The ethics of strategic voting.

Mary and Ann do not take account of other factors, not listed in the OP but none the less operative:

Clinton is assured of adoption as the Democratic candidate
The international standing/reputation of the United States, and the domestic standing/reputation of democracy, are both damaged by the adoption of Trump as the Republican candidate

I think the flaw is in (2). Trying to decide who other people are voting for then strategically casting your vote appropriately makes the error of assuming that your vote is independent of other people’s votes. But other people are going through similar thought processes you are. So other people’s votes are correlated (weakly or strongly) with yours. That’s not to say that by the way you vote you _cause_ them to vote that way, just that starting from the assumption that their votes are mostly fixed while yours is not is a considerable error.

A true story: In my very first election, I was voting for governor. There were four candidates. Three had major traditional power bases. The fourth, a former chief of staff, appeared to have no natural consistency. I liked him, but the polls showed him trailing by some distance, while the others were neck-and-neck. The pre-election polls were something like: 25, 24, 23, 15.

So when it came time to vote, I could choose my ‘ideal’ candidate, or decide he was a long-shot and vote strategically for the one of the three ‘viable’ candidates I liked better. There was one candidate I definitely did not want to see win, so that was a factor in my vote.

On election day, it was indeed very close, but my preferred candidate, rather than being a distant fourth, lost the primary by only 41 votes (after the recount). My ‘alternate’ candidate, meanwhile, was in a distant 4th. It turned out I wasn’t the only one who decided not to vote ‘strategically’.

That opened my eyes to the flaws in the ‘strategic voting’ model.

I think that when people say they’re strategic voters and believe in “voting so as to have the best chance of producing the best outcome”, they don’t quite mean that. There can still be other norms that apply to voting. It’s not like strategic voters think it’s permissible to vote twice if they can get away with it, after all, even though that would help them get the outcome they want.

So maybe there’s a difference between voting for a candidate because you want them to win and voting for a candidate because you want them to lose. The point of a primary is for people in a party to get together and figure out who they’re going to put forward in the general, and “this one’s more electable” vs “this one would do a better job if elected” is a perfectly fine sort of argument to have, according to strategic voters. But that doesn’t mean they have to think that it’s okay to try to sabotage the other side’s process – the other side also has a right to get together and put forward who they consider to be the best candidate. Of course I doubt that this is something most strategic voters could lay out for you, but I’d guess that they feel it’s inappropriate to sabotage the other party’s primary and that the root of that feeling is something like this. One simply doesn’t do this, and I imagine that it hasn’t even occurred to many strategic voters that they might do this, whereas a major argument for Clinton that’s out there in the world is that she’s the safer general election choice.

Like anon, I think Mary is right due to the ‘Trump loses big to either dem candidate” condition.

As for why most people won’t do it, I think disgust may play an important role in distorting judgment here (making judgments irrational). It’s rational to feel a high degree of disgust for Trump and his views; but it’s not rational to let that disgust unduly influence your voting choices. But I think the fact that most people like Mary and Ann (rightly) feel a high degree of disgust for Trump will (wrongly) avoid voting for him under the circumstances you specify.

Correction: “presumabley to vote *against* Trump.”

If the five assumptions are taken as a given and Ann intends to be a strategic voter as described, then Mary is correct about what Ann should do.

The actual Anns you know are balking at this because (1) they don’t hold all five assumptions with equal conviction; (2) are troubled by the idea that if they’re wrong they might have affirmatively contributed to a deplorable outcome; (3) probably a lot of them have a latent sense that when strategic voting crosses from “voting for not your first choice” to “voting for someone you actually despise in an attempt to force the other party into a bad position” is dirty pool in some way; (4) a belief that even if all five assumptions are ironclad truth, the prolonged position of Trump as Republican frontrunner is bad for America in a special way not because of the content of his hypothetical presidency but because Ann probably thinks his particular brand of demagoguery seems to specifically whip up violent and racist sentiment.

I think there has to be an implicit premise here that a vote for Trump does more to ensure that Trump is the nominee than a vote for Clinton does her. Modulo the GOP’s hypothesized willingness to nominate a non-plurality candidate if Trump doesn’t win on the first convention ballot, Ann may not share that premise: Trump’s got a commanding lead of states, delegates, etc.

But since all our normal models for parsing electoral races are slagged, I think we need to make as many implicit premises explicit as possible.

Scott P. @ 2 makes a good point: when deciding whether to vote “strategically”, one has to have a model of the rest of the voting body in order for one’s strategy to be worth anything. Or, if possible, to have actually colluded with others about how to vote (the way many[1] progressives in the South actually did vote for Santorum in 2012 in a coordinated attempt to give the Religious Right enough rope to hang itself).

[1] OK, some. Not nearly enough to make a difference, but more than two.

medrawt @11: “(3) probably a lot of them have a latent sense that when strategic voting crosses from “voting for not your first choice” to “voting for someone you actually despise in an attempt to force the other party into a bad position” is dirty pool in some way”

That’s probably true of many, but then the question becomes: why exactly do they think it’s dirty pool, and is thinking so compatible with their commitment to strategic voting?

It seems superficially like a Prisoners’ Dilemma, but it’s not because this is only the primary they’re not the only voters. In fact, their not being the only voters is essential to the scenario. Unlike the real PD, everyone “cooperating” in the primary still doesn’t ensure an optimal outcome (as they see it) in the general. Cooperating in the primary can only increase the odds of getting a worse matchup than defecting would:

* Any Republican vs. Sanders instead of vs. HRC (worse by rule 2).

* Cruz vs. any Dem instead of Trump vs. any Dem (worse by rule 4).

When an action always leads to negative outcomes, for any combination of factors beyond one’s control, we’re not in a PD. The concept of cooperation becomes inoperative, and we’re left with good old deontology vs. consequentialism. Should we vote our true preference, *knowing* that it’s a consequential negative, or should we jettison principle to achieve better consequences? I don’t think we’ll find an answer here. ;)

Once we’ve gone strategic, is voting for HRC or voting for Trump better? I don’t think enough information is given. We don’t know how the odds compare for Sanders vs. Trump or HRC vs. Cruz, only that they’re both worse (from a Dem perspective) than HRC vs. Trump and better than Sanders vs. Cruz. With no strategic distinction, I think principle makes its reappearance. Voting in the opposite primary smacks of sabotage. If it were the only strategically-best choice, people might do it anyway, but since there’s a strategically equal alternative without that taint people flock toward that instead.

I don’t know how the open primary in Wisconsin works. Here in Montana, you choose whether to vote a Republican packet or a Democratic packet, so if I decide I want to vote strategically for Trump, I’m giving up the chance to vote in the Democratic primary for the Public Service Commission, the challenge to a friend on the County Commission, an interesting open state senate race, etc.

That is, other variables apply.

anon @ 15:

Partly because of uncertainty about the strength of the assumptions in play, partly because of the “Trump is bad even if he has no shot because he whips people up into a willingness to be violent in a way Cruz does not” thing, partly because, as Bloix and others have said, there are limits on what self-styled “strategic voters” are actually capable of bringing themselves to do (which suggests that they aren’t actually pure, cold-blooded strategists).

If we’re talking about ethics, I think we would have to consider the ethical effect of voting for a given candidate on Ann. The actual winner-output of a single vote in a large election is infinitesimal, so it might not involve much ethics. But by voting one attaches oneself as an assistant — or accomplice — of the one for whom one votes. (The word vote comes from a Latin word meaning to vow or devote.)

It is also possible that a large losing vote may yet put the fear of God, or of the voters, into the winners and their associates. Hence, voting for Mr. Sanders may help preserve my Social Security even if he loses. But maybe that’s not an ethical but a business consideration and irrelevant here.

I am a Wisconsin voter so this is not a theoretical problem for me. I prefer Hillary to Sanders — but that doesn’t make any real difference to the problem at hand.

One additional issue that I have considered is whether the best possible outcome for Democrats might not be a contested convention, in which the Republican insanity can be must-see TV for days on end. In that case, voting for Cruz might be the preferred solution.

I think that Mary is right, and I think that most people who hold Ann’s views know that Mary is right. This leaves us with the perplexing question that you posed: why will so few (if any) Anns vote for Trump in their state’s open primary? I think it probably has something to do with maintaining an image. If you vote for a person (or a party) who makes openly racist, sexist, xenophobic, etc. comments, it’s hard to pretend that you play no role in maintaining oppression and injustice. However, it’s easy to wash your hands of oppression and injustice when you vote for a person (or party) who openly supports policies like Affirmative Action and marriage equality. Never mind that the party of Affirmative Action and marriage equality is no less invested than that other party in maintaining an inherently unjust socioeconomic status quo. We’re not allowed to question the value of maintaing our current economic structures, because civilization was built on capitalism and technological progress and investment are inherently good, and blah, blah, blah. As a result of our inability to envision any economic structures other than the ones currently imposed upon us, the average voter overlooks the fundamental and very non-progressive similarities between our two limited political options and focuses, instead, on those social issues that ostensibly separate the “progressive” from the “conservative.” If you want to maintain your standing in one category or the other, you have to vote for the candidate whose public image falls into the desired category OR you have to vote against the candidate whose public image falls into the undesirable category.

I’m not sure your final line is correct – there is no data on this that I am aware of. Anecdotally, I suspect a significant proportion of Trumps vote is what might be called the ‘mischievous’ cohort. Those who take a delight in voting for whoever annoys the particular establishment they dislike the most. You see this all the time in student elections. I’ve no idea if its statistically significant, but I’d suspect a proportion of that ‘mischievous’ vote are people who would identify as progressives or left of centre and for one reason or another (such as in closed primaries) can’t vote in the Democratic primary. I suspect that where its allowed, Dem voters who can vote in the Republican primary are indeed voting for Trump (and also possibly for Cruz, as the polls I’ve seen indicate he would do very badly against either Dem).

The other point about this is that its a strong argument for transferable votes. You see this sort of pattern emerging all the time in Irish elections, where there are many transfers that can only be explained by people thinking several moves ahead strategically. My anecdotal experience from comparing elections in Ireland (where there are transferable votes) and the UK (where there are not in most elections), is that the transferable votes actually encourages people to think and vote strategically, whereas single non-transferable elections tend to be more ‘beauty contests’.

If this argument works (and I’m not saying it doesn’t), couldn’t you ratchet it back to show that Ann can or should strategically vote for Sanders?

“2. HRC is more likely to beat any Republican candidate than Bernie”

This is a good illustration of the problem with strategic voting. Actual relevant data, such as polls, indicates the opposite. This is not a trivial issue with the exercise, the whole argument for voting for Hilary Clinton collapses because one of its factual premises is simply incorrect.

The problem with strategic voting is that people who engage in it are idiots. An actual strategic vote would involve paying close attention to the electoral rules, poling data, and past election results, and then use it to cast the most effective vote. People who actually try to do this, political scientists and professional political strategists, make mistakes about the probabilities of electoral outcomes all the time. Its really hard to do this. I’ve noticed that when people try to make arguments for strategic voting, they tend to not even bother doing the legwork to get an understanding of the political environment people will be casting their vote in.

Not all strategic voters are idiots. Some are liars. I’ve been following this stuff for decades, and strategic voting arguments consistently get deployed on behalf of the local equivalent of the PMDB (https://en.wikipedia.org/wiki/Brazilian_Democratic_Movement_Party) or the PRI, the party or faction that has no policy agenda but all the local political fixers support. They are the party you support because your uncle runs the firm that gets the contract to install the new email servers in the government offices when they win. For people not part of the machine, there are really no good arguments to deploy in behalf of voting for these types of politicians, so crack-brained schemes such as strategic voting are thought up. Strategic voting rationales are never deployed on behalf of somewhat populist left-of-center policies with track records of relevant honesty the few times they do get into office.

Ed wrote:

The problem with strategic voting is that people who engage in it are idiots. An actual strategic vote would involve paying close attention to the electoral rules, poling data, and past election results, and then use it to cast the most effective vote. People who actually try to do this, political scientists and professional political strategists, make mistakes about the probabilities of electoral outcomes all the time. Its really hard to do this. I’ve noticed that when people try to make arguments for strategic voting, they tend to not even bother doing the legwork to get an understanding of the political environment people will be casting their vote in.

Not all strategic voters are idiots.

My 1 experience of strategic voting blew up in my face, fortunately with no adverse long run consequences. In 2008, following Obama’s victory in Iowa, the 3 of the 4 voters in my family, including myself (I think we forgot to discuss this with the 4th), thought the best way to drive HRC quickly out of the race was to ensure that she came in 3rd here in NH. So we voted for John Edwards. Hoo-boy! Not only did she win here in NH (and I am guessing that we were not alone in our thinking), but John Edwards!

It will be a warm day in hell before I try that again.

Ed, you raise a good point, but calling people idiots and liars is a bit spurious. The good point is that strategic voters might seem or feel like idiots due to factors beyond their control – i.e. the actual election results. They’re making themselves vulnerable to all sorts of insults – from people like you – if their strategic goals are not met. I contend that the problem is not with the strategic voters themselves but with the people flinging insults from their glass houses. After all, they never took that reputational risk. Even if things don’t go their way, they can tell themselves and others that the consequences cast no reflection on their pristine principles. It’s so easy.

Harry @29 — I guess I’m failing to see what the _ethical_ question is here. Mary isn’t claiming that Ann is getting the ethics wrong, she’s claiming that Ann is getting the _strategy_ wrong, in other words that Ann is wrong the logical entailment of the strategy she’s (ethically) committed to. But again, it’s not the ethical commitment itself that’s in question. Right?

In other words, premise 2 is true as long as Clinton has a clean win in the primary, but if her win is seen as a fluke, or depends on the superdelegates, she goes into the national significantly weakened. And this will be the result if enough Democrats who would otherwise vote for her strategically vote for Trump.

Casting about in my own mind for a solution, I find I do not have handy much in the way of ideas about the ethics of strategic behavior in general, and that feels strange somehow.

To behave strategically in general requires the capacity to commit. By definition, strategic behavior is chosen behavior that takes account of other players’ likely responses. So, the ability to make known one’s own responses to other players’ behavior is basic to getting strategic behavior off the dime. The classic illustration is the so-called “game of chicken” where two car drivers put themselves on opposite paths toward a collision, to test which will “chicken-out” and swerve first; a winning move for one driver is to somehow bind up the steering wheel so it is not possible to swerve, forcing the other driver to swerve in order to avoid the collision.

As others have noted, an individual vote is insignificant. So, strategic voting is necessarily also a team sport. Fairly large numbers have to vote in concert and commit to vote in concert in the future as well.

When you think about it this way, you realize that a great many people already vote strategically, when they identify with a political party or party faction or particular candidate and commit their vote accordingly. “Strategic voting” on the premises of the OP tries to take advantage of these prior commitments by other strategic voters.

There is a conflict, of course, between strategic voting by fulfilling a commitment and reneging on the commitment to produce a kind of “false” vote either as a protest or in the hope of affecting the outcome.

The Wisconsin Republican Primary is a semi-winner-take-all contest, in which some delegates are allocated in each Congressional district and some (18 out of 42 total) are allocated state-wide, all on a first-past-the-post basis. (I hope I got that right.)

The Democratic Primary allocates delegates in proportion to the vote. (There are some qualifications, but let it go.)

Since Cruz and Trump are so close in state-wide polls, it is conceivable that a small band of false-flag voters could tip the balance, and contribute either to Trump winning or losing a fair number of delegates. The Democratic Primary, with its proportional allocation, does not present the same opportunity.

I conclude that Mary is being strategically astute in trying to get Ann to vote in the Republican Primary, while fulfilling her own commitment to supporting Bernie.

Since electoral behavior is a team sport, as RP has been pointing out, persuading other people to vote in a particular way is more important than one’s own voting choice. So, persuading Ann to vote in a particular way is the (ethically?) more efficacious behavior. And, persuading Ann to vote in the Republican Primary is a twofer. It serves to obscure the actual balance of Republican Party support for Trump and increases apparent Democratic Party support for Sanders, Mary’s preferred candidate.

Ultimately, though, the analysis misses the core function of voting strategies, which is a strategic interaction between candidates and voters. The candidate, too, is presumably acting strategically. That the candidate is better organized than the Party and the Party better organized than the mass of voters is a defect in the system.

Bernie leads Cruz by ~10 and Trump by ~15; HRC leads Cruz by ~3 and Trump by ~10.

I get that HRC seems more ‘electable’ because that is what the fancy suits on the teevee tell us, but I don’t see the evidence for #2.

That being said, the polling data consistently supports the notion that ‘strategic’ blue voters should be propping up Trump against Cruz, as Cruz looks to have better gen-election chances. I dislike strategic voting for a number of reasons, but after reading and listening to a number of ‘this is why I am voting for X’ stories recently, well… let’s just say there are much worse decision criteria out there.

bruce wilder wrote:

To behave strategically in general requires the capacity to commit. By definition, strategic behavior is chosen behavior that takes account of other players’ likely responses. So, the ability to make known one’s own responses to other players’ behavior is basic to getting strategic behavior off the dime.

The last sentence strikes me as incorrect. There are plenty of games in which players’ actions are not visible to each other, and we talk about strategy and strategic behavior in those games. When a pitcher selects the pitch to throw, it is usually considered a problem (except maybe for Sandy Koufax) if the batter knows what it coming. Is a quarterback sneak not a strategic choice?

What am I missing or misunderstanding?

Making known one’s own likely response is a pre-requisite to lying about it.

This is why the analysis of strategic behavior is so dizzying.

Kent and Harry @28-29:

If we’re asking about Careful Kents rather than Arrogant Anns: the strategic vote is for Bernie, because:

1. A vote for Trump is wasted, a Cruz win is too likely (89% says 538) for strategy voters to make the difference. (http://projects.fivethirtyeight.com/election-2016/primary-forecast/wisconsin-republican/)
2. Your vote might make a difference for or against Bernie, since the polls there are close.
3. There is, RNB’s endless BS on this to the contrary, almost no evidence whatsoever that Hillary is more likely to beat the Republicans. Most polls show he wins and wins bigger against *all* Republicans, and in the past 2 decades, early spring polls have proven reasonably accurate. (By margins of up to 24% in some: http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_sanders-5565.html)

More important, HRC’s proven strengths in the primaries have been the Democratic base and minorities, who will show up in big numbers against either Republican candidate and for either Democratic candidate, while her weaknesses and Bernie’s strengths are independents and the youth vote, where HRC will not sufficiently bring out the vote.

I always give my first preference to one of the candidates who is predicted to do terribly, on the off chance that it will give them some encouragement and consolation. I got the idea after reading an article by an author who was writing about a book they wrote that did moderately well, and how their publisher would ring them with updates such as ” you’ve just sold two books in Newcastle. ” To which the author was bewildered “Newcastle?! I don’t even know anyone in Newcastle”.

Clinton is beating Trump by 10 plus. Sanders by a point or two more but he is at his peak. He has never been hit hard. Don’t be naive

The idea of “strategic voting” itself is nutz. First off, you only have one vote and the value of your vote is so close to zero in the overall scheme, it is zero, in effect. So your vote, at best, is really just a gesture (of some form) of support — a futile gesture at that — for the candidate who comes closest to your political views, etc. etc. Throwing it away on playing a cheap imitation Oracle of Delphi is absurd.

Beyond that, it also seems to make people kinda crazy. Those who believe in the power of strategic voting all too often lose their minds, forget how our elections actually — cumulative national totals — and everything they’ve ever learned about maths seem to escape them.

For instance: Gore lost thirty states to Bush in 2000. Thirty. He won just twenty states and DC. It was therefore physically, logically and mathematically impossible for Florida to have been “decisive.” No one state can be, because a president wins on the grand totals for all states won, not just one state alone. It doesn’t matter “if” it appears one state pushed Bush over. That’s simply an illusion — and now an article of faith among strategic voters. If we play the “if” game, any other state could have done the trick, given how close the cumulative totals were (271-266). Gore wins, for instance, if he takes his home state of Tennessee, even with the loss in Florida. But, again, no state can be “decisive.” It doesn’t work that way. They all count together, cumulatively.

Math tells us this. Logic tells us this. Civics tells us this.

Mary is right to vote for Bernie, given her assumptions. She is wrong to tell Ann how to vote at all and shouldn’t. Ann should vote the way she wants to as well, though I personally think “strategic voting” is absurd. In general, people should stop trying to guilt each other into voting one way or another, and just mind their own business.

I’m not interested in debating any of those assumptions

people really aren’t capable of this, huh?

@plume
“In general, people should stop trying to guilt each other into voting one way or another, and just mind their own business.”

Of all the things you’ve ever posted, this sentence is the one I agree with you the most.

Mary is right but Ann may consider an open Republican convention sufficient. More likely though, Ann just can’t help herself.

I am a real time, actual voting citizen of the state of Wisconsin. I have lived here fighting the good fight as a “progressive” Democrat for over 36 years. I have already voted today and have canvased and organized here on behalf of Bernie Sanders. I can tell you that in my experience talking and listening to folks here to date, no one has spoken about “strategic” voting in the primary. There is, however, a sizable number of people who speak of not voting if Mrs. Clinton is nominated. This is important because many of the folks to whom I have spoken have either not voted recently or have registered just to vote for Bernie in the primary. Judgments 1 and 2 above are all that these people care about and most agree with number 1 and a minority accept number 2. So, this information may not be helpful in your academic exercise here but I hope it brings you closer to what is actually happenning out here on the tundra.

And what is the focus of Sanders’ campaign? Pierson and Hacker have made clear who Clinton is–she is a technocrat working to expand equality of opportunity in a reformist manner. But what is the content of Sanders’ socialism–letting financial firms fail and take the economy down with them; erecting protectionist barriers; talking radical with no viable plan for energy; or promoting a single payer system that the Democratic Congress isn’t endorsing; or making demagogic promises to free most of the 2.2 million people in prison today?
Why are people drawn to him? Or is it that they just don’t like Hillary Clinton for reasons they are hiding from themselves–that is, sexism.

No, Aufhammer is saying Sanders can’t make such promises honestly. That bothers him. As it should. As should promises to radically reduce the prison population. At some point, people will catch on that things are RADICALLY more expensive and RADICALLY less effective than you said and you end up alienating people.

Don’t do it, Harry!

Good thread, this. I agree with what appears to be the close-to-consensus view above that _soi-disant_ strategic voters tend not to be thorough-going strategists, at least generally. This still leaves open the question whether this is more often so because they’re in a muddle about the logic or whether they get squeamish about the ethics, as Harry I think is suggesting @63. I am sort of inclined to think that latter plays the more important role, but I wouldn’t count the former either — people who’ve studied logic can be bad about applying it in their lives, let alone the rest.

dilemma is quite different for me

1. sanders would be a terrible president
2. hillary probably will win, but will also face a very negative campaign that would also poison public confidence in government
3. hillary would be a superior candidate in effecting policy change, but may not be the most electable.

is it fair to fault hillary for the ferocity/irrationality of her opposition, from both left and right?

Or is it that they just don’t like Hillary Clinton for reasons they are hiding from themselves–that is, sexism.

OT but then what about reasons Hillary supporters are hiding from themselves;
-ageism?
-anti-semitism (esp. in Christian demographic groups)
-selfishness (among high-income)
-classism/racism/sexism (among comfortably off happy with welfare being gutted)
-American chauvinism (some Americans culpably indifferent to harms their govt causes to citizens of other countries)
Etc

No, Aufhammer is saying Sanders can’t make such promises honestly. That bothers him. As it should. As should promises to radically reduce the prison population. At some point, people will catch on that things are RADICALLY more expensive and RADICALLY less effective than you said and you end up alienating people.

Neither Clinton nor Sanders can honestly promise that their energy plans will be achieved. They might spend most of their time in office fighting for issues they consider even more pressing. They might be thwarted by the Supreme Court or by Congress. That’s an implicit risk for any presidential candidate’s stated aims.

I don’t think that the Sanders plan deserves two RADICALLYs. If the USA actually charges full speed ahead on the Sanders plan I could see it being expensive, but not ineffective. And if the plan is implemented more slowly and partially than Sanders wanted, it would be less effective but also less expensive, the degree depending on how significantly it is impeded.

I have observed over the past 10 years that energy policy “moderates” have systematically underestimated the actually-achievable renewable energy penetration levels and pace of buildout when there are strong policies to support renewable energy. The IEA, the EIA, most of the commenters on the original Berkeley Haas blog article: they seem to be roughly the same sort of people who in 2005 were explaining that solar/wind energy was a luxury for the rich, a luxury that would not significantly penetrate the developing world in our lifetimes, and that a realistic climate policy needed lots of nuclear power. In 2005 that crowd seemed to have the better quantitative arguments and I agreed with them. By 2016 I’ve seen a decade of nearly-continual falsification of the moderate predictions about how rapidly renewable energy can expand and where it can expand. Continually erring toward underestimating what’s achievable is not more sober or realistic than overestimating it.

I’ve read the OP, not the whole thread, so the pt I’m gong to make here has prob. already been made.

ISTM the act of voting contains an emotional/expressive component and even the most ‘strategic’ voter may not be able to disregard that emotional component. Accordingly, a supposedly strategic voter (a la the ‘Ann’ character) may not vote for Trump b.c it feels too yucky, even if purely ‘strategic’ considerations, on the assumptions of the OP, wd dictate a vote for Trump.

I think that the accusation of sexism lodged at those who favor Bernie over Hillary to be a typical smear

I think it’s hard to argue it’s not relevant at the margins (if the average voter is just an incy-wincy bit sexist, in a national vote that will translate to signicant numbers of votes lost to sexism) but I think it’s equally hard to establish it’s more significant than that the prejudices that impinge on Sanders (and on Americans’ famous antipathy to socialism/social democracy more generally). What evidence has the Hillary camp put forward?

PlutoniumKun writesMy anecdotal experience from comparing elections in Ireland (where there are transferable votes) and the UK (where there are not in most elections), is that the transferable votes actually encourages people to think and vote strategically, whereas single non-transferable elections tend to be more ‘beauty contests’.

But first-past-the-post distorts the ballot long before voting day – everyone knows which 2 parties in any constituency have a chance.

This is why the mad idea of open primaries in the OP is real.

What if we weight each judgment? Suppose 1) Bernie’s superiority as president exceeds 2) HRC’s advantage against any Republican. Suppose further that 4) the likelihood of Trump’s loss to either Democrat far exceeds 1) or 2). (Seems to me 3 is altogether superfluous in this exercise. Even if Trump were potentially way more awful as president than Cruz, the strategy might still be to assure Trump’s nomination, based on the other factors.) Given a strong likelihood of 4) and the premium on 1) over 2), I might be willing to risk a vote for *Bernie* over Trump or HRC.

Couple problems. First, this weighting exercise arguably amounts to debating the assumptions, which the rules prohibit. Second, it only explains why Ann doesn’t feel compelled to vote for Trump, not why she votes for HRC.

I actually don’t think it’s at all obvious that Mary’s reasoning is correct. (I’m not sure why everyone else seems to think it is.) As I read the case, if we take (5) seriously, (1) and (3) are more or less irrelevant. What Ann presumably wants—should want—is a Clinton-Trump match up in the general. In that case, all else equal, voting for Clinton makes just as much sense as voting for Trump. Given that Ann also, ex hypothesi, prefers Clinton as a candidate to Trump, it makes perfect sense for Ann to vote for Clinton.

Now I’m not saying real-life Anns are actually thinking in this way (tho maybe at least some of them are), but it’s not at all clear that Mary is correct that the logic of Anne’s position (as laid out in the OP) commits Anne to voting for Trump.

I’m amazed that it’s gone without saying, but of course it is highly unethical and immoral and just plain bad citizenship to vote in a primary for a candidate and party you have no intention of supporting in the general.

Beyond that, a voter is of course completely complicit in all of the actions of anyone they’ve voted for at any stage of the elections. Complicity cares not for one’s strategic reasons, and in this case the potential consequences of complicity approach Pascal’s Wager-levels of overwhelming tiny probabilities. So Ann is absolutely right on the final argument, and which of them has the correct choice overall depends on confidence and magnitude issues on points 1 and 2.

#68 – “There is, however, a sizable number of people who speak of not voting if Mrs. Clinton is nominated.”

Because that will help make Trump president, and boy, will that show Clinton!
A person who won’t vote for Clinton is giving half a vote to Trump. Why not go ahead and cast your whole vote for him?

@36, marcel proust,

” you both explain and give an example where strategy is necessary or improves outcomes from a voter’s perspective?”

Its a bit hard to give examples without explaining in detail the intricacies of Irish politics (life is too short), but an example would be when you see transfers from left wing to right wing candidates which seem designed to ‘block’ another candidate. For example, a generally left wing voter giving a preference after his favoured left wing candidate to a right wing candidate to boost the latter against what is perceived to be a nastier right wing candidate.

Its often been observed that the Irish system of transferable votes in multi-seat constituencies tends to produce elected politicians who are particularly personable – because someone seen as ‘a decent sort’ will get more transfers from opposition parties. A particular obstacle to Sinn Fein, for example, is that right wing voters often give preference to non-radical left wingers to attempt to block Sinn Fein getting the left of centre vote.

js. @92
I actually don’t think it’s at all obvious that Mary’s reasoning is correct.

yes, I also think that Mary’s reasoning about what “the logic of Ann’s position” commits Ann to do is debatable. Mary’s reasoning rests heavily on #4 — that Trump would lose more readily against either Dem candidate than Cruz would. But the OP includes #2, about Clinton being the stronger candidate than Sanders against either Repub, so it’s somewhat unclear, istm, what the logic of the position leads to.

More fundamentally, as I suggested above, *even if* one accepts the OP’s stipulations and accepts that Mary’s reasoning is correct, the answer to why more ‘Ann’s’ don’t vote for Trump — a question explicitly posed by the OP at the end — quite likely is that not that many voters are pure ‘strategic’ voters, even if they claim to be. (I certainly wdn’t want to pull a lever for Trump even if I were convinced that it was the surest way to ensure a Dem victory in the general. It just feels icky to vote for someone one considers an extremely flawed (to put it mildly) candidate. If Cruz were ahead in the delegate count, rather than Trump, the calculation might come out differently, but he isn’t.)

“Beyond that, a voter is of course completely complicit in all of the actions of anyone they’ve voted for at any stage of the elections.”

This is of course the main reason Anne shouldn’t vote for Trump. But when I voiced this view in another thread – moral complicity as a consequence of voting – I found it was a minority view.

“I’m not 100% sure that it is illegal to vote in two primaries (in different states)”

I’m not either, but I suspect it is so long as you’re legitimately eligible (resident) in both. I don’t think primary elections are actually elections per se, certainly not in the caucus states. The rules are determined by the parties, not the states.

I thought I covered that with: “The polls are all over the place, so there is no reliable information, and both think it is best to vote on the assumption that both races will be close”. So they have no reason to believe it is more likely to be decisive in one rather than the other. I haven’t bothered to check whether WI is a winner-take-all state for the Reps — if not, then it is all only about one delegate really (but one delegate might end up being decisive for Trump of course).

it is highly unethical and immoral and just plain bad citizenship to vote in a primary for a candidate and party you have no intention of supporting in the general.

OK, I’ll bite. Why is it “highly unethical and immoral”? I really don’t see this at all. (I can kind of see the bad citizenship thing, maybe.) I mean, I think strategic voting is problematic, mostly for the sorts of reasons that Scott P. outlined way back @3 (that probably needs tweaking a bit, but basically that), but I’m really not seeing the immorality of it.

(A couple of other responses forthcoming, hopefully.)

a voter is of course completely complicit in all of the actions of anyone they’ve voted for at any stage of the elections.

I don’t want to derail the thread, so this’ll be my only comment on this, but here’s the problem: If you take this seriously, voters for GWB are guilty of war crimes and—impracticality aside—ought to be sanctioned as such. I think that’s insane. (feel free to respond, but like I said, I won’t say more about this.)

Ignore the final parenthesis in my comment @109. Is liable to cause massive confusuion. Sorry about multiple posts!

@js: I’m talking moral complicity, not any legal kind. Attaching consequences beyond one’s conscience’s burden to voting decisions is obviously insane.

As to immorality and unethicality, it seems sort of obvious. The purpose of a primary is for the members of that party to choose their candidate. Deliberately interfering with that decision, voting with the intent to harm that party’s chances is a deeply antisocial act. There’s also a case that a vote is an expression of the combination of personal preference and what you think is the most likely choice to make the party whose primary you’re voting in succeed and that a dishonest expression of those preferences is just as immoral as any other dishonesty.

p.s. or saw it, but didn’t fully absorb, which might be worse… (sigh)

I’m on your side with this one Harry. Looks like our state is going to play a very interesting role in things from this point. But unless your county comes through big time for us, looks like a rightward leaning Supreme Court for the foreseeable future.

“Seems odd to me, but if they are inviting me to mess with their decisions, how am I being immoral when I do it?”

It is not a very proper thing to do. Supposing the candidate was not Trump since he makes the whole thing kind of joke-y but there were 2 or 3 decent Republicans I think it is not exactly ethical if you are a Democrat to vote for the Republican you think is less likely to win when you are not a Republican. This is not really the reason for Primaries to vote for the least likely candidates to win in Parties you don’t support. And if you did this in sports it would be cheating so its unsportsmanlike as well.

Although to be honest, I am sort of glad we don’t have Primaries in Australia. Mostly the Parliamentarians choose who is Prime Minister and who is Leader of the Opposition, although we had a recent innovation in the Labor Party due to leadership turmoil where now Labor Party Members get to vote in a ballot. This is good I think since now they can’t change Prime Ministers by counting the votes in a restaurant, they have to have a ballot of Party Members as well as a vote of Parliamentarians who probably will stick to counting votes in Parliament now since the time saving of the Restaurant vote counting method doesn’t save any time when all the Party Members have to vote in a ballot anyhow. Although at least counting the votes in a restaurant was quick, having Primaries for a whole year is very lengthy. Probably it is the lengthiness which makes people think up voting strategies like voting for the Republican candidate most likely to lose, since once you make your mind up who is the best candidate it stretches out for months and months.

In Australia the closest thing to this we have is Preference Whispering.

This is where the Preference Whisperers devise very complicated Preferences Deals between minor party candidates to improve overall the chances of a minor party candidate getting elected.

I think this is okay myself if the minor parties want to do this, and more ethical than voting for the most likely to lose Republican when you’re a Democrat, but the Government recently passed a Bill to Reform Senate Voting to try to eliminate the minor parties doing these complicated Preference Deals.

Harry @30,

Well, you really don’t get to the ethical dimension until you get at least approximate probability values for the underlying practical situation.

I don’t know if Decision Analysis is a method used outside of medicine, where I have some familiarity with its practical application. I should more accurately say its lack of practical application. The reason it is almost never useful in making medical decisions, is that the outcomes for even relatively simple decision trees are too sensitive to even small differences in the probability of different outcomes at a branch point. Even the simplest decisions have too complicated a tree, and even the branch points whose outcome probabilities are best understood, have some uncertainty as to those probabilities that however small, are still too large to help the patient or the provider out of the muddle.

Medicine is particle physics, in terms of our ability to give probabilities to branch point outcomes, and the simplicity of the overall tree, compared to politics. Some of the commenters mentioned a 60% confidence in one or the other of your five propositions. The sensitivity analysis on a tree with even one node at that wide a variance would be fatal to the idea that you’re going to get usable guidance from your analysis. I’ve seen fairly simple trees made useless by a few percentage points of uncertainty at one branch point.

Where I think the problem offered to us originally goes off the rails, is that a situation presented in only five propositions, is actually way more complicated if unpacked into decision points. The decision to vote for Trump rather than Sanders changes the probabilities of an entire array of outcomes, not just that Trump becomes incrementally more likely to beat Cruz in WI, therefore more likely to win the nomination, which it is claimed makes it more likely for the D to win the general, and by a larger margin, which it is claimed helps Ds down-ballot. Just at the first step, Trump doing incrementally better, we have not just a higher delegate count for him, but whatever “momentum” is generated by doing incrementally better than expected. Doesn’t this increase his likelihood of winning the general, this greater credibility he gains by winning WI, or at least not losing as badly as expected? Obviously Ann doesn’t expect to change the outcome just by her one vote, she relies on many others having the same idea to join her, enough others to change the outcome. But enough others join her to make a difference, and suddenly Trump is doing surprisingly well among college-educated women, so the probabilities of outcomes at downstream branch points change. It becomes acceptable for women to support him, because that horde of women in WI followed Ann’s idea and voted for him, so maybe he no longer is such a shoo-in to lose badly to HRC or Sanders. Gad, he might even win, and all because Ann was the butterfly way back dozens of branch points earlier whose flapping made it just possible for him to do incrementally well enough in the next contest that dozens of races later, he was unstoppable.

At least in medicine, we don’t have to worry about microbes or cancer cells getting destructive ideas from reading the news.

Not that it’s even that simple. What if the tactical voters in WI push Trump ahead just enough to cause a contested convention way down the line? The contested convention ends with a walk-out when both sides try to win on a procedural vote (most likely seating challenges, just like 1912), the walk-out results in Trump and Cruz both claiming to be the one true R candidate, entitled to the R ballot line. It can’t result in a third party as in 1912, because September is now way too late to get ballot access for a third party, so it’s duel to the death between Cruz and Trump for the R ballot line, and in the ensuing electoral law confusion (remember, SCOTUS is 4-4, so no unitary legal resolution imposed from above), the R House gets to make its Speaker president, thus ending the 2016 election in the only conceivable way an R could win this year. Oh, and maybe ending the US as well.

This is a way more tangled web than your five propositions make it seem, and the addition of the lie engaged in when voting tactically only promises to make it more tangled.

I have never followed the idea that it would do the patient good in some circumstances if I lied to them, told them some white lie to shield them from some truth that would hurt them to no good end, or that might tempt them to make a bad choice. Adding a layer of lies makes an already difficult to understand situation impossibly difficult. Decision Analysis has never helped me out of a medical thicket, but the general model has helped me to understand the practical reason not to tell white lies that will only serve to make the thicket really impenetrable.

Most voters aren’t going to follow Ann’s idea because they don’t want to start down a path whose end they are wise enough to understand they cannot see, an obscurity they can only worsen by choosing to vote against their true beliefs.

@120 cogent analysis

If you take this seriously, voters for GWB are guilty of war crimes

OT but Jeff R. did write ‘complicit‘ – that’s not the same thing as guilt

I think any decent human being should be very uncomfortable with this new ploy by Democrats, to make it sound like not voting for them is ‘unethical’.

I don’t think it remotely appropriate for partisans to be judging the ethics of others when it comes to voting, because human beings are what they are, and the biases of partisans will always color their judgment too heavily for said judgements to be valid.

In the end, you’ll always be shilling, if even if you don’t know it.

“I don’t want to derail the thread, so this’ll be my only comment on this, but here’s the problem: If you take this seriously, voters for GWB are guilty of war crimes and—impracticality aside—ought to be sanctioned as such.”

Only if there’s no difference between moral culpability and legal culpability.

On the subject of open primaries, they are open in some states because the parties want (or wanted) them to be open. Each party is trying to capture voters from the other party for the general election. An open primary means an opportunity to get a voter from the other party to make an early commitment to a candidate (presumably these commitments prove quite sticky down the line), and may result in a general election candidate more appealing to voters from the other party. This may not make much sense in the current era of stark partisan divides, but things were not always this way.

It amuses me in the extreme whenever I note Americans debating the electability of possible presidential candidates, demonstrating again that they’ve apparently never noticed, despite continual reminders, that because of the intended-from-creation-to-be-antidemocratic Electoral College, and undemocratic-by-definition winner-take-all allocations in nearly all states, elections for the US president are a fecking lottery. Seriously, it’s like arguing about which is the better hand to toss a coin with.

I assume we have open primaries, where we do, to let, say, people registered in the Green Party to vote for a Democratic nominee. I used to have no party affiliation and would vote in the D primary, for example.

In addition to what Glen and Cranky said, in most cases Ann’s decision procedure will be more complicated. She and Mary may not agree how much better Sanders would be, or in what way. They may not agree about the relative chances Trump and Cruz have. The statistics may depend on people not crossing party lines to vote so that the assumptions fall apart. Ann, even, may agree on one level, and on another level distrust Mary and not be willing to follow her reasoning without checking. Etc.

Sorry about the multiple hosts, but one more thought: Ann and Mary both made their decisions mostly on an intuitive basis, but they weren’t satisfied with that. They wanted to be sure they were right, so they wrote down all their reasons for their proposed vote. Mary saw that her reasons mostly fell into the category “belief,” while Ann’s could be described as “strategic”. They could have then discussed the relative merits of deciding on belief and strategy, but instead they stopped right there. My question is, is Mary right to now take Ann’s partial self-analysis and use it to second-guess her decision-making, and is Ann rationally obliged to accept Mary’s asymmetrical view of the disagreement?

I’m not a member or supporter of either party — do you think I shouldn’t vote in either primary

No. Saying the primary is for ‘supporters of the party’ to vote for the leader they think will serve it best was overstatement (although I suppose it’s true of the Labour system). Perhaps I should have said ‘potential supporters’.

Bundling issues together creates distortions: two minority interest groups can (do all the time, actually) get together to enact their otherwise hopeless positions by scratching each others’ backs. The Founders may have been against political parties, although right away they formed some up … no doubt it’s a powerful tactic (Arrow Theorem at the limit), but in the Philosophers’ Congress, everyone should just say what they think.

Defensiveness in general is a problem in this crowded world. People demand an exclusion zone around the things that they actually care about and there isn’t enough room, so they end up arguing to the death about moot points. Because they don’t/can’t trust each other. Begging y’all’s pardon, but Faith can help here.

Just to be clear this has been a great thread, thanks everybody. Even people going ot has been ok.
So I have an OT response to Suzanne’s second comment. I attended a Bernie rally the other day (with some Bernie supporters). His speech was I think his standard stump speech (the same one he gave in June pretty much, though he is a tiny bit more polished). My impression was that he was trying to teach the largely young crowd how politics is done — how movements get results, if you like. He emphasized “this is not about Bernie” and never mentioned HRC, focusing his fire on Trump. Maybe I’m being optimistic, but it seemed like someone who does not expect to be the candidate, and wants his legacy to be a generation of activists like the Upton Sinclair campaign, or “get clean for Gene” generations.

p.s. my state, afaik, has a closed primary: must be registered as a Dem to vote in the Dem primary, registered as a Repub to vote in the Repub primary. Registered Independents or third-party people can vote in the general election, of course, but not in the parties’ primaries.

my state, afaik, has a closed primary: must be registered as a Dem to vote in the Dem primary, registered as a Repub to vote in the Repub primary

But then somebody with Democratic views could still carry out Harry’s plan by registering Repub to back Trump, could they not?

@engels
But then somebody with Democratic views could still carry out Harry’s plan by registering Repub to back Trump, could they not?

Yes, but it can be I think a bit of a pain to register Repub then switch to Dem when the same person decides he/she wants to vote in the Dem primary. Perhaps it’s easier now than in the old pre-computer days. Still, in many or most states one has to register at least several wks in advance of the election date. So while it’s possible, it wd require more time and trouble, and therefore more motivation, than for someone in Wisc., who doesn’t have to do anything except show up at the polling place.

Game Theory and ethics don’t mix very well, because the hypothetical utility-maximising agent postulated by Game Theory is not a particularly ethical actor.

But in any case, I think there is a problem with the “sabotage the other party’s primary” strategy even viewed purely in utility-maximing terms: it ignores the possibility that your opponents can also change their strategy in resp0nse.

The way you’ve set up the game, the Republicans are vulnerable to primary-sabotaging (because they have a candidate that a lot of Republicans really don’t want) but the Democrats aren’t (because all their candidates are viewed as sort-of OK by their actual members). I claim this hypothetical set up would not persist for long if primary-sabotaging became a commonplace tactic. The Democrats would also gain an “as crazy as possible” candidate that was supported by Republicans intent on sabotaging the Democratic primary, Maybe Pigasus makes a comeback from 1968. So if you’re only allowed to vote in one primary, you have to choose whether to use your vote to sabotage the opponents primary or to prevent your own primary being sabotaged. The risk is that it ends up being a Trump vs Pigasus election. (Ok, I think I’d stay up to watch that presidential debate, but it’s maybe not what the hypothetical players want). It seems plausible that “defend your primary against sabotage” wins out over “sabotage the opponents primary”. (I haven’t done the game theory math to check — you’ld need to put some numbers of the utilities)

Harry @104 You’re quite right. Sloppy reading on my part

How about the question of whether it would be ethical for Sanders run as a third-party candidate, something that probably can’t be counted out as a columnist in the Huffington Post suggested in light of his attack on Clinton not being qualified for the White House? Wonder what harry b thinks about his @145 now: ‘He emphasized “this is not about Bernie” and never mentioned HRC, focusing his fire on Trump.’ !!!

Harry, duly noted; LFC, thanks for the explanation

The term “strategic voting” is doing an awful lot of heavy lifting in this post:

1. Voting for a candidate who is not your top choice but nevertheless shares most of your policy preferences, based on electability grounds
2. Voting for a candidate you find loathsome against another candidate you find loathsome in the hopes that both lose

#1 falls under an extremely loose definition of strategic voting. Electability is in fact a valid consideration when choosing a candidate. Put it this way: assuming write-ins are allowed, would you accuse Mary of voting strategically because she is choosing one of the two candidates on the ballot? Why stick to Bernie at all, when she can instead write in the person in the world she believes would be the best president?

Even assuming Ann is a purely strategic voter, her vote would be dictated by the probabilities she assigns to the five judgments. If either the probability of (2) is high or the probability of (4) is low, then voting for HRC is the dominant strategy. There are very good real-world arguments in support of (2) being likely and (4) being unlikely, so this may in fact be the calculation many voters are making.

Finally, let’s not suppose that voters are operate purely logically, as evidenced by the “Bernie or bust” contingent. Voting functions very much as an expression of solidarity or values, and voting for Trump would be a tough pill for most.

Mary wants the “best chance of producing the best outcome” – but I’m not sure the provided list of assumptions gives us her utility function. If (5) is intended to mean that she has decided that maximizing the expected vote margin (D vs R) “trumps” all other considerations… Well, we can set the relative weight of the vote margin to infitely higher than the quality of the President. Then (1) and (3) are completely irrelevant, and yes, then (4) suggests that voting for Trump is the best strategy.

Most people do consider the quality of the eventual President-elect to be relevant to this decision. In their judgment, even “most important” needs be balanced against “still pretty damned important.” A bit more formally, (1) and (3) have a non-negligible impact on whatever bottom-line utility is being considered, and the best strategy becomes a lot less obvious.

Did anybody note that this time, more votes were cast in the Republican than in the Democratic primary? Abut three times more than in 2008. Any explanations?

It’s exciting!

@ 157: It may not demonstrate anything, except that he has a thin skin and is liable to shoot from the hip.

It does suggest what a Sanders/Trump campaign might be like. “Grumpy Old Men III: Race for the White House.”

@145: I’m sure that’s part of what’s motivating Sanders, and very laudable it is. If this doesn’t get ugly, he not only returns to Congress as a greatly empowered voice for his issues, but possibly with a major legacy in the form of young people inspired by him to become engaged with left/liberal politics.

What I think everyone is missing is whether someone will push Chris Christie off the George Washington Bridge, onto the car of the Republican nominee, thereby saving all of the children of the world

@168: True. Trump’s behavior is more juvenile than mature. Sanders and his surrogates whining that “She started it!” is a response that one could have expected from Trump, I regret to say. And beyond that, it’s plain false – by Sanders’ own account, he read a misleading headline in The Washington Post and flew off the handle. It’s entirely possible that he has just managed to disqualify himself.

It doesn’t seem that unethical to turn the primaries into a vote against the candidate that you really don’t want to be president.

Thus, if your preference order is: Sanders, Clinton, Trump, Cruz and you *really* don’t want Cruz to win, voting your 3rd preference (Trump) in the primary with the intention of preventing Cruz making it to the next stage seems reasonable. (And if the Republicans are doing a similar tactic, then the final election is a run-off between the compromise candidates, which doesn’t seem too bad a thing to happen). You’re even (sort of) voting in accordance with your actual preferences, i.e. Trump > Cruz.

On the other hand, suppose your preference order is Sanders, Clinton, Cruz, Trump and you really don’t want Trump to win, but suspect that a lot of Republicans share this sentiment, so you vote Trump in the primary to eliminate your 3rd choice, with the intention of making it more likely that your first or second choice will win. This seems a bit risky, if nothing else.

P.S. It just occurred to me that the Yippies could have nominated an actual Yellow Dog as a presidential candidate in 1968. You could probably still put in a write-in vote for one…

What Adam said @161. Man, I wish I’d written that post.

The first answer (if I’m playing by the rules) is that quantification of the differences matters a great deal. If (3), (4), and (5) are all true but only marginally (i.e., if Cruz would be only slightly worse than Trump, it’s only sorta more important to secure a GOP loss than the best president, etc.), while (1) is true far more significantly than (2), then the math would favor Mary’s ballot for Sanders. If the delta of (2) is significantly bigger than that of (4) (i.e., Clinton’s electoral advantage is much bigger than Trump’s electoral disadvantage) and (5) is at all true, then Ann’s ballot is correct. If delta-(4) outweighs delta-(2), though, Mary’s advice for Ann is likely better.

But the real answer involves my not playing by the rules and observing this: Human brains are not, as an empirical matter, rational outcome maximizers, or at the very least not very good ones. Democratic voters are going to vote in the Democratic primary, irrespective of the mathematics described above, because they feel that it’s the right choice for them.

Depending on your perspective, you could attribute this to a rational lack of confidence in their judgements about (1)-(5), or at least (2), (4), and (5), or whether it’s a morality/fairness judgment, or whether you follow the authors of The American Voter in believe Party ID drives other beliefs, rather than the other way around (is it obvious which alternative I prefer?), you could tell different stories about what “real” internal calculation is going on here. But as Harry already noted, the bare fact remains that very few actual voters are going to engage in Mary’s calculus for Ann. (I would also bet that comparatively few voters are even going to engage in Ann’s wager. Primary calculus is mostly just a determination of “viability” by early primary states, and Bernie Sanders quite evidently cleared the hurdle that a candidate like, say, Kucinich did not. Voters who prefer Clinton will vote for Clinton, and those that prefer Sanders will vote for Sanders.)