Talk:Ranked Robin: Difference between revisions
Content deleted Content added
→Comparison to Black's method: does any form of the Borda count benefit from the use of pairwise matrices? |
→Relationship to Total Vote Runoff?: new section |
||
| (3 intermediate revisions by 3 users not shown) | |||
Line 21:
:::: [[User:Kristomun|Kristomun]] says: "''Ranked Robin uses ordinal ballots, so it can't be Score''". As I read through the description of [[Ranked Robin]] more closely, I see there's a lot that I don't understand about the proposal. As written, it describes a "ranked ballot" (where "1" is at the top) but it seems possible to use a [[STAR voting]]-style ballot (the way my old [[electowidget]] voting mechanism worked). That said, as a drop-in replacement for [[instant-runoff voting]], this seems fine, and moreover, it seems to me that cycles will be rare in practice, so the [[Copeland's method|Copeland set]] seems likely to only have one winning candidate in 99.9% of the cases. Regardless, since the system uses pairwise matrices (like most Condorcet methods do), it's hard to see how this method has much in common with Borda, since adding candidates to the bottom of the ranking shouldn't make any difference in the pairwise comparison between the candidates above the newly added candidate.-- [[User:RobLa|RobLa]] ([[User talk:RobLa|talk]]) 04:46, 13 January 2022 (UTC)
::::: True, if the 99.9% Condorcet rate is a fundamental fact, then it doesn't much matter what Condorcet completion rule you use. It is possible that it's a side effect of the two and a half party domination that IRV encourages, though, or the two-party domination that using Plurality everywhere else does.
::::: I'd probably say the two main drawbacks of Ranked Robin are that it's more susceptible to strategy (than say minmax), and that it's not actually cloneproof. OTOH, Borda's honest VSE is not all that shabby, so at least you're getting ''something'' in return :-)
::::: As for Condorcet and Borda, there's a definite link between the two. Suppose there are no truncated or equal-rank ballots. Then the sum of A>X over all X is A's Borda score, so you can do a Borda count by using the pairwise matrix. That's what makes methods like Smith//Borda summable. Adding another candidate adds another X where A>X can be added to A's sum. In a more general sense, Borda is the "mean" while Condorcet is the "median": see http://lists.electorama.com/pipermail/election-methods-electorama.com/2011-January/092327.html. This gives a good intuition of why Borda is more strategically susceptible: because the median is robust and the mean isn't. [[User:Kristomun|Kristomun]] ([[User talk:Kristomun|talk]]) 10:26, 13 January 2022 (UTC)
== Clone dependence ==
Line 41 ⟶ 47:
and A wins. [[User:Kristomun|Kristomun]] ([[User talk:Kristomun|talk]]) 12:50, 24 November 2021 (UTC)
== Improvement to Ranked Robin method ==
On the Election-Methods discussion forum I'm suggesting that the Borda count tie breaker be replaced with "pairwise support counts" so that the pairwise matrix does not need to be presented -- to voters or to media reporters or to people who take the time to learn how the counting is done. This improvement will yield results that will be similar to the current version, but with less vulnerability to tactical voting.
The resulting improved Ranked Robin method can be described as follows:
"The [improved] Ranked Robin method elects the candidate who has the most number of one-on-one wins against every other candidate. [or whichever Copeland variation is desired] If more than one candidate has the same largest number, the method elects from those tied candidates the candidate with the highest pairwise support count. The pairwise support count for each of the tied candidates is the sum, across all the ballots, of the number of tied candidates who are ranked lower than the candidate whose pairwise support count is being counted."
The words "tied candidate" are important because the pairwise support count should not involve candidates who have already been eliminated using the Copeland method.
Several other election-method experts have given their responses to this improved method. Their responses can be accessed on these two pages:
http://lists.electorama.com/pipermail/election-methods-electorama.com/2022-March/thread.html (bottom of page)
http://lists.electorama.com/pipermail/election-methods-electorama.com/2022-April/thread.html (top of page)
The full suggestion is at the following location:
http://lists.electorama.com/pipermail/election-methods-electorama.com/2022-March/003709.html
The responses can be accessed using the "Next message (by thread): [EM] Improvement to Ranked Robin method" link.
This improvement to the Ranked Robin method will dramatically simplify the explanation of the method, which is expressed as the goal behind creating this method.
[[User:VoteFair|VoteFair]] ([[User talk:VoteFair|talk]]) 17:44, 5 April 2022 (UTC)
== Code to implement it? ==
Thanks - interesting. Where can we get code to implement this? Python preferred. [[User:Nealmcb|Nealmcb]] ([[User talk:Nealmcb|talk]]) 01:46, 22 October 2022 (UTC)
== Relationship to Total Vote Runoff? ==
Thanks! How does this compare to another new method: Total Vote Runoff: A Majority-Maximizing Form of Ranked Choice Voting by Edward B. Fole - SSRN - https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4328946 [[User:Nealmcb|Nealmcb]] ([[User talk:Nealmcb|talk]]) 18:59, 22 January 2023 (UTC)
| |||