Template talk:Comparison of voting systems

Not a general template with special treatment for Schulze

As this table is sorted by compliance, it does not serve as a general template when Schulze is out of sort order, at the top. Correct title would be "Comparison of Schulze to preferential voting systems" which doesn't warrant a "template." We could sort Schulze in natural order and keep title "Comparison of preferential voting systems." — Preceding unsigned comment added by Filingpro (talk • contribs) 21:24, 10 April 2016 (UTC)

Agreed. Alternatively, we could turn this into a template where the top row is the first argument given to the template, or maybe keep natural order and just bold it, so it can be re-used in the other articles. Mvolz (talk) 11:38, 12 September 2021 (UTC)

Column

Trying to add a Column for MinMax, but changing the template is not changing it on the Schulze Method page. Not sure why. Schulze Method passes MinMax criterion but Ranked Pairs (Tideman) does not, and this table should include that information to help differentiate the two. — Preceding unsigned comment added by 71.201.20.135 (talk) 17:53, 28 May 2016‎ (UTC)

LIIA

Ni! The page on Ranked pairs states that it fails LIIA, however in the table it is shown as passing that criterion. There are no references in either the page or in the table, so someone with a bit more knowledge or time and access to search the literature could investigate and fix it. Ideally, adding a reference on the matter to the page. Cheers! --Solstag (talk) 11:07, 3 May 2018 (UTC)

The page is on independence of irrelevant alternatives, not local independence of irrelevant alternatives. No preferential system satisfies the former but Ranked pairs and Kemeny-Young satisfy the latter. 1.126.108.98 (talk) 01:37, 6 October 2019 (UTC)

All right then, thank you. Solstag (talk) 14:15, 30 November 2020 (UTC)

Tideman's Alternative Smith and Schwartz

I have added Tideman's Alternative Smith and Schwartz, described in Collective Decisions and Voting: The Potential for Public Choice. I am uncertain if I have these 100% correct; please verify. John Moser (talk) 03:07, 26 July 2018 (UTC)

Feasibility of Count

Should the feasibility of conducting a vote count in a timely fashion be considered a criteria? In an election with a high number of candidates, some systems are clearly much more feasible than others. For example, if there were 100 candidates for a single position, it would be most impractical to use the Ranked Pairs method if the count had to be conducted manually. 1.126.108.98 (talk) 01:44, 6 October 2019 (UTC)

What you call "Feasibility of Count" is addressed by the computational complexity theory. See the column "Polynomial time". Markus Schulze 05:36, 6 October 2019 (UTC)

Ranked Pairs is not less feasible than any other Condorcet method, in that the pairwise votes must be counted in any case. With 100 candidates for a single position, the same concerns arise for instant runoff voting, possibly to a greater degree: pairwise computations can be made at the precincts and summed; IRV must gather all ballots in one location and then count them, and the count must be restarted if it is discovered new ballots arrive, or else the count must be delayed until every single known ballot is collected (which sometimes takes weeks for large elections across broad geographical areas). If ranked pairs is considered impractical due to candidate count, then the only practical methods are approval and plurality, or MNTV and SNTV. Many-candidate elections for few seats are impractical for voters, though, and cannot produce results representing the voters's choice; they need to be reduced by something like single transferable vote (which has the same issue as IRV regarding centralized counting) and only by a limited distance (based on real-world data, the limit is selecting no fewer than around a fifth the number of candidates, which tops a Condorcet election out at around 9 candidates, whereas plurality breaks at…three). John Moser (talk) 00:08, 19 May 2021 (UTC)

Wright System

You need to distinguish between single member and multiple member elections.

Most of the existing systems in use were designed to facilitate a manual counting process. They are outdated. In the process they introduce distortions in the way the vote is counted and the election result. They do not reflect the voters intentions and are an approximation of proportionality. A reiterative count ensures that each vote is treated equally.

https://en.wikipedia.org/wiki/Wright_system — Preceding unsigned comment added by 1.46.45.243 (talk) 09:53, 22 June 2020 (UTC)

Geller-IRV and single-winner QBS

Dear John Moser, at each stage Geller-IRV, checks whether the candidate with the lowest Borda score can be eliminated without violating the majority criterion for solid coalitions. If this is the case, then this candidate is eliminated (without re-calculating the Borda scores). On the other side, single-winner QBS chooses the candidate with the highest Borda score who can be elected without violating the majority criterion for solid coalitions. Geller-IRV and single-winner QBS seem to be identical. Can you give a concrete example, where Geller-IRV and single-winner QBS choose different winners? Markus Schulze 05:43, 19 May 2021 (UTC)

This is a good question. Those do seem to be identical, now that I think about it, because you would remove all candidates with the lowest Borda scores until your mutual majority appears, which has the highest Borda score. They're not identical for multiple-winner elections, but I hadn't considered this approach for single-winner (and understanding QBS is much more difficult than understanding Geller's system, at least for me). I hadn't considered this and have been focused on Geller, single winner or otherwise. I think this comes down to which is more useful to reference, or notating that both are the same (as with Participation/Consistency). John Moser (talk) 13:06, 1 June 2021 (UTC)