Hacker News Reader

Comment by qwerty456127 | Hacker News Reader

> which system is the best

Can't this be reasoned scientifically?

Yes, and Arrow's Theorem states that there is no "best" - you basically get the choose-two-of-three: https://plus.maths.org/content/which-voting-system-best

Approval voting wasn't considered in that article.

https://en.wikipedia.org/wiki/Approval_voting#Compliance_wit...

Arrow's Theorem only applies to ordinal voting systems.

There are other voting systems that get around this. Kenneth Arrow himself was into non-ordinal voting systems.

Arrow's Theorem only considers single-winner systems. But most politics is done by the Parliament, not the President.

There is no need to use small, single-winner voting districts to choose the members of the Parliament. Make voting districts 5-10 times larger and use proportional voting. Arrow's Theorem is not valid for that kind of systems.

"there is no perfect system" != "no system is any better or worse than any other system"

http://ncase.me/ballot/ and featured here: https://news.ycombinator.com/item?id=13161396

This interactive graph predicting outcomes for different voting methods convinced me that this is much harder to do than it would seem. Very much worth checking out the link. I think maybe I found it on HN.

From blog:

The further to the right a voting method is, the more it "maximizes happiness" for the voters

Why should group satisfaction with the winner be the important metric? Surely good governance should be the most important metric. If the quality of government is too hard to measure then length of time before overthrow or revolution should be a good proxy.

Good governance is a loaded term. Good for who? Different stakeholders will frequently have opposing goals, desires, and needs. Cutting foodstamps program helps some by lowering taxes, but hurts others who need help. Then the argument moves into secondary effects, cutting foodstamps will make poor people better off because the economy will grow, or cutting foodstamps will hurt taxpayers because poor nutrition and financial instability fewer people will escape poverty. The problem becomes complex enough that no one really knows the answer and so even the smartest often just end up restating their preheld belief.

Markets are supposed to be efficient they match demand with production efficiently, or some such. Political demand needs to be matched with equal efficiency. Imposing good governance into a voting system will very likely create distortions.

Good for the country, obviously. If you can't define what good governance is, how can you possible come up with a way to maximize it?

The purpose of elections is not to be happy with the outcome of the election, it's to be happy with the policy that results from the election.

You seem to ignore my point about multiple stakeholders. I believe that the happiness measure is not a measure of outcomes but happiness with getting a representative closer to his/her choice or preference...meaning one that may best represent their views, desires, and needs in government. Tha seems like good governance.

You seem to ignore my point about the collective good. By definition measuring happiness is measuring feelings. Rule by emotion. When has that ever worked out well?

If closer to personal preference is better, shouldn't direct democracy - absolute majority rule - be the ideal? But it's clearly not the case, at least not in any sizeable country.

what does "good for the country" mean?

If welfare helps people in the short term, but reinforces cycles of poverty is that good for the country or not?

If nationalized medicine helps people in the short term, but eliminates the incentive for medical innovation is that good or bad?

> Good for the country, obviously. If you can't define what good governance is, how can you possible come up with a way to maximize it?

It's almost like this is a really complicated problem that the best and brightest have been thinking about for centuries.

It's almost like "happiness" with who won one election is a gross oversimplification and fairly meaningless metric. Sounds like we're in agreement.

Only if you can define your criteria for "best".

Approval voting is biased toward "moderate" candidates. This has all the problems of first-past-the-post in terms of electing representatives who do not allow for a wide range of political opinion. It empowers status quo politicians beholden to special interests.

Is it really? Or is that your take? You could just as easily say that it's biased toward candidates who work effectively with a broader range of people.

We have observed reality to draw conclusions from. Regardless of political affiliation, would anyone really argue that special interests haven't taken over the political process in the US, and made significant inroads in most if not all of the industrialized democracies throughout the world?

Sure, I’ll argue that. “Special interests” is just code for “people whose opinions I don’t like.” But at the end of the day, our government very closely reflects what you’d expect. If you account for the fact that likely voters skew richer, older, and more conservative than the population as a whole, what exactly do you think would be different if “special interests” were not in charge? Would we spend less on defense? 1 in 3 Americans say we don’t spend enough. Would we spend less on social security, more, etc?

Removing wealthy 'special interests would probably have some interesting effects. Off the top of my head, I expect we'd have better regulations on banks and the like. And perhaps Mickey Mouse (and everything else more than 75 years old) would be public domain.

> “Special interests” is just code for “people whose opinions I don’t like.”

It's code for above-the-median-wealthy, enough to influence opinion outside of the populist opinion. Why you choose to go partisan over a systemic issue, is beyond me.

"Special interests" includes many powerful groups that aren't especially wealthy:

- religious groups wanting faith-based educational programs

- farmers wanting crop subsidies (hugely powerful in the US)

- internet users wanting net neutrality

- conservation groups wanting to preserve parks or endangered species

- unions wanting occupational safety regulations

- retirees wanting increased social security

You might even agree with some of these.

The common theme here is "concentrated benefits and diffuse costs".

For the quote of mine being referenced, I'm content with the dictionary.com definition of "special interest group":

http://www.dictionary.com/browse/special-interest-group

a body of persons, corporation, or industry that seeks or receives benefits or privileged treatment, especially through legislation.

The guilty flee where none pursueth. You'll note my original post was carefully non-partisan. Your rush to bring conservative talking points into the discussion suggests some kind of bias, or perhaps a guilty conscience.

Bias toward moderation is a trade-off. Attempting to minimize it seems naive.

At the extreme end, a legislature consisting of communists, anarchists, facists, theocrats, and hippies wouldn't accomplish anything but igniting civil war.

> communists, anarchists, facists, theocrats, and hippies

I would think all those groups together would only make what, max 20% of the US electorate? In a proportional system, they would still be a minority of the Parliament. And the internal disagreements within that minority of the representatives would prohibit them from joining forces and using even their 20% of the power. The rest 80% could mostly just ignore those fringe groups.

Or perhaps it would keep the really extreme and unworkable BS out of the law and government policy. Such a diverse body of legislators is going to have a much smaller set of policies they agree is a good idea, and that would result in a smaller body of law. I see a smaller, simpler body of law that more people can think and agree with as a Good Thing.

I think if you had a national legislature with "moderates" (as defined by California and Massachusetts) and "moderates" (as defined by Utah and Mississippi), you'd have a fairly wide range of political opinion represented. You wouldn't have the extremes of California represented, and you wouldn't have the extremes of Mississippi represented, but that may not be that much of a loss...

A wide range of political opinion is sometimes useful (e.g. in one of the houses of a legislative body). Moderate candidates are also useful. Rather than choose one voting system to use for everything, it might be appropriate to use a couple voting systems that have different strengths and weaknesses. So you might imagine one house of a legislature elected with approval voting, and a different house elected with a proportional method.

> Approval voting is biased toward "moderate" candidates.

Wouldn't that only be true if preferences are roughly normally distributed (or some other fat-middle distribution)?

And if that's the case, shouldn't the outcomes of the voting system tend towards a moderate middle?

If a moderate candidate is the one who is acceptable to the largest proportion of the voting public, why shouldn't they win?

Approval voting allows voters to vote for 3rd party candidates without "throwing their vote away". That would be a huge improvement over first-past-the-post, as 3rd party candidates would have a decent chance of winning if they can make a compelling case to voters.

People disagree on what the purpose of government even is, so... no.

And this is probably the origin of most political disagreements.

Reply on news.ycombinator.com