User talk:Kristomun

From Electowiki
Jump to navigation Jump to search

Welcome![edit source]


Welcome to the new Electowiki! See User:RobLa and Electowiki:About for notes about the migration.

Let me know on my talk page if there's anything I can help with, if you know of any deleted articles on Wikipedia that could be resurrected here, etc. There's a To do list if you're looking for something to help with.

This runs on Mediawiki, so it has a lot of the same features as Wikipedia, but is missing a few (like automatic citation generation) and has a few that Wikipedia doesn't (like support for inline YouTube videos). Images from Wikimedia Commons can be used directly without needing to be uploaded. Templates can also be transcluded from other wikis, though it's better to export and import them.

The wiki is 14 years old and has never had a ton of activity, so the goals/policies were never really solidified. Your input is welcome, especially on how to handle the separation of biased advocacy from neutral informative content (which are both welcome). See Electowiki:The caucus for the discussion topics and Electowiki:Policy. — Psephomancy (talk) 18:24, 22 January 2020 (UTC)

Condorcet IRV page[edit source]

Thank you so much for going through & cleaning that up (removing LNH bits)! I had meant to go do it after our discussion but got sidetracked. Much appreciated! 🙏

--Marcosb (talk) 20:56, 24 September 2021 (UTC)

Category:Consensus multiwinner methods[edit source]

Does this mean they try to elect a bunch of utilitiarian winners instead of PR? — Psephomancy (talk) 04:32, 7 February 2020 (UTC)

Suppose you're in a party list setting and there are four parties (R, D, Greens, and Libertarians). A consensus method would choose an equal number from each party to sit on the council even though vastly more voters vote for R and D than Greens and Libs. The way this ties in to consensus is that if the council uses supermajority or unanimity for its voting process, then it's more important to have a wide variety of representatives than that they are proportionally represented. In the extreme case of unanimity, it doesn't really matter if the Democrats have 50% of the council or 10%, because a single Democrat can block a decision. Instead it's more important that every shade of Democrat (or Republican, Green or Libertarian) who could block the proposal if he were part of the population in a direct democracy, can block the proposal on that council.
They can also be used for settings where variety is more useful than proportionality, e.g. if you're going to give everybody in a group the same set of movies, it may be more important that nobody finds the whole selection completely useless than that the selection is proportional, even if that means that a majority might only have a few movies they're interested in.
This should probably go on a page, but I haven't found enough sources... and I don't think I can source the name "consensus method" anywhere. Kristomun (talk) 09:14, 7 February 2020 (UTC)
Hmm, ok. To me, "consensus" is a term used in single-winner methods to indicate that the winner is a good representative of the entire electorate, like a Condorcet winner or utilitarian winner, "centrist" relative to the electorate.
Borda count is described as a consensus method, for instance:
Parker Friedland has a somewhat different definition here:
And so a multi-winner consensus method would choose a group of "centrist" candidates like this instead of trying to achieve proportional representation. So what Clay Shentrup prefers: "If you think PR is good, you say the three most outer. If you're more like me (skeptical of PR), you say the three center (just barely partisan) are best." (talk) 17:19, 7 February 2020 (UTC)
Centrist multiwinner methods, I'd call majoritarian (although that might not work for Range as it doesn't pass the majority criterion to begin with). A method like Minimax approval is rather "anti-centrist" - compared to proportional representation, centrist ones favor large blocs and minimax/consensus ones favor small ones.
I don't have much sunk into the name, so if you could think of a better one, I wouldn't have a problem changing it. I guess the problem is that it's hard to summarize just what these councils are (nonproportional, but not majoritarian/centrist), so I just went with consensus due to its use when the council uses a consensus/supermajority/unanimity method to vote on issues. Kristomun (talk) 17:40, 7 February 2020 (UTC)
Perhaps, for the lack of a better term, we should just call them minimax multi-winner methods. I'll think more about it. Kristomun (talk) 10:08, 8 February 2020 (UTC)
On the other hand, that may be too narrow since weighted methods like Chamberlin-Courant tend to elect these sorts of councils, then assign proportional weights to each winner's voting power. "Small-bloc" / "large-bloc" methods, perhaps? Kristomun (talk) 11:55, 8 February 2020 (UTC)

Quick intro[edit source]

Hi User:Kristomun, I'm 99.9% sure I know who you are from your username, but I want to respect the privacy of users who wish to remain anonymous (e.g. just like I respect the privacy of User:RRichie over on Wikipedia, even though the username, the talk page, and the edit history all give a hint or two). It's helpful (but not required) to know identities to patrol for conflict of interest on open wikis like this. Regardless, could you fill something in at User:Kristomun, even if it's "I'm User:Kristomun and I sometimes edit Electowiki", just so that your user page isn't a redlink? :-) -- RobLa (talk) 00:35, 31 March 2020 (UTC)

I thought it'd be pretty obvious who I were from my nick. If I wanted to be anonymous, I would've chosen something more... opaque, like "the anonymous mouse" :-) Anyway, I've created a page now. Kristomun (talk) 11:19, 31 March 2020 (UTC)

Voting theorists Page[edit source]

Kristofer, I would support your inclusion on the Voting theorists page if you would flesh out your user page with your contributions to make them clear. I know much of it has been simulations which may not really count but there is one part you mention I am super interested in. You mention the tradeoff between representativity and utilitarian efficiency of multiwinner methods. Can you make an electowiki page detailing the results? Is it possible to add some of the more modern systems to that analysis like SSS, SMV and Allocated Score? --Dr. Edmonds (talk) 16:56, 26 August 2021 (UTC) 🙏

To not toot my own horn too much, I've asked someone else what he thinks are my most important contributions. I haven't yet heard back from him and so I sort of forgot about this. But I would say (in no particular order): the proof that reversal symmetry, Condorcet, and DMTBR are incompatible; the exact minimum manipulability calculations for ranked voting with few voters and three or four candidates; and my "manual DSV" posts about the limits of methods that seek to circumvent Gibbard-Satterthwaite by failing universal domain. I don't think any of these are on the scale of, say, Green-Armytage's contributions, though.
Regarding your question, it would take a pretty deep refactoring of my code, because it's based around ranked voting data structures. There's no theoretical barrier to applying the tradeoff calculations to cardinal methods, but in practice a lot of coding would have to be done to make it work. Making a page on the results should be relatively easy; I'll try to remember to do that :-) I could also add those other results to electowiki (e.g. the strategy results or the rev. sym. proof). Kristomun (talk) 20:58, 14 October 2021 (UTC)