Discussion of voting systems

(Jesse) #1

[Admin: There’s been some discussion of voting systems on one of the keyboard layout threads. I’m reparenting that discussion over here]

Runic keycap layout
(Tim Neumann) #3

@howellbridger: I generally agree with your suggestion. However only asking the “satisfaction” question is not sufficient in my opinion: I, for example, would accept both Bamum and Mende, however I do have strong preference towards Bamum which I would like to have counted somehow.

So maybe just ask two Questions, both allowing multiple answers to be selected:

  1. Which are your preferred Layouts?
  2. Which Layouts would be acceptable to you?

(Michael Richters) #4

These questions can both be answered by ranking your preferences, and simply not ranking ones that you disapprove of.

(Bart Nagel) #5

Indeed – a system such as instant runoff voting can be effectively used for this.

In short, you rank your preferences. All the first choices are added up. If there is a majority, that’s it: you have your winner. Otherwise, the options with the least votes are eliminated. Then you add up the first choices again, obviously not taking into account any votes for eliminated options. You continue in this loop until there’s a majority or a tie.

(Gergely Nagy) #7

I have not read everything in this thread yet, but… this all seems far too complicated. In the end, the decision lies with @jesse and @kaia, so any vote is just an indication for them, something to help. So the primary purpose of any vote would be an outcome that helps them. It doesn’t have to be all that fair, or exact, or whatever. Just something that gives enough of a hint about what we’d prefer.

So a simple ranking of options sounds easy to implement, and the results can just be recorded, and a few stats run: which option appears first the most? which option wins if we add a weight for each rank? etc.

Or, even simpler: which option would you buy? It does not capture everything, but it captures enough to be useful, and yet, is simple.

Nevertheless, when it comes to voting systems, the primary goal would - in my opinion - have to be how useful the outcome is for @jesse & @kaia. Everything else is secondary.

(James Cash) #8

Agreed - I think a simple ordering of preferences that J & K can look at is probably the best. This isn’t actually a democracy, so we don’t really need a codified system for how votes will work; J & K just need to have an idea what do most people like/do any people hate a particular option.

(Jesse) #10

Yes! In the absence of an obvious consensus, we are going to pick up the one that we like better. If the two of us disagree, Kaia has one more share of stock. So she wins.

I think we’re going to be able to do additional rounds of keycap sets in the future. So, if your favorite choice doesn’t get selected, don’t freak out too much

(Michael Richters) #11

Very good point. Approval voting would seem to make the most sense from the point of view of the seller. Assuming perfect economic rationality. =P

(Chris Conley) #12

Yeah, IRV produces distortions in the results that Approval does not.

(Gergely Nagy) #13

Another thing to keep in mind, is that Discourse (the forum software) has some poll stuff built in, things like:

Single choice

  • Choice One
  • Choice Two

0 voters

Multiple choice

  • Choice One
  • Choice Two
  • Choice Three

0 voters

Number rating

Choice One

  • 1
  • 2
  • 3

0 voters

Choice Two

  • 1
  • 2
  • 3

0 voters

Choice Three

  • 1
  • 2
  • 3

0 voters

I’d think that using one of these may be best, because they are built into the forum anyway, and all of them are reasonably simple. Although I have no idea how useful their results are yet. Feel free to vote on the above polls, and we’ll see :slight_smile:

Runic keycap layout
(Jesse) #14

Oh my god yes. Built in polling is a huge win.

(Arlo James Barnes) #15

My favorite version of this is where you have two ranking columns, “explicitly approve” and “explicitly disapprove” (negative weightings) and “nobody” is an option (of course, Arrow’s :bow_and_arrow: impossibility theorem [see below] limits all voting systems in one or more of five ways). But @james.nvc makes an excellent point with

and even in a ‘democracy’, whatever that means, voting is not the best way to build consensus in small groups (sure, there may be 178 users, but only 16 of y’all so far have contributed a fair amount).

*Despite the forum guidelines’ insistence that this is a “public park”, the forum is clearly a “private thing”, not a res publica / republic / “public thing”.

(Arlo James Barnes) #16

(Jesse) #17

The guidelines suggest that you should treat this space like you would treat a public park. “Please treat this discussion forum with the same respect you would a public park.” Put another way: This isn’t 4chan or reddit. Please be nice and friendly and don’t be abusive or disrespectful to others. It’s ok to disagree with others ideas’ or statements, but only when done respectfully.

(Arlo James Barnes) #18

I had missed this nuance, but you are right, it is saying something a little different.

Why would I want to be anything else? That would be unpleasant for me as
much as for anyone else, and I wish no unpleasantness on the talented and
seemingly good-natured people in this forum :slight_smile:

(Michael Richters) #19

Not quite, as far as I can tell. It provides the average, and the number of voters, so it’s possible to calculate the total, though I don’t see a way for the poster to get the raw vote data (maybe an admin can; I’m no expert on Discourse). But I actually don’t care; despite some poorly-defined metrics that conclude that score voting is better for overall satisfaction, it’s really not the correct model for a product with no competition. Approval voting is obviously superior in this case, since there really are only two options: “Do I buy it, or do I not buy it?”

As far as I’m concerned, it’s only worth looking at preferences if there’s a tie on an approval vote first.