STAR vs. Score

Jack Waugh

@jack-waugh said in STAR vs. Score:

Americans seem to have mastered strategy for FPtP pretty well.

I don't see that as a positive.

No doubt you have had more important things to do in your life than to remember where this discussion was. I wasn't arguing that any effect of the use of FPtP was a "positive." I was arguing that mastery of its strategy is evidence that voters eventually catch on to the strategy that works for a given voting system.

Score may or may not have similar issues. But it certainly won't always pick the Condorcet candidate if one exists, and I see that as a big problem.

Then demonstrate the problem. Show a case where STAR does better.

The bigger issue, assuming you are interested in actually seeing any of this stuff in action as opposed to seeing another 20 years of debating and nitpicking, is that Score has been around a long time and made no progress. While the things that bother me about it may not bother you, they clearly bother other people.

This is not a debate about what to try to market. It is a debate about whether a complex system works better than a simpler one. How can any of us think clearly about comparing these systems if we don't address even the simplest questions related to them?

rob

@jack-waugh said in STAR vs. Score:

I was arguing that mastery of its strategy is evidence that voters eventually catch on to the strategy that works for a given voting system.

And... I don't see the need to master strategy as a good thing. Especially for a new system. The method should do the work for the voters, and shouldn't give an advantage to those who are best at guessing who front runners will be.

My point is, yes I am aware that is how it works under FPtP. It's one of many reasons I think that system is awful.

Jack Waugh

"I don't see the need to master strategy as a good thing. Especially for a new system. The method should do the work for the voters, and shouldn't give an advantage to those who are best at guessing who front runners will be."

This is not a choice we have. Allow me to introduce you to the Gibbard Theorem.

"My point is, yes I am aware that is how it works under FPtP. It's one of many reasons I think that system is awful."

I agree. The problem with FPtP as compared to alternatives is not that it encourages strategy, and that they don't. The problem is what global consequences result from the strategy it encourages.

Dear, beloved comrade, I'm asking you to stop beating around the bush and if you know of a case where STAR outperforms plain Score, exhibit such a case.

rob

Edit: I just noticed this is a reply to something from 24 days ago. Weird. Anyway, in case you still think this...

I'm well aware of the Gibbard Theorem. Did Gibbard also provide the profound insight that even if you wear a seatbelt, you can die in a car crash?

Just because perfection isn't attainable doesn't mean everything is equal. Some systems are highly resistant to strategic voting, some far less so. I believe any Condorcet method would result in the vast majority of voters putting no effort into strategically altering their vote from what it would be if they were simply trying to honestly express their preferences. And for those few who do try to be strategic, the vast majority of them would not gain any benefit from doing so.

Also, and probably more importantly, under such a system parties would have little incentive to strategically nominate, by reducing the number of candidates from their party to one.

beating around the bush and if you know of a case where STAR outperforms plain Score, exhibit such a case

I've given one, please stop saying I haven't. Nader vs Bush vs Gore is an obvious one. I feel like I already explained it pretty well, but in case you didn't get it, I'll have another go.

Under Score, Nader being on the ballot would have caused many "N>G>B voters" to lower their score for Gore, so they could express that they prefer Nader to Gore. In other words, he'd be a spoiler just as he was under FPTP. (this would further mean that candidates like Nader would have felt pressure to do like all the others on the left did, which is run as a Democrat. Same would apply to right-leaning candidates. Which would leave the two party system entrenched)

Under STAR, all people whose preferences were N>G>B would be able to a) express that they like Nader more than Gore, in case Nader became a front-runner against Gore, and b) give 100% of their voting power to Gore, if Gore and Bush were front runners.

The problem with Score would be an even bigger problem in elections where the three front runners were closer (e.g., where lots of people thought Nader could actually win), or in elections where voters were less informed as to the likely outcome, such as local elections.

In the case of Perot - Bush Sr - Clinton, STAR would likely have allowed Perot to win, I suspect it would have been a much tighter 3-way race. People such as myself would have given Perot a 5, Clinton a 2, and Bush a 0. A lot of other people would have given Perot a 5, Bush a 2, and Clinton a 0. (note that Perot was a true centrist, in that he was almost equal in his appeal to Republicans and Democrats) Under Score, voters would have probably been foolish to not give one o the major party candidates a 5, since they'd be scared that Bush and Clinton would be front runners and they'd be wasting voting power. (as happened with FPTP)

In both of those races, under STAR, there would be very little incentive to vote with anything other than honest preferences. They would not have to worry about which two of the three would be front runners, since they'd be able to give the full power of their vote to their preferred of the front runners. Not under Score.

(all that said, I think STAR doesn't do so well if there are more than 3 viable candidates, which is why I prefer Condorcet.... but still, STAR is dramatically better than Score at reducing strategic incentives)

Jack Waugh

@rob said in STAR vs. Score:

Some systems are highly resistant to strategic voting, some far less so.

I suspect that voting strategy has not only a magnitude but also a direction. The measure, however you take it, of how "highly resistant" a system is to it would only reflect the magnitude, I suppose. That would leave the possibility that system A is more highly resistant, but the strategy it does allow pits individual interests against collective interests (like the Prisoner's Dilemma), but that another system B is less resistant, so allows strategy of greater magnitude, but the direction of the strategy would align the same way with individual interests as with collective interests. It is when those directions are not aligned that the harm happens.

As for your three-candidate scenarios, yeah, intuition says I could probably come up with an example illustrating your point, given the difference in number of rounds of tallying. But systems that lead people to think (correctly or not) that strategy in nomination is no longer helpful to their cause, will likely lead to races having more than three candidates.

rob

@jack-waugh said in STAR vs. Score:

But systems that lead people to think (correctly or not) that strategy in nomination is no longer helpful to their cause, will likely lead to races having more than three candidates.

You are right that STAR isn't as good with more than three candidates.

What it comes down to, from my perspective, is that to reduce the incentive to strategically exaggerate (*), you need to minimize any reliance on "strength of preference."

STAR does this by doing a pairwise comparison as the last step. A pairwise comparison by its nature doesn't consider strength of preference (as you can see when it is a 2-candidate race in simple FPtP).

Condorcet methods try to do it all with pairwise comparisons. This reduces incentive to exaggerate even further. But since we can't guarantee there will be a Condorcet winner, we'll never get it to zero.

However, my position all along has been that getting it all the way to zero would be nice, but isn't necessary. If you get it close enough to zero, attempts to be strategic will have just as much chance of backfiring as they have of helping. A Condorcet method, including one with a very simple "tie breaking" formula, is good enough. STAR may or may not be good enough. Score is not good enough. Again, this is my opinion, but I it does come from a pretty solid game theoretical foundation.

* technically, incentive to exaggerate isn't the only thing we are trying to reduce. We also want to reduce vote splitting, which creates the incentive to strategically nominate, which in turn causes partisanship and polarization. Finally, we also want to aspire to "one person one vote", so each person has equal voting power. All of these things are accomplished by reducing the consideration of "strength of preference" in the tabulation.