

Anyone who voted for the war gets a zero from me.

I agree wholeheartedly. Any politician who voted for the war based on political calculation, inability to understand the evidence, or just plain stupidity doesn’t deserve to sit in the oval office. And I put no stock in after-the-fact apologies concerning how they voted and whining about being “misled” by the Bush administration.

That should winnow the field.

I would whole-heartly welcome the idea of instant run-off voting. I love the idea of having many choices and having these choices remain in the race longer if possible so that their ideas can be heard.

In the meantime, I will watch the French elections dreaming of sitting in a cafe, sipping an espresso and talking politics.

i just want to say that i'm really bothered that this voting exercise wasn't structured exactly how i preferred. next time, please try to read my mind and make it better...

thanks.

i wound up roughly ranking the candidates, but no one who voted for the war got more than a 5.

lestatdelc,

agreed about kucinich, except at least in the debates he is good at bringing up things other candidates are afraid (or wary of) to address. i kind of feel the same way about edwards on the iraq war vote--he's saying most of the right things right now, but i can't take it seriously.

it'll be interesting to see if 2007 has kucinich posters plastered all over lawns throughout southeast portland, as in 2004, or if people have moved on.

As far as voting effectively, we needn't look farther than France...this is what a true democratic system looks like.

Well, no. They use an extremely poor voting method - plurality with a top-two runoff.

Condorcet, Borda, Approval, Range and many many other methods are massively better.

http://RangeVoting.org/vsi.html

Ugh. I can never quite get why people like Kuccinich.

I agree with your points, lestadelc, but I would like to see Kucinich achieve more exposure. He does rock the boat, and if ever a boat needed rocking it is the current ship of state. As for flipping on issues, that charge can be laid at the feet of almost every candidate, including Hillary. Fortunately, for Kucinich, he won't be elected; otherwise, the oligarchs of both parties and their satraps would gang up on him just as Tip O'Neill and his buddies in both parties ganged up on Carter.

In addition to giving a zero to candidates who voted for the war, the same score applies to candidates in the pocket of AIPAC and Israel's Likud party. That probably means all candidates except Gravel and Kucinich.

i just want to say that i'm really bothered that this voting exercise wasn't structured exactly how i preferred. next time, please try to read my mind and make it better...

Range Voting actually comes extremely close to the same social utility efficiency you would get if you picked the winner based on reading the voters minds and picking the most utilitarian choice.

http://RangeVoting.org/vsi.html

I've been able to converse with a few presidential candidates, past and present, and Kucinich is the only one who impressed me with his/her combination of awareness and straightforwardness. I have no problem with his change of position abortion rights. It may have been announced just before he announced his candidacy, but that does not mean it was something other than an honest reappraisal. Dennis has more more integrity than most of both houses of Congress put together.

The greatest problem with his candidacy is the deliberate sabotage of the corporate press of him and anyone else who is a serious threat to corporate hegemony.

His years of being anti-choice then suddenly becoming pro-choice a month before he announced his candidacy for President in 2004 is all you need to know about the guy.

If your point is "to hell with anybody who has disagreed with me on any single important issue, even if they become convinced they were wrong," then you might as well toss your own hat in, since you'll likely find issue with every candidate.

If your point is "he only changed his stance out of shallow, political expediency," then please market yourself as the incredible mind reader that you must be.

Again, there's no top-tier candidate on the left side that i can't find fault with, 'cause i'm just so damned spot on, perfect. But i'll be cheering come November 2008 when one of them is elected President, regardless of who it will be. Moreover, i'll be working hard to elect that person, first trying in vain to effect the D nomination, then in the effort to elect that nominee.

Contrast that with the dogs on the other side of the aisle. Why does the perfect always have to be the enemy of the good in our tent?

Range Voting is a defective voting system whose proponents openly reject majority rule in favor of "social utility efficiency."

This isn't a flaw, once you understand the logic behind it. Social utility efficiency is economics speak for greatest voter satisfaction. Why would voters want to be less happy with election results just so the "majority" could always have its way? Most people realize they often won't be in the majority (especially when you consider primaries).

Majority rule is based on the false belief that candidate A is a better candidate than candidate B, if more than half the people prefer A to B. But this axiom is easily disproven by a simple example:

% of voters, their order of preference 34% A > B > C 34% B > C > A 32% C > A > B

66% of the voters prefer A to B, so the Majority Axiom says A is a better candidate than B. 68% of voters prefer B to C, so the Majority Axiom says B is better than C. So if A is better than B, and B is better than C, then obviously A is better than C, right?

Well, not according to "majoritarianism", since 66% of voters prefer C to A. So your "majority is best" belief contradicts itself, and is obsolete.

Our voter satisfaction metric doesn't fail like this. With Range Voting, if A has a higher average rating than B, and B has a higher average rating than C, then A obviously has a higher average rating than C.

While this is an interesting experiment, the results of this poll won't tell us much about politics in practice.

Since you don't support this argument, I can only imagine that you are implying the following: since this is a poll, and voters gain nothing from the outcome, they will not be strategic, like many voters would be in real life. To this I respond by simply pointing out that Range Voting degrades far less under strategic voting than systems like IRV - half as much compared to IRV specifically.

A preferential system - kind of like the second choice polls that Hotline runs - would tell you which candidate had the strongest initial bases of support, and then which candidate had the broadest compromise support.

Range Voting is a preferential system, which not only allows you to specify the order of your preferences, but the intensity of them. It's possible for a voter to say that he likes candidate A a little bit more than B, or a lot more than B.

I suppose that what you meant to say is "a rank-order" system. But rank order systems

1) Produce less satisfying results than Range Voting 2) Are generally more complicated (more work to use and tally) than Range Voting 3) Almost always lead to two-party domination, largely because they... 4) Usually give voters a strategic incentive to "betray" their sincere favorite candidate in order to get the "lesser of two evils".

So rank-order methods are clearly inferior, and full of problems, which make them unsuitable for use in political elections.

A misfire, as happened when Le Pen, with merely 17% of votes in the first round, proceeded to the runoff at the expense of a leftist. Le Pen's vote total did not increase past 18% in the runoff round.

The same may happen when France holds its runoff presently.

A better system is instant runoff voting...

Wrong. IRV has lead to two-party domination in all four countries where it has seen long-term widespread use. But most of the 27 countries that use a genuine top-two runoff have broken free of two-party domination.

http://rangevoting.org/TTRvIRVstats.html

If there's one thing I can say about IRV, it's that it survives off these kinds of unfortunate myths.

We have good choices in the candidates at present. They represent the cross section of the party and the current policy thinking of the party. They are mostly progressive in outlook. I think Hillary's unfavorables make her not as strong as Obama, Edwards, or Richardson.

Re Bill Bonner: "...candidates in the pocket of AIPAC and Israel's Likud party."

Why is it that some people are unwilling to acknowledge that most of these candidates essentially agree with AIPAC and "the lobby", and see Israel as a "strategic asset" that has served its imperial master well?

The theory that they are somehow "forced" by those conspiring Jews to act in opposition to their own strong moral underpinnings is not a progressive one. Mearsheimer and Walt are not progressives, and neither are you.

May I just say that it drives me nuts when I hear Hillary Clinton referred to as "Hillary", and the other candidates referred to as "Edwards", "Obama", etc. I hear journalists doing it constantly. IMHO it's demeaning and sexist.

That said, I do not think Clinton is electable, so do not support her candidacy. I admire Obama, but think he could use another 4 to 8 years before he's ready for the job. Edwards isn't a bad choice, but am concerned his personal life may get much more complicated than he realizes (this from a breast cancer survivor). But bottom line, any of the dems would be an huge improvement over what we have to deal with now. Hell, most of the republicans would be a relief!

That said, I do not think Clinton is electable, so do not support her candidacy.

Exactly the reason that systems like plurality and IRV almost always lead to two-party domination. You don't just care about whether YOU like a candidate, but whether OTHERS will like him, because you don't want to waste your vote.

With Range Voting, this problem is eliminated, so you truly can show as much support as you wish for your sincere favorites - because it doesn't diminish how much your scores for other candidates help them.

i"His years of being anti-choice then suddenly becoming pro-choice a month before he announced his candidacy for President in 2004 is all you need to know about the guy."

Correct me if I'm wrong, but didn't Al Gore also do a 180 on the pro-choice/pro-life issue? That is, he actively pressed for legislation to restrict abortions before he "saw the light" and decided it was more politically advantageous to pander to feminists. If so, guess you wont be on board with Big Al either when he officially announces his candidacy.

I too am rather unimpressed when a candidate appears to change his spots a little too easily on some of the more hot-button issues. Talk about unscrupulous politicians. Seems to me that something like abortion warrants a rather solid, unwavering position regardless of which side you happen to come down on, much like the second amendment vs. gun control. For example, if a putz like Chuck Schumer suddenly began speechifying on the rights of gun owners and voted accordingly, he'd appear to have even less integrity than he does now. Not that I'd mind, but I'm never going to like, trust or respect the man.

I think Hillary's unfavorables make her not as strong as Obama, Edwards, or Richardson.

If she gets the nomination and runs against Chuck Hagel he would have a fair chance of coming across as the lesser evil despite the negatives he has.

It's hard to keep up with Clay Shentrup's attacks on IRV ("IRV elects extremists!" "IRV leads to two-party domination" etc), but in the meantime, it's proven extremely effective in all kinds of elections in accommodating voter choice. Sure, it's a majoritarian system, because we're talking about voting methods to pick one winner, but "darkhorses" and independents get a fair shake, and win if the frontrunners/major parties falter.

Check out www.fairvote.org for the latest news on IRV around the country, including a strong editorial endorsement last week from USA Today and fun new examples of IRV on campus elections -- IRV is now used by a majority of colleges ranked in the top 30 by US News and World Report.

Interesting to see range voting tried out here, although I suspect some voters will figure out the tactical incentives created by the system that put "insincere voters" at an advantage over sincere ones. Thanks to this site for continuing to showcase different voting methods.

My first vote for President was in 1972, and circumstances had me in a work uniform, taking a bus home and already hearing the election called for Nixon after 4 pm Pacific, and nevertheless I was incredibly proud and frustrated to cast my mark for McGovern.

Casting negative comments on the candidates because they do not reach our own personal level of issue and emotional perfection really is a disease of the left, I've had it as bad as any in the past.

This time I am really ready to see the back of the Republicans. Obama would be fine, Edwards would be fine, Joe SixPack registered Democrat would probably be fine, even though none of them probably come up to my own level of political and spiritual philosophy. Could I vote for Hilary, who I like personally but detest politically? I probably could vote for her, but it would be tough to work for her (and I am someone who has worked in the last 3 presidential campaigns).

Yet the challenge for me and you and all of us, if Hilary or whomever fails to meet our level of expectations on fair trade or anti-imperialism or whatever, is to create our movement and make our argument and our backers so irresistible that even Hilary will have to come on board our train, as opposed to seeing our casting a vote for a (less-favored) Dem in Nov 08 (such as Hilary would be to me) -- which is really being done defensively to keep a Repub. out -- yet is interpreted by the less-subtle thinkers as being my "approval of" or even "allegiance to" Hilary.

Good job Kari, though I'd like to see the other poll run also - for comparison's sake.

I cannot and will not support Sen Clinton, I didn't mind Hillary, but this iteration offends me as a lefty Democrat, completely. I might give her a vote rather than a Republican, but I sure wouldn't like voting for the lesser evil with this field available.

You fans of the French model might do well to read French history. No thanks.

While we are considering potential presidents, let's not forget those candidates who voted to suspend habeas corpus when they voted for the Military Commissions Act.

Re: Hillary being called Hillary and not Hillary Clinton or just Clinton---it may be her wish, but it is a mistake. People looking for someone to lead this country, especially in such turbulent times as these are looking for the ideals of strength and forcefulness. Right or wrong, words do matter and so do names/titles. If Hillary runs as Hillary and wins, will she be President Hillary or President Clinton? Hillary for President makes it sound like she's running for Student Council...not President of the United States of America.

RE: Hillary. She may choose to call herself Hillary (or rather her marketing people may choose that as part of their election strategy and I agree it's a mistake), but professional journalists should not. They can certainly report on her decision to use her first name, but should use the same salutation form for each of the candidates, be it last name only, first and last, title, etc. It is sloppy and biased (possibly thoughtless) journalism to do otherwise. I cannot recall another presidential candidate that was refererenced by their first name only by the press.

The use of "Hillary" is not new -- the press used it while she was the First Lady. Can you imagine journalists (and I'm not talking about tabloid writers) referring to Barbara or Nancy or Eleanor? I've always suspected the "Hillary" thing was started by Republican handlers as a subtle way to diminish her credibility -- they so resented having an educated, intelligent, opinionated woman as First Lady. We (dems) should insist she be addressed in the same manner as the other candidates, even if she foolishly does not.

1.“Love”/“hate”: I agree with the first post (Jamie). I know it’s on the survey to identify the far end of the ranges. But several posters have discussed their “love” or “hate” or “like” for certain candidates. This emotional stuff drives me nuts!! The search for someone likeable has made Americans prey to an actor (Reagan) and a sociopath (Dubya). Both men had political agendas that many of the people who voted for them detested – but the voters didn’t care because they just liked ‘em “as people.” Wouldn’t we be better off if we tried to make decisions like adults? I vote for the ideology and for the resume. 2.Speaking of the resume, Bush is going to leave the US in a horrible fiscal shape, with hollowed-out institutions, and a failed foreign policy. As we discuss hiring his replacement, we ought to emphasize two minimum qualifications: a)foreign policy expertise and b)experience managing a large bureaucracy. That leaves three viable candidates: Bill Richardson, Al Gore, Wesley Clark. This rational approach could benefit Democrats. No current Republican front-runner even comes close to these three men in the “resume” department.

It shouldn't be hard for Rob Richie to keep up with the serious problems with IRV, considering they have been pointed out to him several times. Yet he reliably ignores the corrections, and repeats the same false and misleading statements for new audiences. Here are some examples: www.blueoregon.com/2007/03/dpo_straw_poll_.html#c63423062 rangevoting.org/RichieRV.html www.nysun.com/comments/17647

Consider the comments to Richie's recent op-ed, in which he, once again, repeated several false statements about IRV. Clearly the respondents (most of whom have advanced math/science credentials) had no patience for it. One of them writes:

Instant Runoff Voting is a recipe for disaster. I was a member of the state committee of NYS Green Party for 4 years, where IRV was used for elections, etc. It's a terrible system.

-- http://www.opednews.com/articles/genera_fairvote_070426_instant_runoff_votin.htm

And the Libertarian Reform Caucus calls IRV a "bullet in the foot". http://reformthelp.org/issues/voting/runoff.php

So, don't take my word for it.

Richie's claims that IRV has proven "effective" are vague and subjective, and rely on ad populum appeals and other marketing strategies. We at the Center for Range Voting, on the other hand, have used extensive computer modeling of elections in order to calculate precise figures of voter satisfaction index (http://RangeVoting.org/vsi.html). Who do you trust, someone who uses scientific principles and discusses the facts and figures, or someone like Richie who uses these "sweet lemons" marketing tactics, and ignores the wealth of evidence in contradiction to his claims?

Interesting to see range voting tried out here, although I suspect some voters will figure out the tactical incentives created by the system that put "insincere voters" at an advantage over sincere ones.

As Rob well knows, honest voters are always at a much greater disadvantage with IRV - making this one of his more bizarre claims. We actually created this page to address it. We have presented it to Rob, but notice he just continues, undeterred by facts.

To add to the irony, Richie doesn't even know the proper strategy that a cunning Range Voter would want to employ. In a recent online post he assumes the strategy would be to give "the top score to their favorite and a zero to everyone else."

Wrong. Imagine your favorite candidate is Ralph Nader. Would you really want to give him a 10 and give everyone else a 0, or give him a 10 and also give your favorite "electable" candidate a 10?

Think a little harder Rob.

Rob is right

Nope.

The two-party outcome is more a function of the "magnitude" (the number of seats in contention) than it is the translation rule (votes -> seats). With magnitude M=1 you tend to see M+1=2 viable candidates.

This is a common myth - the belief that single-winner elections will be two-party dominated under any election method (so it's not IRV's fault that it leads to two-party domination).

The counter-evidence is that 21-23 of the 27 countries that use a genuine runoff, for example, have three or more viable parties in contention in their major single-winner races. This is because of subtle but important strategic differences in the methods.

This page explains why IRV (and most ordinal voting methods) leads to two-party domination.

Think of it in simple terms. If voters were not strategic with IRV, we would expect that it should lead to more than two parties, because it would not matter what party a candidate ran under - he should still have the same chances of winning. But clearly voters understand that there is a strategic incentive with IRV to top-rank your favorite front-runner, not your favorite candidate. And of course, to appear to be a front-runner requires you to be in one of the same two parties from which the winner always comes. And so, you have to be in one of those two parties to win with IRV (save for some extremely rare exceptions, which occasionally happen under plurality as well).

So I repeat: No, Rob Richie is not right.

The article Clay cites is pretty badly distorted. For one thing, it attempts to dispute the idea the IRV always produces a majority winner! Given that by definition the process stops once a candidate secures a majority of votes, this is a fairly absurd argument.

It's not absurd at all. Here's a simple example of a 3 candidate election where IRV fails to elect the candidate who would win by a majority to either of his opponents.

% of voters - their vote 34% Obama> Edwards> McCain 17% Edwards> Obama> McCain 15% Edwards> McCain> Obama 34% McCain> Edwards> Obama

The other argument made is that IRV can make spoiling worse--and then skews the conclusion from their example. http://rangevoting.org/IRVpartic.html

In the example, 8 voters used the same Nader-Gore-Bush combination. For an unknown reason, they are separated into two rows, 5 votes and then another 3.

The reason is pretty clear. It is to distinguish the two groups. While both groups may have the same ordered preferences, one of them stays at home and the other votes.

The authors attempt to show that the 3 voters who "spoiled" by voting Nader cost Gore the election in this example, because if they hadn't voted at all, Gore would have won.

This is absurd of course, because it wasn't three voters who "spoiled," but EIGHT.

Or you could say that it was the 7 Bush voters who spoiled the election, by showing up and voting for Bush. But you are missing the point. The point isn't to lay "blame" to any one group of voters for the outcome. The point is that IRV exhibits a severe and bizarre pathology, in which voters get a better result by staying at home than by showing up and adding more support for their sincere favorite candidate. I think this is a behavior that a reasonable person can agree is absurd, and yet another of the myriad of problems with IRV that make it unsuitable for use in any important elections.

Furthermore, what we have here is a non-exhaustive list of outcomes.

Of course we do. We are specifically discussing a problematic outcome that can occur with IRV.

Putting aside the fact that it's utter nonsense that most Gore voters would put Bush in front of Nader in 2000...

These are just hypothetical candidates whose names are familiar to people. You can insert a more realistic set of options. That is inconsequential to the mathematical reality of this pathology, and it is troubling that you mention this even in passing.

the entire table is worthless for analysis because it is assumed there are NO Gore-Nader-Bush voters (when in fact it's likely they would be the majority of Gore voter profiles).

We don't make that assumption actually. We just simplify the example, by leaving out negligible factions. Since you are having problems conceptualizing this with the given candidates, here's a more "realistic" example that might help you.

#voters their vote 7 Dean > Kerry > Brownback 6 Kerry > Dean > Brownback 5 Brownback > Kerry > Dean 3 Brownback > Kerry > Dean

This seems plenty realistic to me.

Dean wins after Kerry is eliminated.

But if 3 of the Brownback fans had not voted (the bottom row), THEN Kerry, their second choice, would have won. They would get a better result by not casting votes.

So we have one outright untruth (IRV doesn't guarantee a majority winner), and one horribly skewed argument unreflective of reality. Not much of a rebuttal.

Taking more time to study these issues before you make these false assessments will save us both a lot of time. I hope my response here has been helpful and informative for you and other readers.

After watching the debate I want to change my vote. Mr. Gravel- give that guy a 10!!

The only real question, is what psychological pathology gives Clay S. such a overblown view of the alleged superiorities of range voting. Did he invent it, as no one has ever heard of it before this year?

Give it up Clay, there is simply no need for the negative aspersions and labored attempts to find examples where IRV might not be perfect. IRV is a perfectly fine system, it saves money and increases choice compared to a top-two runoff system, and really does not have any bad consequences -- IT REFLECTS THE VOTERS' WILL, all 'negative' results attributed to IRV flow from that. And range voting seems to reflect the voters' will also, and it might work too. There is just no need to exalt one to heaven and cast the other to hell, there's gotta be some untold emotional backstory to account for Clay's need to act that way.

Anyway, I came looking because it's after 6 pm on Fri and Kari had promised to post the results of the poll, DONE BY RANGE VOTING which we all agreed would be an interesting experiment (esp. if we do it again next week by IRV). So, where's results, team Blue?

Clay, your revised example doesn't change a thing on the first point--the candidate who wins will be the one with a majority of votes at the end.

In my first IRV example above, Edwards is preferred to both his rivals by a 66% "super-majority" of the voters. So clearly Edwards is the majority winner - yet IRV eliminates him in the first round.

Sure, at the end, after you've eliminated the true majority winner, Edwards, in an earlier round, so McCain is the majority winner among those remaining. But that's not saying much. For instance, consider the hypothetical Shentrup voting method: Two candidates are picked at random, and then a plurality election is held between them. It always elects a "majority winner" in the same way IRV does. But most people realize this is a ridiculous definition of "majority winner".

To refute that is to not even understand how a winner is reached by IRV. In the example you present, there is no winner, since no one has a majority.

In my example, Edwards - the majority winner - is eliminated in the first round, and McCain beats Obama in the second round. So McCain is indeed the IRV winner with a fake "IRV majority".

The only distinguishment is that you've decided a particular portion will stay home, without any reasoning as to why.

Let's just say to play their Nintendo Wii. If that's not an exciting enough reason, just use your imagination.

It's arbitrary, controlled to give you the result you are looking for.

Well, the result was there to be found. If IRV didn't have this problem, we wouldn't have found examples of it no matter how much we tried.

You don't provide any rationale as to why only 3/8th of the spoilers might stay home.

Same rationale that millions of Americans use when they stay at home. Busy, have a date, gotta take the kids to soccer practice. You know...life is busy.

My question for you would be, can you give me any rationale for why this situation would not be possible?

Your point was that spoiler voting creates a worse result, when in fact the result of the example doesn't change whether you have spoiler votes or not.

Yes it does. By staying at home, those 3 voters get a better result. Voting honestly hurt them.

Or you could say that it was the 7 Bush voters who spoiled the election, by showing up and voting for Bush.

Quite so--thus correctly identifying the true agent of change here: Bush voters, not Nader voters.

No, both. Those 3 Nader voters could stay home, or they could show up and the Bush voters could stay home, and Bush would lose. There's no single faction to blame. And again, blame is not the issue. The issue is the flawed behavior of IRV.

I beg your pardon? Inconsequential?

Yes. Inconsequential. I posted the exact same example with the names of candidates that would make more sense to you, and the math is exactly the same.

The only reason Bush wins in the scenario you have is that because after the first round, Gore drops out instead of Nader...and when that happens, Gore's votes go to BUSH instead of Nader.

Yes, this is indeed how the flaw takes place.

Aside from being completely stupid in a real-life analysis of that election, as I pointed out it fails to consider all possible voter combinations and thus can't adequately represent a methodologically sound hypothetical. Nice try.

So tell me which voter combinations are missing here, that would be likely to occur in significant proportions so as to have affected this scenario?

7 Dean > Kerry > Brownback 6 Kerry > Dean > Brownback 5 Brownback > Kerry > Dean 3 Brownback > Kerry > Dean

The only real question, is what psychological pathology gives Clay S. such a overblown view of the alleged superiorities of range voting. Did he invent it, as no one has ever heard of it before this year?

I certainly didn't invent Range Voting. It has been used in the Olympics for a long time. But in any case, I have stated copious evidence in support of Range Voting here. If you think my conclusions are overblown, I invite specific criticisms of the arguments.

Give it up Clay, there is simply no need for the negative aspersions and labored attempts to find examples where IRV might not be perfect.

The point isn't that it's not perfect, but that it's actually quite poor - a claim for which I've provided a great deal of evidence here. Democracy is too important to be conducted with a poor method such as IRV.

IRV is a perfectly fine system

I have presented substantial data to show that it is not a fine system in fact. It produces a ver lower voter satisfaction around 60% - http://RangeVoting.org/vsi.html

it saves money and increases choice compared to a top-two runoff system

I wouldn't say it increases choice, since all four major IRV countries have achieved two-party duopoly in their IRV elected posts, whereas most of the 27 countries which use a genuine runoff have escapd two-party duopoly. See http://RangeVoting.org/HonestRunoff.html

and really does not have any bad consequences -- IT REFLECTS THE VOTERS' WILL

Not very well it doesn't. http://RangeVoting.org/vsi.html

all 'negative' results attributed to IRV flow from that.

I don't know what this is supposed to mean. IRV's negative results flow from the fact that it is an ordinal, not cardinal, voting method. They flow from the fact that it uses different ballots and needs different voting machines, etc. etc.

And range voting seems to reflect the voters' will also, and it might work too.

It reflects the voters' will better than any other known single-winner voting method.

There is just no need to exalt one to heaven and cast the other to hell, there's gotta be some untold emotional backstory to account for Clay's need to act that way.

Or it could just be the rather extensive evidence that I've laid out, such as these social utility efficiency figures: http://RangeVoting.org/vsi.html

Range Voting's benefits are definitely worth exalting, while IRV is an extremely poor voting method that leads to two party duopoly, increases spoiled ballets (by a LOT), and may become fixed where it is implemented, such that even if voter have a bad experience with it, they are unlikely to have the stomach for a further upgrade to Range Voting. So IRV can be damaging in this sense.

Eagerly awaiting results...

I'm curious to see how well they'll match my ongoing 2008 U.S. Presidential election using Range Voting, which has 4,000 votes cast so far. Here are just the candidates you ran in your Range Voting straw poll, from my poll (which used a 1-5 range).

Barack Obama : 3.10 (WINNER, by far) Al Gore : 2.76 John Edwards : 2.59 Hillary Clinton : 2.34 Wesley Clark : 2.26 Dennis Kucinich : 2.17 Bill Richardson : 2.08 Joe Biden : 2.00 Christopher Dodd : 1.68 Harry Reid : 1.60

A more recent poll I posted (with only 43 voters so far, so don't read too much into it) had the candidates like this:

Al Gore : 3.15 John Edwards : 2.75 Barack Obama : 2.62 Ron Paul : 2.60 Rudy Giuliani : 1.74 John McCain : 1.74 Hillary Clinton : 1.68

So hopefully this is an indication that people are losing faith in Hillary. Her sheen of inevitability is quite gone.

But IRV would seemingly fail here. Obama would get 34% of first-place votes, McCain 34%, and Edwards 32%. Of course, it's an unrealistic example with these names, but it's not so unbelievable given other names. It's this example that has made me think twice about IRV.

Here's probably more what you had in mind.

<u>% of voters - their vote</u> 26% Obama> Edwards> McCain 23% Edwards> Obama> McCain 2% Edwards> McCain> Obama 49% McCain> Edwards> Obama

Edwards has 25% of the first round votes, and is eliminated. McCain then wins 51% to 49% against Obama. But Edwards is preferred to Obama by 74% of the voters, and preferred to McCain by 51%. Not just the wrong candidate wins, but the wrong party.

You're making the fallacy of looking beyond the numbers for a result you can't get. The cold fact is that Edwards is the least popular choice for office in that example--68% want someone else, and he is the first choice of the fewest people.

You're making the fallacy of thinking that there's something special about "first choice". It is entirely possible, for instance, that the Obama voters like Edwards even more than the Edwards voters. In that case, if Edwards won, the Edwards fans would "want someone else" even more than the Obama fans would "want someone else". So your belief that first choices are somehow more important than subsequent choices is false.

And IRV agrees with me. See this example, where a candidate with a tiny fraction of the first place votes wins with IRV.

#voters their vote 1000 Z 512 P>A>Y>X>W 256 Q>A>Y>X>W 128 R>A>Y>X>W 64 S>A>Y>X>W 32 T>A>Y>X>W 16 U>A>Y>X>W 8 V>A>Y>X>W 4 W>A>Y>X>V 2 X>A>Y>W>V 1 Y>A>X>W>V 2 A>Y>X>W>V

In this "nightmare" IRV election, the "no hopers" based on their top-rank vote counts would seem to be A,Y,X,W, and V, who combined own less than 1% of the vote. The lowest of the low would seem to be A, X, and Y. Meanwhile the "top dog" would appear to be Z with 49.4% of the vote. However,

* A wins the election.
* But if two more voters came and voted for Y, then Y would win.
* On the other hand, if the two extra voters went for X, then X would win.
* On the third hand, if W got the two extra votes, then W would win.
* But if V (or U, or T, or S, or R, or Q, or P) had the two extra votes, then Z would win.
* But even if Z were granted 20 more votes, then Z still would not win!

Let me try to make this even simpler. Given the preferences stated in my example, IRV would say that Edwards was the better candidate if he ran against Obama. And IRV would say that Edwards was the better candidate if he ran against McCain. But when they all three run together, IRV suddenly "changes its mind" and decides that McCain is suddenly better than Edwards.

And perhaps even more bizarre is what happens if we look at it in reverse. If we are looking the IRV ballots and about to "call" the election, McCain is winning, according to IRV. But if Obama drops out right before we call it, then Edwards wins. Here's a humorous anecdote I've heard about this paradox: After finishing dinner, Sidney Morgenbesser decides to order dessert. The waitress tells him he has two choices: apple pie and blueberry pie. Sidney orders the apple pie. After a few minutes the waitress returns and says that they also have cherry pie at which point Morgenbesser says "In that case I'll have the blueberry pie.

That's not how IRV works

My point exactly.

and it's not appropriate to try to define a hypothetical that literally can't happen.

You are wrong. I have discussed a perfectly realistic example of this in my previous post. IRV fails to elect the majority (Condorcet) winner in 4%-24% of 3-candidate IRV elections, depending on how far they degrade toward plurality voting under the effects of strategy. With 5 candidates, this becomes 10%-44%. So it can and does happen, and making unsupported and patently false claims like this does not help you.

IRV is not DESIGNED to give the spoiler a shot

The spoilers were Obama and McCain, not Edwards.

it is designed to leave no vote waste

Wrong again. In my example, all of the Obama voters prefer Edwards to McCain, yet those preferences are totally ignored by IRV, because it only looks one layer at a time. So those "Edwards > McCain" votes were entirely wasted. IRV does not consider them - it does not even "see" them.

And now, all of those Obama voters are thinking, "Gosh...I wish I would have just gamed the system and voted for the liberal that I thought had the best chance against the winning conservative." And then next time they vote, they use this strategy, which shuts out third parties (because they aren't electable, and will either lose or just act as spoilers) and causes IRV to rapidly degrade toward plurality, and causes it to result in two-party duopoly.

Don't believe me? Look at the Australian House of Representatives, where they have been using IRV since 1918, and their 500+ seats are two-party dominated. Look at the Irish presidential election, which isn't even a powerful post, and sits alongside a multi-party PR legislature and still results in, not duopoly, but monopoly - there has been only one exception to single-party rule since the post began in 1938, and it was a phenomenal fluke.

Malta and Fiji experienced the same effect with IRV as well.

So this is reality. So you need to stop quarreling with me and realize that what I'm saying is the truth, and that we need a better voting method than IRV.

You don't get the right to see Edwards win by coming in third

Which is another reason why IRV is unfit for use in elections, and produces a poor social utility efficiency.

IRV REFUTES ITSELF

Here's a case where IRV does my work for me. It actually refutes itself.

<u>Number of voters - their vote</u> 9 Clinton > Obama > Edwards 8 Edwards > Clinton > Obama 7 Obama > Edwards > Clinton

In this 24 million voter election, Obama is eliminated first, and then Edwards wins in the second round.

But now let us say that we wished to ask IRV who the worst candidate is, by reversing every voter's preferences.

<u>Number of voters - their vote</u> 9 Edwards > Obama > Clinton 8 Obama > Clinton > Edwards 7 Clinton > Edwards > Obama

Clinton is eliminated first, and Edwards still wins. Edwards is both the best and worst candidate.

Once again, another hypothetical vote that can't ever happen. Worst candidate? WTF?

What do you mean "another". You said the same thing before, but you were wrong. IRV can and does fail to elect the Condorcet winner, often.

The point that this more recent example is intended to illustrate is that regardless of how common this problem is, it is a severe pathology that is telling of what a poor fundamental model IRV is, structurally.

And don't forget the other refutations I provided, which were not based on any particular example election.

I'm not particularly interested in weird-ass hypothetical examples

Maybe they don't interest you, but they are important, because they often expose fundamental problems with IRV that do happen in the real world. Calling them "hypothetical" does not diminish their importance in this regard. Indeed these problems apparently are quite realistic, because the net Bayesian regret they cause in elections is quite large - meaning less voter happiness with election results.

nor am I interested in giving the third-place guy in IRV a shot in a Range Vote.

Why not? He may well be the utility maximizer, in which case you are being economically imprudent not to elect him - because it will probably be to your detriment not to.

To me, the question isn't which voting system serves the candidates best - but rather which one serves the voters the best.

That would be Range Voting, as expressed here.

I'm still open to thinking about this stuff (especially since a real proposal ain't nowhere in the cards), but it seems to me that a simple rank-ordering doesn't necessarily reflect the fullness of a single person's opinions.

Right. Range Voting both captures more information than other systems, by including intensity data, and uses that data more intelligently.

For example, in the range vote, I set the following values:

See that giant gap between Joe Biden and Wes Clark? That's how I feel - there's a big gap between them.

And so your Range Voting ballot will cause that much of a point difference to be expressed between them. None of the information you express will be ignored by Range Voting. Whereas IRV can waste most of what little information it does use. And then of course there are even worse methods, like plurality, which obtain almost ZERO information about voter preferences.

If I had to merely rank them top to bottom, the Biden/Clark gap would be the same as the Edwards/Gore gap

And most likely IRV would decide the winner before even getting that deep into your vote - thus ignoring what little information you gave it to begin with.

but why shouldn't I be able to express my true opinion?

Especially when the strategic incentive to "polarize" your scores to all 10's and 0's isn't huge, and calculating the best way to do it is probably more trouble than it's worth - and you may pick a bad strategy, where as honesty is a pretty good one.

If the purpose of a voting system is to most fully express the views of the individual voters - and, in aggregate, the will of the electorate... well, it seems to me that range-voting does that the best of the systems I've explored so far.

And that is confirmed by our social utility efficiency calculations, which compared some 60 different methods.

http://rangevoting.org/vsi.html

Clay, you still haven't proved a darn thing except what an un-justified Johnny-One-Note you are.

Your "examples" are so poorly written no can understand what they're supposed to mean. You are hand-picking your examples.

I, on the other hand, have had the experience of numerous actual experiences, at Conventions of an Oregon political party, where dozens of people played out scenarios of voting for ice cream flavors under the leadership of people who understood IRV. You could watch the votes of the losing candidates move to the bigger groups. IRV works. It's clear. It expresses the will of the voters. I have no hatred for range voting, but YOUR OPINION -- which we are still not clear where it's coming from since your math examples and rhetoric are off-putting junk -- that range voting is so un-imaginably orders of magnitude better does not prove anything for the rest of us.

Clay, you still haven't proved a darn thing except what an un-justified Johnny-One-Note you are.

It's not true that I haven't proven anyting. I have proven that IRV fails the Condorcet Criterion, the Monotonicity Criterion, and Independence of Irrelevant Alternatives. These are serious flaws that election method experts widely recognize as significant.

Your "examples" are so poorly written no can understand what they're supposed to mean.

No one ever said that understanding the relatively complex math involved in analyzing voting methods was simple. I have done my best to use language that would be comprehensible to the layperson. Where I have failed, I think it only fair to assign some of the blame to "lazy" readers, who have responded before doing the necessary research to gain a deeper grasp of the concepts. If in fact you cannot understand what the examples mean, I think it would be better to say, "I do not understand what you mean, so please elaborate" than to prematurely claim victory. I'm always willing to elaborate further, as I have quite a passion for democracy.

I, on the other hand, have had the experience of numerous actual experiences, at Conventions of an Oregon political party, where dozens of people played out scenarios of voting for ice cream flavors under the leadership of people who understood IRV.

I'm not questioning your experience. I'm saying that you are wrong about several things you believe concerning voting methods, namely IRV and Range Voting.

Furthemore, understanding how IRV operates is not the same as understanding election science. I highly doubt that most of those alleged IRV experts would be able to describe to me the Gibbard-Satterthwaite theorem for example. How many of them would know what social utility efficiency even is?

You could watch the votes of the losing candidates move to the bigger groups. IRV works.

Sure, IRV works. You could also draw the candidates' names out of a hat, or elect the tallest candidate, and that would "work". But it wouldn't pick very good winners, which is the problem for IRV. IRV has a social utility efficiency of only 76.32% for honest voters, and 39.21% for totally strategic voters. For Range Voting it is 94.66% and 77.01% respectively. So IRV starts much lower, and suffers twice as much from strategy. It is simply a very poor voting method - the second poorest of the common methods, next to plurality.

IRV fans continue to ignore this fact, and simply repeat their false belief that IRV is "perfectly fine". But claims without evidence are a dime a dozen.

It's clear. It expresses the will of the voters.

Nope. With somewhere between a 39% and 77% social utility efficiency, it does not express the will of the voters very well at all.

I have no hatred for range voting, but YOUR OPINION -- which we are still not clear where it's coming from since your math examples and rhetoric are off-putting junk

Calling it off-putting junk is not a refutation, unfortunately. If you have any specific rebuttals to offer, I would appreciate hearing them, and would be happy to reply.

that range voting is so un-imaginably orders of magnitude better does not prove anything for the rest of us.

It's not my opinion - I have cited the social utility efficiency figures, and you are free to grab our source code and add additional utility and strategy generators if you like, in case you don't believe that ours were sufficiently realistic (and if you additionally have a gut instinct that that lack of realism favored Range Voting over IRV).