Forum:"The voting method is bad!!!"

From Meta Wiki
Jump to: navigation, search
Forums: Index > "The voting method is bad!!!"
Archive
This page or section is an archive.
Please do not edit the contents of this page.
This thread was archived on 10 March 2019 by Cqm.

Hi all,

During some of the discussions regarding the future governance of the company, a lot of users brought up the opinion mentioned in the title. While I agree that it has its flaws as any voting system does; I find it misinformed that said users insist there must be better voting methods due to perceived problems. I would like to use this thread to clear up some of the misinformation and provide insight into the mathematical model behind it. I would also ask any commenting user to read through this post and the sources provided before commenting as there will not be a point to a discussion if facts aren't laid down first.

Background

There is a lot of mathematically well-defined criteria to determine the fairness of a voting method out there. I do not expect anyone reading this thread to read through all of them but I recommend taking a quick look at a couple of them at least as I think you will appreciate that it would seem common sense that these should be respected by any voting system.

I have some bad news. As proven by Arrow's impossibility theorem it is literally mathematically impossible to have a perfectly fair voting system. And when I say mathematically proven, I mean mathematically proven. This is not a matter of opinion but of fact. As long as we are using voting to make any decision, there are going to be major issues. You could, of course, argue that we should not use voting at all but this thread is not intended for that.

The Schulze method

The voting system we currently have place for the elections is called the Schulze method. As you probably already know, it involves the voters submitting the list of options ranked by preference. This is then converted into paired comparisons of options in the background based on the ranks the voter provides. In a nutshell, what it actually uses looks somethinmg like A>B, A>C, B>C rather than the A>B>C input. While it is also possible to have the former to not lose as much information in the voting process, it is a lot simpler to the voters to understand what they are supposed to be doing with a ranked list. Given the confusion over the method that's already here, I think it was the right call to use such a method.

The aforementioned pair comparisons are then added together from the votes of every voter and a winner is determined for every pair match-up. The amount of votes that ranked the winner higher than the loser in any of these match-ups is then considered the strength of this victory. What we now have looks something like A>B (25), A>C (20), B>C(15). This is then converted back into the final output of A>B>C calculated from the strength values of each pairing. This is how we can use it to choose any number of options from a higher number of potions.

I have skipped some of the finer details of how the background calculations are done as this thread is intended as a simplified explanation of the process. Anyone interested in the specifics can visit the Wikipedia article for the Schulze method linked above. However, I can assure you that each step of this method is mathematically sound.

But it doesn't have...

I intend to use this section to clarify why the deficits brought up by certain users aren't actual problems. I will update this part as more are brought up.

A way to oppose - Only in the sense that it also doesn't have a way to support anyone, strictly speaking. Otherwise, you are essentially opposing any option you rank low much like you are supporting any option you rank highly - this is literally the point of any ranked voting method. In the Schulze method specifically, any option you rank low will perform poorer in terms of the pairwise match-ups and be less likely to win as a result. I do not know how exactly people want an oppose vote to work, but I also do not know how anything could be satisfactory if this isn't.

Conclusion

While there are definitely valid alternatives, the Schulze method is a great method to use for voting. Mainly, it satisfies a ton of fairness criteria while remaining relatively simple for anyone to vote on -- as the complexity lies instead in the calculations. As such, I find it incredulous there is such discontent over it.

I hope this was informative and thanks for taking your time reading this! bad_fetusTalk 16:46, 19 December 2018 (UTC)

Discussion

Comment - I may have missed the conversations, but what are the complaints here? Reading through the explanation, I get the feeling some voters feel some candidates are wholly unsuitable and are having difficulty expressing that. I think a review of the voting will be helpful after it has been completed to ensure it works well for us, but until that time I'm not sure what the problem or solution is here. cqm talk 16:56, 19 December 2018 (UTC)

It might indeed be that people thought some candidates are unsuitable that led to this thread. It was repeated a few times that the voting method is poor in yesterday's Discord discussions however and there seems to be a general lack of confusion as to how it actually works. I thought it was a good idea to try and explain a few things. bad_fetusTalk 17:56, 19 December 2018 (UTC)

Comment - I feel like this should have been an essay... Kate msg 17:43, 19 December 2018 (UTC)

It was indeed intended more as an informative piece than a discussion. It still seemed appropriate to place it here since there was some strong disdain against the voting method during yesterday's heated discussions. bad_fetusTalk 17:56, 19 December 2018 (UTC)
I don't suppose you have the chat log of what happened? You have made me interested in what went down. Kate msg 18:59, 19 December 2018 (UTC)

Comment: Agreed that this should've been essay as opposed to discussion. However, I would like to point out that a lot of United States board elections use this same method. I don't see this method as a problem. The community of course should have a say in which how voting should happen. This is our first election, and we are following guidelines and examples, with future elections I believe we, as a community, will come together to determine the best method for us. TylerJarret (Talketh) 18:16, 19 December 2018 (UTC)

Comment - with regard to opposing; the place to raise concerns if you think a candidate is not suitable to be on the board would be in the discussions which are ongoing alongside the voting. I get that people not feel comfortable publicly stating these things; but if there is that much of a cause for concern about a candidate then it is important that other voters should be made aware of it. IsobelJ (talk) 19:38, 19 December 2018 (UTC)

Comment - I said this over on Forum:2018 Weird Gloop board elections: results and post-mortem, but I'll repeat it here: This voting method makes sense and seems fine to me, especially after reading some of the Wikipedia links/explanations. However, I agree that this would be good to be written up as an essay (with any additions, as needed), so it can be easily referenced for future elections. -- F-Lambda (talk) 22:04, 2 January 2019 (UTC)

Closed - Discussion has long since ceased. cqm talk 23:42, 10 March 2019 (UTC)