How should voting work?

[Valkenar]

Details are in the article.

Interesting article with similar information. One thing of particular interest was the mention of "Arrow’s Impossibility Theorem" I'll need to look that up.

Couple points. First, despite their generally somewhat fanatical praise of the Condorcet method they do have a reasonable counter argument to the simplicty and cyclical inconsistancy problems.

The first, which I fully agree with is that while the ballot may be large and cumbersome it doesn't actually need to be convincing. Just a simple thing like the following would do

On each line, fill in the circle next to the candidate you'd rather be president.
()George Bush or ()John Kerry
()George Bush or ()Carrot Top
()John Kerry or ()Carrot Top

It could be rather long, but is straightforward. As foro the length, the question is whether that would drive more voters away than our present system does? I don't know, but if this were the only problem I'd say it's worth having the more sensible system.

However I can agree that the cyclical resolution system, particularly the Schwartz Sequential Droppin method which provides the best results, will be confusing to a lot of people. This could be a problem if rumors surface and a believed that voting in a certain strange way is likely to benefit you somehow. While the system is fundamentally more fair than the other methods, I think it's a mistake to ignore the complexity issue the way my originally posted website does.

The bottom line for me though is that the benefits outweight the drawbacks, and some kind of public education campaign will help to make sure people understand enough to avoid voting in a way opposite to heir intention.

[Bill Glasheen]

The bottom line for me though is that the benefits outweight the drawbacks, and some kind of public education campaign will help to make sure people understand enough to avoid voting in a way opposite to heir intention.

I know what you are thinking, Justin. Correct me if I am wrong. The thing that upsets you the most is the scenario in 2000 where Nader took so many votes. Allegedly he prevented Gore from winning. This is way more complex than that. Many assumptions are made, including that the same number of people would show up to vote if there were only 2 candidates.

Clearly there is a problem with Plurality voting (the existing system). However education as you suggest may not give you the result that YOU want. Quite frankly a good third of the electorate are pretty ticked off that the major party candidates aren't really that different from each other on most issues. Many "faithful" on either side of the political spectrum express themselves by not voting at all, or making a "protest" vote like Nader or Bandnarik.

And you thought Florida 2000 was a nightmare? Picture the poor senior citizens who were confused with butterfly ballots or who couldn't punch a chad to save their lives. (I already voted BTW, via absentee method. My voting machine used the punch cards. And yes, I checked the card before putting in the machine.) This isn't an election with just 2 or 3 candidates, Justin. This election has 5 possible candidates, not counting the legal right to enter a write-in. Exclude the write-in, and you have 10 possible combinations. Include a write-in (just a SINGLE one) and that ups it to 15 pairs.

And that's just for ONE office!

Borda's method works really well. Approval voting also works well to get the least objectionable candidate. Either of these methods also would give a voice to 3rd party candidates that they presently don't have. And I view that as a very good thing.

- Bill

[Med Tech]

One direction I've been leaning a long time, would be to have run-off elections.

"Once again, beware of the law of unintended consequences. The problem with a straight voting Democracy and runoff elections is that the remaining parties have to "dance with the devil in the moonlight." Fringe parties get extraordinary power because they become bargaining chips for the votes they represent in the runoff."

Bill, do you have any examples of this happening? I have some niggling reservations about runoffs in the back of my head somewhere, but it's something I've only recently begun to doubt. I still tend to think that if nothing else, the major parties will just adopt just enough plank issues of the 3rd parties to nullify their uniqueness on the ballot. I don't know that newer parties would be able to swing their votes to the cause of another party effectively enough to become bargaining chips.

[Bill Glasheen]

They were discussing this as a "public service" story on XM radio a few days ago.

I believe Israel uses such a system.

- Bill

[Med Tech]

Borda Voting: Voters rank all candidates from first to last. A voter’s first choice is awarded a number of points one less than the number of candidates. The second choice receives one less point than the first, and so on. (The final choice receives no points.) Points are totaled and highest number wins.

I don't think I favor this one at all. Winning by mediocrity? Least objectionable? No, our candidates already seem to come from the bottom of the barrel as it is. We'd end up with another Jimmy Carter, or worse...

Approval Voting: Voters give a single point to as many candidates as they consider acceptable, and the candidate with most points wins. This process often leads to electing the least objectionable leader.

While this one may lead to the same mediocrity, at least it'd work for Florida, huh?

[Bill Glasheen]

Arrow’s Impossibility Theorem

[Bill Glasheen]

Here's an example of a modified Borda's method in use.

Associated Press Top 25 College Football Poll

It works. In this case, the coach or sports writer voting in the poll will rank his/her top 25 teams. First place team gets 25 points, second 24 points, ..., and 25th place team gets one point.


Note how well it correlates with team performance. It passes the "sniff" test.

- Bill

[Valkenar]

I know what you are thinking, Justin. Correct me if I am wrong. The thing that upsets you the most is the scenario in 2000 where Nader took so many votes. Allegedly he prevented Gore from winning.

No, not really. I do think that the result of that election has been an absolute disaster, but my biggest problem with our voting system is that it means there's no point in voting if you don't like the major party candidates. And the way it does this is two-fold. Both the plurality vote and the electoral college do this.

Combined these factors make it so there's all but no point to my voting. I will anyway, because I feel obliged, but what's the point? To send a message?

This election has 5 possible candidates, not counting the legal right to enter a write-in. Exclude the write-in, and you have 10 possible combinations. Include a write-in (just a SINGLE one) and that ups it to 15 pairs.

Yes, it's a hassle, but it's still extremely straight-forward.

And that's just for ONE office!

So use the best, most laborious process for the most important, national office and let the states decide how they want to send up senators as they see fit.

Borda's method works really well.

I have trouble picturing a ballot that implements Borda's method that is less confusing than the condorcet method. How do you suggest people go about doing the ranking? If they have to write numbers 1-10 down next to the names there will be an awful lot of duplicate and skipped numbers. Furthermore, ranking things like that is *hard* it would take me an hour to figure out how I want to rank the candidates. The condorcet method is tedious, but you're faced with a simple decision each time.

But the biggest problem I have with Borda's is that it rewards strategic voting. That's a very, very bad thing. Remember you asked what I don't like about the plurality vote? That's it, strategic voting.

That said, if the choice is plurality or Borda's, I'll choose Borda's every day.

As far as the electoral college goes, I think it's outdated and pointless, but it's still compatible with a single national election. Just multiply the values from each state by the weight you want that state to have.

[Bill Glasheen]

Some people like strategic voting. They feel that negatively about certain candidates.

Borda's method is very easy with a touchscreen electronic ballot. Teach people how to shuffle the selections from top to bottom with the touch of a finger. It's easy.

Approval voting is very easy to do, and leads to the least objectionable candidate. And considering how folks feel about all politicians, that about sums up their selection by any method. Using the Approval voting method, very few card ballots would have been thrown out in the 2000 election in Florida.

I agree with your comparison of Borda and Plurality. Borda really does work better.

But do not dismiss Plurality in combination with the present Electoral College system. It works pretty well. It acts to some extent as a low pass filter to states with extreme views (preventing tyrrany of the majority). And it makes politicians really pay attention to states that feel equivocal about the available candidates. That can be viewed as a good thing.

- Bill

[Valkenar]

Some people like strategic voting. They feel that negatively about certain candidates.

That's fine. Believe me, I understand well about feeling strongly opposed to a candidate. But these feelings are something that should be reflected by an honest vote and not doctoring your vote. Condorcet allows people to vote very strongly against someone, without allowing them to manipulate the system through voting other than they truly feel, which Borda does.

It should not be possible to use your knowledge of how others are going to vote to enhance your own voting power. This is gaming the system, and undermines the premise of fair elections.

Borda's method is very easy with a touchscreen electronic ballot. Teach people how to shuffle the selections from top to bottom with the touch of a finger. It's easy.

Yes, that could be done. However, despite my general opinion that technology should be used for everything it can be, I have major reservations about electronic voting. Also, as long as Diebold is being given the contracts I'll be against implementing it.

Approval voting is very easy to do, and leads to the least objectionable candidate. And considering how folks feel about all politicians, that about sums up their selection by any method. Using the Approval voting method, very few card ballots would have been thrown out in the 2000 election in Florida.

I think approval voting is agreat short-term fix. It could be implemented immediately and would be a vast improvement. There are worse things than getting the least objectionable candidate, though ultimately I think we can do better.

But do not dismiss Plurality in combination with the present Electoral College system. It works pretty well.

Sorry, but I'm dismissing it. Yeah, it could be worse, that's true. We could play giant games of twister, with the winners getting to cast their state's vote for president.

It acts to some extent as a low pass filter to states with extreme views (preventing tyrrany of the majority).

How do you define an extreme view? I agree that tyrrany of the majority is a very serious consideration. However, if the majority holds a certain view, then by definition, i t is not the extreme view. It can be wrong, it can be downright insane, but the word extreme fundamentally means outside the norm. If the majority holds a certain view, then that is the norm, not the extreme.

And it makes politicians really pay attention to states that feel equivocal about the available candidates. That can be viewed as a good thing.

I disagree. If I move to Wisconsin, have I become a better person, more deserving of political promises? Which states happen to be swing states is a statistical accident of population distribution. There is some extent to which people choose to live in a certain state, but to a large extent people end up where they end up.

The fundamental idea of protecting minority opinion from majority oppression is more than good, it's critical. But doing it by state is all but completely arbitrary at this point in history. And there's a point where it doesn't make sense. Should we give more voting power to african-americans, to these in the top 1% of wealth? Or the bottom 1%? Muslims? Jews? People in Rhode Island? Which minority attributes make a person deserving of extra voting power?

I don't have a really good answer to the question of tyrrany of the majority, but I don't think the electoral college is doing a sensible job of it anymore. In the past, perhaps, but now the US is much more homogenized.

(please don't construe this as an argument against State's rights. One of the things that makes this country work so well is that you can move to a state with somewhat different laws if you decide you can't stand those where you live)

[Bill Glasheen]

Justin, I think we agree more than we disagree. Our language is just different. Perhaps one or the other of us is just being less precise. (I'll take credit...)

It acts to some extent as a low pass filter to states with extreme views (preventing tyrrany of the majority).

The fundamental idea of protecting minority opinion from majority oppression is more than good, it's critical.

Same thing. I'm an engineer/physiologist, and you are a psychologist. Words have different connotations in our respective domains.

Condorcet allows people to vote very strongly against someone, without allowing them to manipulate the system through voting other than they truly feel, which Borda does.

Actually you are absolutely wrong on this, Justin. Condorcet alloows for greater flexibility. I can do anything you can do in Borda with Condorcet. But I can create some paradoxical outcomes in Condorcet (A > B, B > C, C > A) that you cannot possibly recreate with Borda. Indeed it's Borda's limitations that make it less likely to confuse and/or allow manipulation.

As I stated before, I now know how to manipulate the results of that Condorcet-based personality test so that I can show the test makers what I really feel about their stupid instrument, and their attempt to force me to say one thing (integrity) is more important than another (achievement). Trust me, I've researced this a lot. You cannot manipulate Borda the same way. Indeed Condorcet invented his method to ALLOW FOR the greatest possible flexibility on the part of the voter.

I don't have a really good answer to the question of tyrrany of the majority, but I don't think the electoral college is doing a sensible job of it anymore. In the past, perhaps, but now the US is much more homogenized.

The US is much more homogenized? You have GOT to be kidding. How much do you travel, Justin?

I do quite a bit of travel from Virginia to New England to Minnesota to Arizona to California to New York. Going from one state to the next can feel like going from one planet to the next. Do you have any idea what the average Virginian thinks of a New Englander's or New Yorker's point of view? Have you any idea how unique Minnesota is? Ever been to the wild and free state of Arizona, home of Barry Goldwater? Ever traveled to the superficial land of fruits and nuts? (South California) The ethnically mixed Asia outside of Asia? (San Francisco)

Indeed this election is a classic example of why the electoral college system is so important. Take a good look at that "Red and Blue State" map. Listen to Gene use the label "bible thumpers," and talk glowingly of his hero Kerry. And then see how those "bible thumpers" vote w.r.t. the choice of New England.

And by the way, ask folks in rural England how they feel about the folks in London preaching to them about how to handle animals. Ask Kami what it's like to live in rural Oregon, and have to put up with the politics of the city. The phenomenon is ubiquitous across all geographic scales.

Plus ca change, plus c'est la meme chose.

- Bill

[Valkenar]

Same thing. I'm an engineer/physiologist, and you are a psychologist. Words have different connotations in our respective domains.

Whoa, whoa, whoa. I am *not* a psychologist. I work as a computer programmer, and I happened to take some psychology in college. I also took philosophy, and was going to add that, so it's just as fair to say I'm a scientist or philosopher as it is to say I'm a psychologist. Incidentally, the #1 thing I got out of my study of psychology is that 90% of it is BS, in my estimation (though I didn't have much psychiatric/abnormal psych exposure which I suspect is better).

Otherwise I agree that maybe it's just a terminology.

Actually you are absolutely wrong on this, Justin. Condorcet alloows for greater flexibility. I can do anything you can do in Borda with Condorcet. But I can create some paradoxical outcomes in Condorcet (A > B, B > C, C > A) that you cannot possibly recreate with Borda. Indeed it's Borda's limitations that make it less likely to confuse and/or allow manipulation.

No, you're absolutely wrong.

In Borda it is an effective strategy to place likely winners lower on your list than your actual preference. For example, imagine a race between Kerry, Bush and Hitler. Kerry and Bush are close, Hitler has no chance of winning. My true preference is 1. Kerry, 2. Bush, 3. Hitler, but it is better for me to vote 1. Kerry, 2. Hitler, 3. Bush, because I know Hitler isn't going to win, and putting him before Bush will hurt Bush's chances. This is the strategic voting aspect, and the problem I mentioned. This is how Borda works.

Condorcet voting does not have this problem. It will not hurt Bush's odds at all to say that I prefer Hitler to Bush, because they are independant measures. As long as I say Kerry>Bush, it doesn't matter what I say about Hitler, because he cannot possibly win. This is why it is not possible to game the condorcet system this way. Can you think of any way to vote other than your actual preference that results in a greater likelihood of your true preference occuring? I would honestly like to hear it.

Finally, the paradoxical outcomes you keep mentioning are 100% resolved by the SSD method. Did you read that part? Repeat, the A>B, B>C, C>A situation is not a problem. It is fully solvable. Should I quote the methodology here?

As I stated before, I now know how to manipulate the results of that Condorcet-based personality test so that I can show the test makers what I really feel about their stupid instrument,

This is a totally, totally different. A personality test is nothing like the voting system in two important ways. Your ability to game a personality test is accomplished by recognizing what the test is really measuring with the question. With voting, the measure and the question are identical, there's no subterfuge.

More fundamentally, a personality test asks a number of different questions which are then aggregated to generate a vector in a smaller degree space. Condorcet voting does not do this, at all. If I have my terms right, and am not bungling my linear algebra (entirely possible), the condorcet method creates pairwise matrices of degree N, and thus N orthogonal vectors in N-space. Totally and utterly different from what a personality test does.

Indeed this election is a classic example of why the electoral college system is so important. Take a good look at that "Red and Blue State" map. Listen to Gene use the label "bible thumpers," and talk glowingly of his hero Kerry. And then see how those "bible thumpers" vote w.r.t. the choice of New England.

But that's just it. The individuals in those states would feel just as strongly if you put them on the moon. If I lived deep in Atlantis, I'd *still* be me, and still be against Bush. Put me right in the bible belt and it's still me voting.

But like I said, you can still just weight the votes of people in states you want to give preference to differently. You don't need the electoral college to do that. If anything creates tyrranny of the majority, it's the fact that in each state, the majority's opinion decides where all of that state's votes go, and the minority opinion is not represented nationally.

[Bill Glasheen]

My apologies for calling you a psychologist.

the #1 thing I got out of my study of psychology is that 90% of it is BS

Perhaps Dr. Giella might feel differently... But in conversations with you on other subjects, I can see that you feel this way.

Oye!

You're wrong!

No, YOU'RE wrong!

No....YOU are wrong!!



Sorry, Justin, I can manipulate with Condorcet with maximum effect. They give all possible combinations; that's the way the method works. In your example above, at some point I would have the option of choosing between Hitler and Kerry. And indeed I would tank the likely winner candidate I didn't want to win with every possible match - even if the alternative was my neighbor's dog who bit my mother. The net effect is the power to give the (wo)man who is most likely to defeat my (wo)man the lowest possible score.

I know a lot of people who feel this way about their candidates, Justin, and against the candidate they dislike.

There are variations on all these methods that can mitigate the negative effects. But the bottom line is that Condorcet gives the voter maximum flexibility to achieve his/her wishes (without added manipulation on the back end, which can be done with all methods).

- Bill

[Bill Glasheen]

But that's just it. The individuals in those states would feel just as strongly if you put them on the moon. If I lived deep in Atlantis, I'd *still* be me, and still be against Bush. Put me right in the bible belt and it's still me voting.

You are assuming no correlation between where you live and how "you" might vote. Completely statistically independent, eh? This doesn't even pass the sniff test, Justin. Do you know how we people who own property in Virginia feel about New York buying garbage dumping rights in rural Virginia counties? Do you know how Kami feels about the animal rights activists in Portland telling her how she can deal with the cute little fuzzy mountain lion that is mangling her goat out on the farm?

I can take some data and easily show the statistical interaction effect if you want to get geeky on me.

I'll let you in on a secret, Justin. When I was around your age, I voted for McGovern for president. Do you know anything about George McGovern?

I'm fed up to the ears with old men dreaming up wars for young men to die in.

It very much had to do with my situation in life, and not so much to do with Bill Glasheen's DNA.

- Bill

[Valkenar]

Perhaps Dr. Giella might feel differently... But in conversations with you on other subjects, I can see that you feel this way.

You don't have to appologize for calling me a psychologist, I just wanted to set the record straight so nobody would think I have credentials I don't. Note also that I don't think all of pschology is BS, and I think that psychiatry in particular does pretty well, to a good extent. I'm not trying to bash the field; Obviously I appreciated it enough to study it. But this is a topic for another time.

You're wrong!
No, YOU'RE wrong!
No....YOU are wrong!!

Glad you caught my point with that.

And indeed I would tank the likely winner candidate I didn't want to win with every possible match - even if the alternative was my neighbor's dog who bit my mother. The net effect is the power to give the (wo)man who is most likely to defeat my (wo)man the lowest possible score.

First, let me acknowledge that any voting system is going to be flawed, according to Arrow's Impossibility Theorem. However, my point is that the SSD form of condorcet results in a system that would succesfully prevent an individual from benefitting by insincerely voting for candidates who are hugely unpopular over a candidate they dislike. Under this system, for example, if you rank Nader over Kerry, it is not going to adversely affect Kerry's odds, because Nader is not going to defeat Kerry anyway.

Because of the way the vote is tallied, a strategic vote can only work if it is done by enough people that the weak candidate actually does defeat the strong candidate. But then it would be a very poor and risky strategy, because your preferred candidate might lose also in that case. Any voting system is going to be screwed up if everybody is trying to cheat it.

I probably overstated my case before, but the benefit of the SSD method is that it gives very good protection against strategy. It's not perfect, no method can be, but at the very least, it's less susceptible to strategy Borda.