by Michael Ossipoff
The strategy of voting is important because—depending on the voting system—strategy can be useful or, as in Plurality, even necessary. In Approval, you needn’t vote strategically. As I said in my previous article here, “Some Problems with Plurality”, you can, in Approval, just approve your favorite(s), or all the candidates you like. Or you can vote strategically. In that article, I described some voting strategies for Approval.
As was mentioned in the article that introduced the new second Approval poll (which has just begun and can be found in the left column of this page and DC’s front page), the purpose of this new Approval poll is to demonstrate how information from a previous Approval election can provide information for strategy in the next Approval election.
So I suggest, let’s vote strategically in this second Approval poll, based on the information from the first Approval poll. Though I’ve already discussed Approval strategy, there are things to say about it, in relation to this current Approval poll, and the available information.
Before I start on that, let me say that all methods share the problem of over-compromising. When ICT and Symmetrical ICT get rid of the Chicken-Dilemma, they largely get rid of the over-compromise problem; however, not entirely because the over-compromise problem doesn’t entirely depend on there being a Chicken-Dilemma.
Need for intensive computation – usually requiring machine counting – are too vulnerable to count-fraud. The more simply and easily implemented point systems, such as Approval, and maybe Score, can be conducted with a count fraud–secure hand-count.
Approval Poll Strategy Revisited
An advantage of Approval, Score, ICT, and Symmetrical IC versus Plurality, IRV, and unimproved Condorcet, is that the former set of methods requires twice as many mistaken over-compromisers, in comparison to the latter set of methods, to mistakenly give away an election. That’s a significant advantage, and it’s true regardless of what kind of strategy those over-compromisers are using.
So I’d like to discuss strategy for voting in the current Approval poll: It’s difficult to discuss that in detail without saying something about political preferences. It probably wouldn’t be appropriate for me to express political advocacy here, but detailed strategy discussion requires mentioning political preferences. I’m not trying to promote political positions, even if I might unintentionally imply what mine are, or discuss how you should vote if you have certain specified political preferences.
In a previous article, I defined the u/a election. It’s an election in which there are unacceptable candidates who could win. So unacceptable that voting so as to keep an unacceptable candidate from winning is much more important than the matter of which acceptable, or which unacceptable candidate wins. I claim that our official public political elections are u/a elections. That means that our poll is one too. You might or might not agree with me on that, and that could influence your voting strategy.
One thing about a u/a election is that the strategy of Approval and Symmetrical ICT are uniquely simple. If it’s a u/a Approval election, then just approve all the acceptable candidates and none of the unacceptable ones. I don’t know of any method—other than Symmetrical ICT and Score—that can match the strategy-simplicity of Approval, in a u/a election.
Here’s where I disclose, for illustrative purposes (but avoid promoting) my preferences. In our poll, I consider the Green candidates, Jill Stein and Roseanne Barr to be acceptable. I consider the Democratic and Republican candidates to be entirely and emphatically unacceptable. I encourage you to judge such matters for yourself, based on performance in office, and on the platform proposals (available on the web) of the GPUS and Democratic parties.
That leaves two other candidates, the Libertarian Gary Johnson, and the Constitution candidate Virgil Goode. Because of the lack of information, I counted them unacceptable and didn’t approve them. Moving forward, my point is that I regard the election as u/a, with Roseanne Barr and Jill Stein as the only acceptable candidates; therefore I approved only Barr and Stein. Although candidate-promotion is not my purpose here, I should say this: Many people do not seem to believe that Roseanne Barr is a serious candidate; however, looking on the Internet at what she has said regarding her candidacy could change that view.
For instance, read her answer to a Green candidate questionnaire: You’ll find that she’s quite serious, and remarkably outspoken. She’s my favorite candidate in this poll—she has my Plurality vote in the Minguo Plurality poll (Minguo is conducting four 2012 presidential polls: Plurality, Approval, 0-10 Score, and 0-2 Score). Minguo is working together on some new projects with Democracy Chronicles; you can vote in their polls (http://minguo.info), in addition to the ones here at Democracy Chronicles.)
If, as do I, you regard the poll as u/a, your strategy needn’t take any win-ability information into account. Just approve (only) the acceptable candidate. I mention that in case some of you consider the poll to be u/a. If you do, then maybe you assign the “acceptable” and “unacceptable” status as I do. But what if you feel that the Democrat is acceptable and that the only completely unacceptable candidates are the Republican and worse.
If you think that the Democrat is acceptable, we could debate that, but it would probably be inappropriate for me to say more than I already have about that topic. Anyway, I reluctantly have to admit that, if you consider the Democrat acceptable, and the Republican (and anyone worse) as completely unacceptable, then you should approve the Democrat and everyone who is better than he, and not approve the Republican or anyone who is worse than he—and, if you have those perceptions about acceptability, you should do so without regard to any win-ability information.
However when I say, “…if you consider the election u/a, with the Democrat acceptable and Republican unacceptable…” that’s actually a very strong “if”. It would mean that the acceptability difference between Democrat and Republican is so great that you must vote the Democrat over Republican, to ensure that the Republican doesn’t win, even though Republicans have always consistently polled last in Internet polls, and therefore the Republicans don’t have any significant chance of winning in this poll.
Realistically, if you accept certain undesirable attributes of the Democrats, then don’t you, at least a little, accept those same attributes when they are found in the Republicans? This preference for Democrats over Republicans—isn’t it really more in the nature of partisan loyalty, or home team–winner preference, rather than genuine outrage at unacceptable attributes of Republicans that aren’t possessed by Democrats? Aren’t those good old boys really good old, and similar, friends, when you look at the great similarity of their policies and deeds? And if you prefer Greens’ policies to Dems’ policies, then how strong or important, really, could that Democratic Party loyalty or Democratic home team–winner preference be?
So my point is that I claim that you probably don’t regard this poll as a u/a election with the Democrat acceptable and Republican unacceptable. Remember that, for the election to be u/a, you must consider the acceptability-difference between the two sets of candidates to be so great that any differences within the sets are negligible–so that all that matters is avoiding the worse set, and it doesn’t matter which member of the top set, or which member of the bottom set wins.
So, as I said above, I suggest that you, for you, this poll is not u/a, with the Democrat acceptable and the Republican unacceptable. Now, maybe you still like the Democrat better than the Republican (but still prefer Greens to Democrats). But now the Republican’s low win-ability has great strategic relevance. As I said, Republicans always, consistently poll at bottom—last, in Internet polls, including the first Approval poll here.
Therefore I say this: If the Democrat isn’t your favorite, then you have no reason to approve the Democrat in this second Approval poll. This is because the only reason to approve a non-favorite is if you need him to beat someone worse. But you don’t need a Democrat to beat a Republican. The Greens can out-poll Republicans, as we found in the first Approval poll, with 53 ballots. In fact, in the current Approval poll too, Stein is far ahead of the Republican, who seems to have no chance of beating Stein.
As of last night, Stein and Obama are tied for the top spot. You don’t need the Democrat; if you prefer a Green, then you can get a better outcome than a Democratic win—you can get a Green win, if you approve the Greens and not the Democrat. Note that right now, Stein is in a tie-for-top with Obama in our current new Approval poll. If you prefer the Greens, then you have no reason to do anything other than help them against the Democrat.
Of course if the Democrat is your favorite, then, in the interest of nonpartisan-ness, I should admit that—because Obama won the first Approval poll—if you feel that he’ll be as well-supported this time as last time, then you don’t need the Greens as an anti-Republican compromise. Of course, those who prefer the Green candidate, having noticed that they don’t need the Democrat to beat Republican, no longer have to approve the Democrat.
Furthermore, what if that means that Obama doesn’t out-poll Romney? Ok, I admit, that isn’t likely. As I said, Romney, being the Republican candidate, didn’t do well, and it doesn’t seem too likely that he is the favorite of more of our voters than is Obama. But what about Johnson? He might do better than Romney. If Romney’s poor showing causes Obama to lose compromise-approvals from the Green supporters, and then might Johnson not out-poll Obama? And might you not need the Greens as a compromise against a Johnson win?
Of course, one could use a similar Johnson argument to say that those with a Green preference could need the Democrat to beat the Libertarian. But how important is that to someone who prefers the Green? If you prefer Greens to Democrats, do you think that the merit differences between the Democrats and Libertarians are really great enough to justify voting Democrat, and thereby not helping Greens against the Democrat?
To fairly evaluate the merits of candidates and parties, I suggest that you read (on the Internet) the parties’ platforms, and things that the candidates have said. For instance, I suggest that you read the 2008 and 2010 platforms of GPUS, as well as what that Roseanne Barr and Jill Stein have said, and compare their policy proposals to the policies and deeds of Obama. Then decide for yourself about the difference between Greens and Obama, and what you count as acceptable and as unacceptable.
Approval Poll Strategy – Summary of Strategic Voting Suggestions:
If you consider it to be a u/a poll, then approve all of the acceptable candidates and none of the unacceptable candidates.
If not, then approve your favorite(s), and those you genuinely like, and approve others only if it’s likely that they’re the only candidates who can beat someone who is even worse than they are.
To put it in terms of the better-than-expectation strategy that I described in a previous article: Approve the candidates who are better than what you expect from the election. Maybe or maybe not approve a candidate if s/he is exactly as good as what you expect.
Ok then, what do you expect? Well, Obama won last time. There’s a good chance of his winning this time too. From the point of view of someone who prefers the Greens, the candidates worse than Obama did very poorly last time, and the candidates better than Obama did very well. So it’s safe to say that the expectation in this election is Greens-ward from Obama. If you’re a Greens supporter, then the expectation is better than Obama. That means that better-than-expectation strategy says to not approve Obama.
It’s been shown that the better-than-expectation strategy maximizes your expectation for the election. It’s your best strategy, the way of voting that is most in your interest. So if you prefer one or more Greens to Obama, and even if the Democrats aren’t unacceptable to you, then, strategically, you should approve the Greens but not Obama. If such are your preferences, then that’s your best strategy, based on the results of the first Approval election.