A Final Report on the Election

[Justice Soothsayer]

I refer you to my post on ThePrincess Speaks thread

http://forums.slcds.info/viewtopic.php?f=1&t=1591

Me too. Merge threads?

[Cindy Ecksol]

I'll also note that you're only analyzing one case: increasing the number of parties certainly clarifies the picture in a different way.

Not in the least. All my examples use four parties. Perhaps you mean increasing the number of voters?

Jon, you are trying to convince us that the system is broken by analyzing only one case: the case in which there are four parties. Future elections may have more parties...or fewer. As you point out, the impact of this strategy is stronger when there are fewer parties. My point is that if someone is truly concerned about this, it would probably be at least as effective to go out and found another faction with a unique agenda as to spend time arguing over whether the system is flawed and trying to repair it.

Now obviously with only 70 citizens, the idea of having six or seven factions is fairly ridiculous. But that just demonstrates that even in a very simple society like CDS, key issues are not independent, they are interlinked. Perhaps we should be dispensing with arguments over the details of the voting system and going ahead full-steam on a project to add a new sim or two and attract more citizens. That would increase the diversity of the population and probably generate opportunities for more factions to emerge.

In the meantime, I challenge you to come up with a system that is significantly "more democratic" than the one we have now. The only way I can think to do this would be to increase the "direct representation quotient" by permitting ranking of all candidates within all parties not eliminated.

How do I know that strategic voting did not change the outcome of this election?

You don't. And you also can't prove that it did without asking everyone to reveal their votes and state a faction affiliation. Furthermore, whether it did or not is not particularly important except to those who will have to figure out how to advise their faction members to vote in the next election. They will have to decide what they believe about the effectiveness of such "bullet voting" and faction members will have to decide whether they wish to vote strictly according to their faction's advice or pursue their own preferences. No coercion or ethical violation is involved either on the part of those who give advice on voting strategies nor those who actually cast votes.

So shall we get on to discussing something more productive? Personally I'm looking forward to hearing some interesting proposals at the RA session this weekend and watching to see how an RA with nearly equal representation from every party deals with forming effective coalitions to accomplish some real work.

Cindy

[Jon Seattle]

Jon, you are trying to convince us that the system is broken by analyzing only one case: the case in which there are four parties. Future elections may have more parties...or fewer. As you point out, the impact of this strategy is stronger when there are fewer parties. My point is that if someone is truly concerned about this, it would probably be at least as effective to go out and found another faction with a unique agenda as to spend time arguing over whether the system is flawed and trying to repair it.

Its an interesting theory, but I already founded one faction, and did my time building it. No need for another.

In the meantime, I challenge you to come up with a system that is significantly "more democratic" than the one we have now. The only way I can think to do this would be to increase the "direct representation quotient" by permitting ranking of all candidates within all parties not eliminated.

I continue to focus on the faction level elections. The trick NuCARE found is actually very well known and there are also a number of standard ways of dealing with it. All have plusses and minuses. I continue to support the basic idea of our system, that it should be parties that compete, not individual candidates in a popularity contest. Lots of people have tried to change this for a more US-style system, so far with very little success.

How do I know that strategic voting did not change the outcome of this election?

You don't. And you also can't prove that it did without asking everyone to reveal their votes and state a faction affiliation.

Oh I do, and its not hard to get a pretty precise estimate of this. You should be able to figure it out.

But I do want to make things clear, that I am using these small cases to help people understand what is going on. I've done a number of larger models at this point, and the results are not substantially different.

[Bromo Ivory]

How do I know that strategic voting did not change the outcome of this election?

You don't. And you also can't prove that it did without asking everyone to reveal their votes and state a faction affiliation.

Oh I do, and its not hard to get a pretty precise estimate of this. You should be able to figure it out.

Jon - It is pretty clear that if you really are doing more than speculating at this point (Cindy asked if you could prove it, and you are claiming you can), then you had names attached to votes you tracked. And this I do not believe you are trying to say.

If it is, I have misjudged you quite a lot.

[Jon Seattle]

No names are necessary at all, its a simple statistical test. You see, there are only so many ways you can fill in eliminated votes. So, the solution is trivial: have the software fill in the missing votes and look at the results. You can do almost the same thing with the posted frequency table.

One thing that is basic to the current software design, it does not actually assign Borda counts except in the reporting modules. That means that any number of ways of counting can be tested -- just plug in a new module. Thats what I did, and thank goodness, the elimination did not change the results. At all. This attempt failed.

But is also opens the door to some interesting experimentation going forward. The method of stochastic sampling, that is testing our proposed voting systems with made up voting populations, that nevertheless are generated by the computer in such a way as to have the very very similar statistical properties as the actual vote.

Now, how do I know it was a real attempt at manipulation? All the one-faction votes all arrived within 45 minutes, except for one, that arrived early next morning. (That, to say the least, is unlikely to be random.) And then, to my surprise, one of the people involved told me about the whole scheme. Shall I post a transcript of that discussion? (with permission.) S/he does a great job of explaining how it works.

[Jon Seattle]

I want to add one other correction for the people who have been spreading rumors. I really truly never seen the names attached to votes. I am not talking about some number of three-factions-eliminated votes among some others. There were no others.

1. The vast majority of voters did not do this. Thank goodness. Most of us were generous with our points. It is remarkable in my mind that not many people eliminated any.

2. All the three-factions-eliminated votes fell into one cluster of 45 minutes, followed by one early the next day. There were no others. period.

[Cindy Ecksol]

I want to add one other correction for the people who have been spreading rumors. I really truly never seen the names attached to votes. I am not talking about some number of three-factions-eliminated votes among some others. There were no others.

1. The vast majority of voters did not do this. Thank goodness. Most of us were generous with our points. It is remarkable in my mind that not many people eliminated any.

2. All the three-factions-eliminated votes fell into one cluster of 45 minutes, followed by one early the next day. There were no others. period.

Time to get on with it, Jon. If you object to the use of these tactics, how are you proposing to change the voting process so that they are no longer available? You can't keep saying things like "Thank goodness only a few people did this!" if you (a) voted for the law that defined the process, (b) agree that there is nothing in the law to prevent a voter from eliminating factions, and (c) like the idea of faction-oriented politics (which implies that a faction will plan and act as a group) vs. individual-oriented politics. No one who eliminated other parties was acting "unethically" by any definition that I'm aware of: they were acting within the law and voted as they saw fit. That's not to say that you might not agree, but if you DO think that voting in this way is "unethical" then it's up to you and the others who agree with you to propose alternatives. But it's time to stop with the politics of slander and innuendo (which I and many of MY friends consider "unethical") and with discussing data in a way (chronologically sequenced) that you have already said you would like to eliminate and move on to fixing the problems you perceive.

I'll also note that you're still appealing to authority in your arguments and assuming that we'll fall for the fallacy and accept your analysis as the only possible way of understanding the election results. You're still the only one that I'm aware of who has access to the raw dataset. Perhaps it would be a good idea for Sudane to authorize you to publish it to the community so that we can all draw our own conclusions rather than relying on your self-assumed authority.

Cindy

[Jon Seattle]

I'll also note that you're still appealing to authority in your arguments and assuming that we'll fall for the fallacy and accept your analysis as the only possible way of understanding the election results. You're still the only one that I'm aware of who has access to the raw dataset. Perhaps it would be a good idea for Sudane to authorize you to publish it to the community so that we can all draw our own conclusions rather than relying on your self-assumed authority.

What does Sudane have to do with it?

Here is an excerpt from my 13 Jan letter to the SC. As you can see, I have been asking to release data since before the voting period. I have not yet gotten permission and I think it may take some discussion. Of course, the SC has access.

from Jon Seattle <[email protected]>
to Gwyneth Llewelyn <[email protected]>,
Claude Desmoulins <[email protected]>,
Justice Soothsayer <[email protected]>,
date Jan 13, 2008 6:11 AM
subject Election Software Attached
mailed-by gmail.com

Hi All, ...

With your permission, I would like to release the vote database
without voter identifiers as soon as the election is over. This will
help with analysis, and may also increase confidence in the outcome.

Best,
Jonathan

[Cindy Ecksol]

I'll also note that you're still appealing to authority in your arguments and assuming that we'll fall for the fallacy and accept your analysis as the only possible way of understanding the election results. You're still the only one that I'm aware of who has access to the raw dataset. Perhaps it would be a good idea for Sudane to authorize you to publish it to the community so that we can all draw our own conclusions rather than relying on your self-assumed authority.

What does Sudane have to do with it?

Sorry, meant Gwyn, but you know how it is with newbies getting these things wrong all the time.

Here is an excerpt from my 13 Jan letter to the SC. As you can see, I have been asking to release data since before the voting period. I have not yet gotten permission and I think it may take some discussion. Of course, the SC has access.

from Jon Seattle <[email protected]>
to Gwyneth Llewelyn <[email protected]>,
Claude Desmoulins <[email protected]>,
Justice Soothsayer <[email protected]>,
date Jan 13, 2008 6:11 AM
subject Election Software Attached
mailed-by gmail.com

Hi All, ...

With your permission, I would like to release the vote database
without voter identifiers as soon as the election is over. This will
help with analysis, and may also increase confidence in the outcome.

Best,
Jonathan

Great! Hopefully the SC will discuss it soon and release the data. Although you've already released quite a bit without authorization, of course.

Cindy

[Jon Seattle]

I'll also note that you're still appealing to authority in your arguments and assuming that we'll fall for the fallacy and accept your analysis as the only possible way of understanding the election results. You're still the only one that I'm aware of who has access to the raw dataset. Perhaps it would be a good idea for Sudane to authorize you to publish it to the community so that we can all draw our own conclusions rather than relying on your self-assumed authority.

What does Sudane have to do with it?

Sorry, meant Gwyn, but you know how it is with newbies getting these things wrong all the time.

Claude, in this case.

Great! Hopefully the SC will discuss it soon and release the data. Although you've already released quite a bit without authorization, of course.

To the contrary. I have been releasing data with authorization.

[Cindy Ecksol]

Great! Hopefully the SC will discuss it soon and release the data. Although you've already released quite a bit without authorization, of course.

To the contrary. I have been releasing data with authorization.

You lost me here, Jon. I thought you were saying that you had not received authorization from the SC to release the dataset and were still waiting for it. Are you saying that you had authorization to release data that you had massaged but not raw data? That strikes me as extremely odd, and not good for the health of our nascent democracy. There must be a misunderstanding somewhere between you and the SC...

Cindy

[Jon Seattle]

Great! Hopefully the SC will discuss it soon and release the data. Although you've already released quite a bit without authorization, of course.

To the contrary. I have been releasing data with authorization.

You lost me here, Jon. I thought you were saying that you had not received authorization from the SC to release the dataset and were still waiting for it. Are you saying that you had authorization to release data that you had massaged but not raw data? That strikes me as extremely odd, and not good for the health of our nascent democracy. There must be a misunderstanding somewhere between you and the SC...

To the contrary. The SC is concerned about the problem of individual voters being identified, especially by people watching the polls. This issue is still being discussed among SC members. I presented the aggregates to the SC before releasing them.

The fact is that while I gave the SC quite an earful, I was very very careful with handling the data. Altogether, with software development, and detailed analysis it was huge volunteer effort, documented, an supervised at every stage by the SC. My proposed security bill is mainly a description of what was actually done to protect the integrity and privacy of the vote.

You seem to imply that there was some impropriety, but there was not. The SC and their designates, as the people who run the election, are allowed to hear about the status, progress, and have any question they want answered in full detail. Anything less would allow future elections to be manipulated, far beyond strategic voting, in the dark. I am sure that is not your goal.

[Claude Desmoulins]

Any electoral system tries to balance allowing the voter to express their preferences as accurately as possible with trying to shape the process to encourage certain outcomes. The Saint-Lague method encourages coalitions and makes it hard to get a single party government if there are more than two factions. Letting voters exclude pushes the pendulum a bit the other way. It's a matter of where we decide to strike the balance.

[Jon Seattle]

Any electoral system tries to balance allowing the voter to express their preferences as accurately as possible with trying to shape the process to encourage certain outcomes. The Saint-Lague method encourages coalitions and makes it hard to get a single party government if there are more than two factions. Letting voters exclude pushes the pendulum a bit the other way. It's a matter of where we decide to strike the balance.

After thinking and experimenting with the system quite a bit I will say that I have to agree with an important part of this: that having a more expressive system is an important goal, especially given the Saint-Lague method.

The question is how can we do that and still keep the system comprehensible, not only in the simple way of how many points are distributed to each faction, but the kinds of issues that come up when you think about the interaction between each donation of points, and also between different voting strategies. Right now we are asking a whole lot of complex calculation on the part of each voter if they are really to know the consequences of their vote.

Borda with elimination has a very strange property, that your votes against a faction (though elimination) get stronger and stronger the more factions you eliminate. I would be a lot more comfortable if eliminating faction A, withheld the same number of points from A, as eliminating A and B. I think DNate wanted elimination to work that way (thus the assignment of equal low scores), but because elimination cuts the Borda point sequence in different places, it gives more power to those who do mass elimination.

One approach to this that could be interesting, might be voluntary. Start a movement of factions to promise not to organize strategic votes, and a movement of voters to promise not to eliminate factions who pledged and have a history of not using strategic voting. In effect this creates a bubble of voters and factions who can use the preferential system with some level of confidence. Of course a faction might break their promise, but if the votes end up being released after the election that should be pretty clear.

[Beathan]

Jon --

You have made the point I have been trying to make exactly. I think that we should point out the ethics of voting without changing the law. Thus, ethical action should be voluntary -- not forced by legal constraints. It should be enforced through electoral action -- voting against unethical parties (any party refusing to pledge not to abuse the system).

Beathan