Policymic has this essay by Daniel Kamerling, advocating approval voting. The article specifically says the contest for the 2012 Republican presidential nomination would be improved if presidential primaries used approval voting.
Policymic calls itself “an online news platform to engage millennials in debates about real issues.” The term “millennials” generally means people born after 1982, who were graduating from high school at the dawn of the current century. Thanks to Rosa Barker for the link.
Thanks for sharing the Approval Voting article.
I’m curious how approval voting advocates think approval voting would work in this contest. If you like one of the top four best, but really don’t like one of the four much, how would you vote?
The typical advice is “vote for your favorite frontrunner” and anyone you like better. So… do you vote for only one of these people? How is that different than plurality voting?
@Rob
There’s a technical answer to this here: http://www.electology.org/threshold
Approval Voting would be the “maximally strategic” option.
For a rougher approach:
You know you vote your favorite no matter what. It sounds like this candidate is lucky enough to be in contention (lucky you). Then you’d consider how likely it was for the unliked candidate to win and how much you dislike that candidate.
If this unliked candidate is likely to win and is pretty bad, then you’d vote for the other candidates who are also in competition to win but who are acceptable. You’re hedging your bet against the unlikable candidate this way.
Alternatively, if the unliked candidate is not so bad or unlikely to win, then you’d just vote your favorite. Here, there’s not enough risk to hedge your bet, and you want to make sure your favorite wins–even over the acceptable candidates in contention.
Of course, you can also lend support for any other likable candidate not in contention in both scenarios.
As much as you love to pick apart all these hypothetical examples, the reality is that we have a great deal of Approval Voting data from large contentious elections. E.g. the German Pirate Party, which has been very successful and recently won 10% of the seats in the Berlin parliament, widely uses Approval Voting. You can see trends here:
http://www.electology.org/pirate-elections-germany
Notice the party chairman got an average of 1.73 approvals per ballot. Clearly it doesn’t make any sense to ask, “How is that different than plurality voting?”
Of course we know about your complete lack of objectivity about this issue, as witnessed here:
http://www.electology.org/fact-check#TOC-Rob-Richie-criticizes-Approval-Voti
There is nothing particularly novel or unusual about this GOP contest. Santorum and Romney are the clear frontrunners. So here are a couple of scenarios:
Voter prefers Paul over Gingrich over Santorum over Romney
Voter approves Paul, Gingrich, and Santorum
Voter prefers Romney over Gingrich over Santorum over Paul
Voter approves Romney
So for some voters it’s like Plurality. For other voters it isn’t. In the aggregate, the results are substantially different than Plurality Voting.
Approval Voting is a step on the road to
Head to Head Condorcet voting — i.e. Approval Voting as a tiebreaker if there is no Condorcet Winner.
How about — IF no majority Approval person, then declare a vacancy — especially needed if the candidates are all airhead morons.
DemoRep,
Approval Voting is superior to Condorcet methods in basically every way.
1) Radically simpler for voters and election officials.
2) Better outcomes (lower Bayesian Regret) under most circumstances.
3) Plausibly more likely to elect a Condorcet winner than REAL Condorcet methods:
http://ScoreVoting.net/AppCW.html
Here in uber-liberal San Francisco, we’re currently seeing a strong push for a repeal away from Instant Runoff Voting, back to Plurality Voting. The most effective talking points are about the complexity of the ranked ballot, including increased spoilage rates. This is a pretty clear indication that ranked systems are just not politically viable in the USA.
In my view, Approval Voting is a “step on the road” to Score Voting, which would be the ultimate in terms of performance and simplicity.
Clay,
You’re simply wrong that “Romney and Santorum are the clear frontrunners” and that, correspondingly, votes for Gingrich and Paul can be expressive votes. It all depends on the state, but Paul has his share of close second-place finishes, Gingrich has won states and this month Gingrich, Santorum and Romney are all fighting to win in Mississippi and Alabama.
So again, if the advice is “vote your favorite frontrunner and anyone you like better,” what are you going to do if you are an Alabama voter who likes Gingrich best, but prefers Santorum to Romney? Your advice would be “vote only for Gingrich”, as he’s your favorite frontrunner.
As to comparing this to the Pirate Party’s internal elections where I’m sure it was nothing like the intensity of this contest, you’re deluding yourself.
On San Francisco, there’s no push to go back to plurality voting. The talk is go back to two-round runoffs for mayor and other citywide offices (and keep RCV for Board of Supervisors). For thoughtful people backing that, the case is tied to their interest in a “second look” at the top candidates — not to the alleged complexity of the ranked ballot. In nonpartisan elections with large candidate fields, that can be an issue under ANY one-winner system.
For lower tech folks —
A > B
Along comes Choice C.
Condorcet Head to Head math —
C may beat both A and B.
C>(A>B)
C may lose to both A and B.
(A>B)>C
C may beat A — BUT lose to B.
C>A>B>C
34 ABC
33 BCA
32 CAB
This was noticed by Condorcet in France in the 1780s (repeat 1780s).
Condorcet was one more victim of the devil robot party hack morons in the French Revolution — because he had some math brain cells.
Political SCIENCE moves on – building on the works of our predecessors.
Condorcet math can be used for ALL legislative, executive and judicial elections.
Test Winner(s) versus Test Loser — All others are deemed Test Other Losers.
The votes for the TOL go to a TW or TL.
A Condorcet Winner wins in all combinations.
A Condorcet Loser loses in all combinations.
Would need computers to do ALL of the combinations in any larger election.
Approval Votes can be tiebreakers if all CW are not directly chosen.
—
Pending utopia –
P.R. and App.V.
@Rob
I don’t know how many times we have to point out that bullet voting is not an issue with Approval Voting. This frequently doesn’t make any sense tactically. That’s because it’s good to have the option to hedge your bets against opponents by voting for other acceptable contenders. You may have multiple favorites. And you always want to show support for likable candidates not in contention.
Admittedly, it occasionally makes sense to vote for a single candidate (bullet vote) with Approval Voting. But, much more often than not, it makes more sense to select multiple candidates. And it’s important that voters have that option.
Not real world enough? There’s also plenty of real-world evidence against bullet voting here: http://www.electology.org/bullet-voting
Interestingly, as you look at that page, IRV has a relatively worse record with bullet voting. Perhaps we should start responding that not only is Approval Voting reasonable with bullet voting, but IRV is actually worse.
We have other options, of course, when we talk about IRV. We could talk about how in practice IRV encourages betraying one’s favorite and hurts voters when they rank their favorite candidates as better. Or we could talk about how relatively complex IRV is for its dubious benefit (a small buffer against vote splitting).
We like our pick of Approval Voting. It’s incredibly simple and comes with a wide range of benefits (as if wisely choosing its winner wasn’t good enough). We’re sorry you don’t like it, too.
Aaron. C’mon, please address the situation of the Alabama voter. By your own strategic recommendation, backers of Santorum, Romney or Gingrich should vote for only one of those three. And I suspect most Ron Paul voters would bullet vote for him as well. Walk me through a real scenario of what might happen in Alabama with approval voting in the primary.
Also, you and Clay also deny the reality of what happened with Bucklin voting in states like Alabama that tried it. Bucklin is basically a kind of “slow motion” approval vote. When used in U.S. elections, bullet voting increased over time, results were erratic and the system was rejected everywhere it was tried precisely because of concerns about bullet voting. Those simply are the facts and you can find them too.
Then of course you do your hate-IRV thing, not acknowledging that in the many US cities and nations that use it there isn’t any evidence that voters start voting insincerely (unlike plurality voting, where we all know many abandon their first choice) and there are many examples of where it allowed the Condorcet winner to win after trailing in first choices.
And finally, of course, you continue to do little to promote proportional representation. FairVote backs campaigns for it, does frequent writing for it and so on. We could actually cooperate on that if you wanted to be productive.
Rob Richie tells his typical story, where selective examples from history are asserted to mislead. Where to begin?
Bucklin voting is Instant Runoff Approval, not “slow motion” approval. It simulates repeated approval polls — an ancient application — where voters gradually add approvals, depending on their preference strengths and a desire to complete the election. Voters will start out, in the first rank, approving just their favorite, usually. Then, in lower ranks, they *may* add additional approvals. The method seeks majority approval. Under some conditions, such as, apparently, party primaries in Alabama in the 1940s, there may not be a lot of additional approvals! But even there, Bucklin garnered something like 11% of voters adding them, which is often enough to fix the spoiler effect.
There is no particular evidence I’ve seen that Bucklin increased in bullet voting over time, under constant conditions. There were early public elections with high use of additional ranks, which were politically killed, rather obviously, because they worked. Voters loved Bucklin. Ask Richie about Bucklin in San Francisco! There were then some the straggler implementations that lasted for much longer, but which were running under conditions where low use of additional approvals could be expected. They were replaced, I think, by top two runoff. What would have been better, long term, would have been to add a runoff to Bucklin if a majority were not found. That would have motivate people to add more approvals, so they wouldn’t have to vote again! (But they won’t vote again if either option is satisfactory to them, which it will often be. Richie, as a Top-Two-Runoff-Killer, cites low turnout in runoff elections as a problem, when it is the opposite. When a really bad choice is present in a runoff, voters turn out in droves, that’s known.
The claim was made here that “Approval” would improve party presidential primaries. “Approval” is simply Counting All the Votes. There is no doubt that, in a party primary situation, this would improve overall satisfaction with results, but what is debatable is *how much*? If it were a costly reform, that question would be crucial. But it’s cheap, practically free.
Richie won’t mention that in primaries, where more than two candidates with some chance of winning is common, IRV falls totally on its face, and can very badly err. Approval will do no harm, and may help, maybe even a lot. But even better would be a ranked approval method, like Bucklin, or Score voting. Even better than that, methods which elect a highly representative convention and let the convention work out who gets votes and how.
Richie may say he’s promoting proportional representation, but the cumbersome STV method that he proposes can be highly unrepresentative (it depends on details). A method is known that would produce practically perfect representation, and it was designed to be easy to vote, almost 130 years ago. I doubt that he’s interested.
Richie built a career attacking the most successful long-term voting reform in the U.S., top two runoff. It has some problems which could easily be fixed. But instead of fixing it, he works to dump it and replace it with a method that pretends to emulate IRV, but definitely worsens results, history shows. He doesn’t care.
I’m tired of the deceptions. I’m glad to see that voting system reformers are waking up. One size does not fit all, and snake oil can make you sick, setting back reform for a long time.
Abd —
I’m still waiting for your or another approval voting enthusiast to explain what you would advise to backers of one of the three GOP candidates essentially tied in the polls right now in Alabama and Mississippi — with each polling around 30%, the odds that anyone would get even 40% with approval voting, let alone a majority.
As to me, we’re clear that in our ordering of single-winner voting methods that might be won and used in contested governmental elections, it’s IRV, followed by runoffs and plurality last. As to forms of PR, we also have not been an STV-only group, as even a cursory visit to our website would show.
On Bucklin, your history is simply wrong, and you and others just aren’t confronting the reality of what happened. Here are a few excerpts from a longer account at approvalvoting.blogspot.com:
* Bucklin voting in practice led to a massive dropoff between the number of voters indicating a first choice and second choice. In Alabama, for example, sixteen statewide primary election races were used with Bucklin voting between 1916 and 1930. On average, 87% of voters cast only a single bullet vote for one choice
* San Francisco in 1916 adopted Bucklin voting, but stopped using it soon after. One issue was changing to voting machines that had trouble tabulating the system, but another apparently was a massive rejection of casting additional choices, with only 3% of voters doing so.
* Portland (OR) used Bucklin voting from 1913 to 1932, when it was repealed by 81% to 19%.
* Reviewing this history of Bucklin voting, Ernest Bernard Shulz in his American city government: its machinery and processes (1949, pages 216, 217, and 370) said Bucklin’s most serious defect was the least surprising to someone familiar with the later-no-harm criterion – that a voter’s second choice could count against a first choice. The National Municipal Review fingered this same problem far earlier. It critiqued Bucklin as having the “essential defect” that “the expression of a second-choice may harm the first-choice” and that “as soon as the largest group of electors discovers that it does not pay to record second preferences, the system will tend to break down.”
@Rob
-On Approval-
I gave you the approach to take with Approval Voting on any scenario. If you choose not to process it, then that’s not my responsibility.
I’ve already pointed out that it is sometimes appropriate to bullet vote with Approval. Most of the time it isn’t. And most people choose multiple candidates when given the choice. I don’t know what kind of concession you’re looking for.
-On IRV-
IRV, on the other hand, has some serious issues that are more frequent. Your strategy appears to be denying that these IRV debacles have happened at all.
For instance, you mentioned that many IRV voters have voted honestly. But that honesty should not put us at ease. In point, many Burlington IRV voters were honest. And because Burlington’s conservatives voted honestly, IRV elected their least favorite.
That kind of atrocity is never a risk with Approval Voting. Worst case scenario with Approval Voting, you wind up electing an acceptable candidate rather than your favorite. If only voters in Burlington were so lucky.
Your later point on “majority” with IRV and Approval is off. No voting system can guarantee a majority with more than two candidates running (minus a contrived majority via endless runoffs such as by Robert’s Rules).
The fact that IRV has a winner with greater than 50% of the remaining votes means nothing. Eliminate enough candidates even randomly and you will reach this same outcome. This “majority” says nothing about the remaining candidates’ quality. And IRV easily eliminates good candidates–particularly Centrists.
-On Bucklin-
As the different name suggests, Bucklin is not Approval Voting. Bucklin also has a more complex ballot and its algorithm is different. This and other factors make it difficult to know why bullet voting discrepancies exist between Bucklin’s previous use and Approval Voting’s modern use.
Sure, Bucklin (a system that is not Approval Voting) was repealed. But voting systems get repealed all the time. Just look at IRV. It’s been repealed all over the US. The same repeals happened for STV in the first half of the last century.
Whether a system is repealed for the right reasons is another question. I strongly disagree with some of the simplistic rhetoric that occurred against IRV (e.g. “one person, one vote” nonsense). And I’m quick to correct others when I see this come up. Of course, there are very good reasons to replace IRV as well.
-On PR-
I know FairVote has lots of money and that you are paid a full time salary. But we have neither million-dollar resources nor the ability to focus full-time on this. The fact that many of us spend so much time on this issue speaks only to our will. Consequently, we’ve chosen to focus on Approval Voting because it needs an advocate and we have to recognize our current resources.
This says nothing about any of our opinions about PR. Personally, I love PR. I’ve read much of Douglas Amy’s work as well. I’ve even stood up for crappier PR systems like Cumulative Voting and SNTV. And I’m just as saddened as you are over STV not being around the way it used to be. I think there’s better than STV, but STV is such an improvement that I would be eager to have it.
If anything, I’m saddened that you don’t spend more time trying to implement PR. Despite our friction, I don’t see a future collaboration as being impossible here. That discussion would warrant more private correspondence, however.
Rob Richie,
Your argument seems to be that, for instance, Green Party voters would vote Democrat AND Green as long as the Greens weren’t frontrunners. But if they became strong enough to be frontrunners, then Green Party voters would bullet vote for the Greens. I hate to break it to you, but that “nightmare scenario” is unlikely to sway anyone reading Ballot Access News. I think most readers would LOVE to see third party and independent candidates getting accurate representation of their support.
The other interpretation of your argument is that Approval Voting “degenerates” into SINCERE Plurality Voting. E.g. somehow that Nader supporter who bullet voted for Gore with Plurality Voting would have instead bullet voted for Nader if he had been allowed to vote for as many candidates as he wanted to. As bizarre and totally implausible as that seems, it would at least be a huge improvement over ordinary tactical Plurality Voting.
FairVote’s web site actually specifically cites favorite betrayal as a major problem with Plurality Voting. So surely you would see Approval Voting as a HUGE improvement over Plurality Voting regardless of which of the above interpretations we go with.
> You’re simply wrong that “Romney and Santorum are the clear frontrunners”
That was actually the HEADLINE on a national newspaper I saw a couple days ago. Of COURSE there are states which deviate from that national polling, and voters will adjust accordingly.
> On San Francisco, there’s no push to go back to plurality voting. The talk is go back to two-round runoffs for mayor and other citywide offices (and keep RCV for Board of Supervisors). For thoughtful people backing that, the case is tied to their interest in a “second look” at the top candidates — not to the alleged complexity of the ranked ballot.
I have actually LIVED in San Francisco for eight years, and I have recently attended meetings at local Democratic clubs, and at City Hall. People generally DO NOT like runoffs. In fact, even the detractors of the current Instant Runoff Voting implementation generally like not having runoffs. By far the most common complaint I hear about IRV is the complexity and confusion.
In my discussions with local progressive leaders about a possible switch to Approval Voting, one thing I bring up is the question of whether to include a runoff if no candidate is approved by a majority. Here’s what one local activist (who nearly was elected to the Board of Supervisors) said: “Approval plus a runoff takes you down the road of admitting that runoffs are popular, which they really aren’t.”
I wish you could see how your lack of objectivity or humility creates and motivates your opposition.
Rob,
Also, your claims about Bucklin voting are just wrong.
http://groups.yahoo.com/group/RangeVoting/message/6229
You can see a very detailed discussion around FairVote’s strange Bucklin claims here. Notice the Wikipedia edit shenanigans.
https://groups.google.com/forum/#!msg/electionscience/4u4W1GppT70/qAZbjitlWnsJ
Gentlemen:
Amidst the long messages, the ongoing failure to accept the implications of what happened with Bucklin elections (read the citations at the approval voting blogspot and follow up on them if you have questions, Clay) and the odd statements about San Francisco (like not admitting that no one is talking about plurality voting there), let me just zero in the original question.
To repeat, Alabama has a very competitive, very important primary coming up on Tuesday. The latest polls collected at RealPolitics.com have the averages of recent polls this way
* Romney – 26.3%
* Gingrich – 24.0%
* Santorum – 22.7%
* Paul – 7.0%
The two most recent polls have it even closer.
So, __by Clay’s own logic (and to a certain extent Aaron’s far more tortured logic)_ you’re in a bind if you have a distinct preference order, which in fact is normal. Clay says “vote for your favorite frontrunner and anyone you like better.” That means you vote for one of the top three and add in Ron Paul if you like him better (which most backers of the top three candidates won’t, so Paul’s vote liklely won’t grow much, but at least Paul voters could weigh in on the choice among the top three if and only if they accept they decide their guy can’t win).
Any partisan for one of the top three would be saying “vote just for my guy.” All the 24-hour cable pundits would be explaining how screwy this can be if you vote for your second choice along with your first choice.
I guarantee that no candidate would come to winning a majority. And… quite plausibly the winner would be the Condorcet loser among the top three – exactly as would have happened in Burlington if approval voting had been used there.
And at the end of the day, yes, the results would look FAR more like plurality voting than a runoff or instant runoff.
If I’m wrong, please zero in on the Alabama example. Try to game out approval voting being used right now in Alabama. Tell a reasonable story about what might happen that would avoid a plurality-type outcome.
Rob,
> the ongoing failure to accept the implications of what happened with Bucklin elections
I just showed you numerous examples where a substantial portion of the population did not bullet vote with Bucklin voting. The implications that you keep asserting are totally baseless.
And the elections in Alabama DID NOT EVEN USE Bucklin voting!
http://www.electology.org/fact-check#TOC-Attacks-on-Bucklin-voting
> the odd statements about San Francisco (like not admitting that no one is talking about plurality voting there)
I quoted a statement by a candidate who come in second place in a San Francisco supervisor race, that runoffs are unpopular. And here’s an article titled “Ranked Choice Voting Confusing, poll says”:
http://www.sfgate.com/cgi-bin/article.cgi?f=/c/a/2011/03/09/MNG11I6QNO.DTL
> quite plausibly the winner would be the Condorcet loser among the top three – exactly as would have happened in Burlington if approval voting had been used there.
We have a nearly bullet-proof argument that the Condorcet WINNER would have won in Burlington.
http://ScoreVoting.net/Burlington.html#UngerApp
>The latest polls collected at RealPolitics.com have the averages of recent polls this way
>* Romney – 26.3%
>* Gingrich – 24.0%
>* Santorum – 22.7%
>* Paul – 7.0%
>The two most recent polls have it even closer.
So I’m sure you can imagine that some Ron Paul supporters will be voting for someone other than Paul. Whereas with Approval Voting, NONE of them would have an incentive not to vote for Paul. The only exceptional aspect here is that Paul supporters tend to be so passionate that they are pretty honest, even with Plurality Voting.
> So, __by Clay’s own logic (and to a certain extent Aaron’s far more tortured logic)_ you’re in a bind if you have a distinct preference order
It’s extremely probable that IRV would have a Burlington-like result with these preferences, in which voters who prefer X>Y>Z will get Z, but could have caused Y to win if they had insincerely ranked Y in first place. That sounds like a bind to me.
> Any partisan for one of the top three would be saying “vote just for my guy.”
Just like Nader’s supporters did. And look how well that worked.
> All the 24-hour cable pundits would be explaining how screwy this can be if you vote for your second choice along with your first choice.
With IRV, all the 24-hour cable pundits would be explaining how screwy it can be if voting for your first choice causes your third choice to defeat your second choice. And so on, and so on.
But you lack objectivity, so you ignore IRV’s problems and hyper-exaggerate Approval’s.
> And at the end of the day, yes, the results would look FAR more like plurality voting than a runoff or instant runoff.
“Look like”. IRV provides the ILLUSION of a majority winner.
Clay. My last message just so people can realize your argument haven’t won the day.
1. You are absolutely wrong about Bucklin and Alabama’s Democratic primaries for governor and other offices. Not the first time you’re wrong, of course, but please, do some research.
2. You are absolutely wrong that any serious person is talking about plurality voting in San Francisco. It’s either RCV or runoffs or some combination. Period.
3. Approval voting in Burlington would have elected Wright (the candidate who would have lost to either of this top two challengers if matched against them one-on-one). Jameson Quinn looked at the numbers and was enough to admit I’m right on this.
4. You haven’t answered the Alabama question — which is just an update of the same question that could have been asked going into any number of recent GOP nomination contests.
Onward.
Rob,
1. I cited the results from two of those Alabama primary elections, proving they were not Bucklin. What is your evidence for this bald assertion that I’m “absolutely wrong”?
2. I never claimed there was a push for Plurality Voting WITHOUT RUNOFFS. My point is that a substantial fraction of the people who prefer Plurality to IRV are primarily motivated by IRV’s complexity/confusion. As far as I can tell, many of them LIKE not having runoffs with the IRV implementation — but they think having runoffs is WORTH trading IRV for Plurality.
If they really liked runoffs, then presumably the progressive wing of the board would try to put forth a counter-measure of IRV followed by a runoff between the top two finishers from the IRV contest. I actually heard David Chiu speak about that at the D3 Democratic Club. But I’ll bet you that doesn’t happen. Because people DO NOT generally like runoffs. They want to get rid of IRV because they prefer Plurality to IRV, regardless of whether it entails runoffs.
3. We have a nearly bullet-proof argument that Democrat Andy Montroll (the Condorcet WINNER) would have won in Burlington with Approval Voting.
http://ScoreVoting.net/Burlington.html#UngerApp
You keep asserting Wright would have won, but without citing so much as a single flaw in our model. You have a habit of making assertions without facts to back them up.
4. We have answered your “Alabama question” thoroughly. Once again, here’s the grade school level math.
http://www.electology.org/threshold
Wow. So now I’m a noted authority? Great!
Fact is, Approval voting in Burlington could plausibly have elected any of the top three candidates; I suspect that Wright could be the most likely result the first time, due to the “chicken dilemma”. But if similar elections happened in the future, voters could easily learn how to elect the best candidate, depending on the situation, but without ever needing dishonesty. In IRV, voters would soon learn to be voting dishonestly, leaving a two-party monopoly; exactly the problem with plurality that IRV was supposed to fix.
Meanwhile, Bucklin would probably get the right result on the first try; though, like approval, it is flexible enough to elect any of the three without voter dishonesty, if preference strengths shifted.
As to Richie and Clay’s repeated assertions that each other are simply wrong about Alabama… it’s a matter of the definition of “Bucklin”. Even if Richie is right about the definition, and Alabama counts as Bucklin, his point is irrelevant to the larger issue, because whatever you call it, the system in Alabama had nothing to do with Approval.
The fact is that real-world experience with Bucklin and Approval give no support to the idea that these systems devolve to plurality. Even if some, or even most, voters reasonably decide to vote for only one, real world Approval elections with more than two serious candidates have consistently had enough non-bullet voting to significantly change the dynamics.
Seems like Richie is kind of in his own little world of make believe.
Re Jameson Quinn’s first paragraph in comment 21:
I don’t think the primary goal of IRV is to fix the duopoly problem. The main push of FairVote, the first thing you see in all its literature, is that it fixes the “spoiler problem”. By its very definition, the spoiler problem is that dividing votes between a small-party candidate and a more moderate large-party candidate, the opposing large party candidate can win.
This is a very short-sighted goal. It is tantamount to saying that things would be all very well and good if the fringe candidates would bow out of the race when they know they don’t have a chance, but since they won’t, we’ll give their proponents a way to vote for them anyway without hurting the compromise candidate, because they will be eliminated in early rounds for lack of broad first-place support.
Ted Stern,
IRV does NOT fix the spoiler problem, it only reduces it.
http://www.electology.org/spoiler
> It is tantamount to saying that things would be all very well and good if the fringe candidates would bow out of the race when they …
Well, I don’t agree. My view is, the spoiler concern can convince voters not to support a VIABLE (non-fringe) candidate, just because they think he or she “can’t win”. A candidate doesn’t have to be fringe to be portrayed as unelectable. Maybe the candidate isn’t one of the fundraising leaders, or didn’t get the right endorsements. Regardless, the problem is that a candidate with enough support to be the legitimate winner CAN LOSE if the voting system punishes voters for throwing away their vote on a candidate who APPEARS to have a chance. So the monied opposition can cause a less monied winner to actually become a loser, just by convincing the supporters that the candidate “can’t win”.
If you truly fix the spoiler problem, that goes away. That’s why I support Score Voting and Approval Voting.