Re: [EM] Definitions of "completion method" and "aggregate method"
I thought a "completion method" is what we revert to if there is no Condorcet winner to elect. Methods like Schulze and Ranked Pairs or Bottom Two Runoff don't need completion methods because the are decisive, they elect someone whether there's a Condorcet winner or not, but they always elect the Condorcet winner when such exists.Powered by Cricket Wireless------ Original message------From: Rob LanphierDate: Wed, May 4, 2022 3:16 AMTo: election-methods@lists.electorama.com;Cc: Subject:[EM] Definitions of "completion method" and "aggregate method"Hi folks,
There are two new articles on electowiki that may not make sense to
those of you who are are intimately familiar with election-methods
jargon:
"aggregate method":
https://electowiki.org/wiki/Aggregate_method
"completion method"
https://electowiki.org/wiki/Completion_method
An "aggregate method" is jargon I made up to describe the methods that
we've been describing all along on this mailing list (e.g. the
"Smith//Condorcet" method was popular on the EM list1).
What I mean by "aggregate method" is a method where a series of
methods (Copeland, Smith, etc) are chained together. Then the "winner"
(or "winners") pop out at the end of the chain.
What I mean by "completion method" is the method at the end of the
chain in an "aggregate method".
Am I using those terms the way that people who make a living
researching/studying this stuff use those terms? I don't want to
redefine terms, but if no one agrees on what "completion method" means
(or has a really sloppy definition of it, or just uses the term
without knowing what it means to sound smart), then I don't feel bad
about redefining the term. Regardless, I would really appreciate some
help coming up with good definitions.
Thanks
Rob
Election-Methods mailing list - see https://electorama.com/em for list info
Hi r b-j!
Thanks for the quick reply! Is a "completion method" specific to
aggregate methods that use the ballots in an election to select a
Condorcet set at the beginning of the process (e.g. the Smith set),
and then uses another method when more than one candidate (or at
least, less than the desired number of candidates) is selected? Can
the term be used when NOT speaking about Condorcet-winner compliant
methods?
Rob
On Wed, May 4, 2022 at 12:30 AM robert bristow-johnson
rbj@audioimagination.com wrote:
I thought a "completion method" is what we revert to if there is no Condorcet winner to elect. Methods like Schulze and Ranked Pairs or Bottom Two Runoff don't need completion methods because the are decisive, they elect someone whether there's a Condorcet winner or not, but they always elect the Condorcet winner when such exists.
Powered by Cricket Wireless
------ Original message------
From: Rob Lanphier
Date: Wed, May 4, 2022 3:16 AM
To: election-methods@lists.electorama.com;
Cc:
Subject:[EM] De finitions of "completion method" and "aggregate method"
Hi folks,
There are two new articles on electowiki that may not make sense to
those of you who are are intimately familiar with election-methods
jargon:
"aggregate method":
https://electowiki.org/wiki/Aggregate_method
"completion method"
https://electowiki.org/wiki/Completion_method
An "aggregate method" is jargon I made up to describe the methods that
we've been describing all along on this mailing list (e.g. the
"Smith//Condorcet" method was popular on the EM list1).
What I mean by "aggregate method" is a method where a series of
methods (Copeland, Smith, etc) are chained together. Then the "winner"
(or "winners") pop out at the end of the chain.
What I mean by "completion method" is the method at the end of the
chain in an "aggregate method".
Am I using those terms the way that people who make a living
researching/studying this stuff use those terms? I don't want to
redefine terms, but if no one agrees on what "completion method" means
(or has a really sloppy definition of it, or just uses the term
without knowing what it means to sound smart), then I don't feel bad
about redefining the term. Regardless, I would really appreciate some
help coming up with good definitions.
Thanks
Rob
Election-Methods mailing list - see https://electorama.com/em for list info
I don't know of any methods besides Condorcet that result in Condorcet-style ties requiring a completion method.
--pH
-----Original Message-----
From: Election-Methods election-methods-bounces@lists.electorama.com On Behalf Of Rob Lanphier
Sent: Wednesday, May 4, 2022 2:58 AM
To: robert bristow-johnson rbj@audioimagination.com
Cc: election-methods@lists.electorama.com
Subject: Re: [EM] Definitions of "completion method" and "aggregate method"
Hi r b-j!
Thanks for the quick reply! Is a "completion method" specific to aggregate methods that use the ballots in an election to select a Condorcet set at the beginning of the process (e.g. the Smith set), and then uses another method when more than one candidate (or at least, less than the desired number of candidates) is selected? Can the term be used when NOT speaking about Condorcet-winner compliant methods?
Rob
On Wed, May 4, 2022 at 12:30 AM robert bristow-johnson rbj@audioimagination.com wrote:
I thought a "completion method" is what we revert to if there is no Condorcet winner to elect. Methods like Schulze and Ranked Pairs or Bottom Two Runoff don't need completion methods because the are decisive, they elect someone whether there's a Condorcet winner or not, but they always elect the Condorcet winner when such exists.
Powered by Cricket Wireless
------ Original message------
From: Rob Lanphier
Date: Wed, May 4, 2022 3:16 AM
To: election-methods@lists.electorama.com;
Cc:
Subject:[EM] De finitions of "completion method" and "aggregate method"
Hi folks,
There are two new articles on electowiki that may not make sense to
those of you who are are intimately familiar with election-methods
jargon:
"aggregate method":
https://electowiki.org/wiki/Aggregate_method
"completion method"
https://electowiki.org/wiki/Completion_method
An "aggregate method" is jargon I made up to describe the methods that
we've been describing all along on this mailing list (e.g. the
"Smith//Condorcet" method was popular on the EM list1).
1996-August/065995.html
What I mean by "aggregate method" is a method where a series of
methods (Copeland, Smith, etc) are chained together. Then the "winner"
(or "winners") pop out at the end of the chain.
What I mean by "completion method" is the method at the end of the
chain in an "aggregate method".
Am I using those terms the way that people who make a living
researching/studying this stuff use those terms? I don't want to
redefine terms, but if no one agrees on what "completion method" means
(or has a really sloppy definition of it, or just uses the term
without knowing what it means to sound smart), then I don't feel bad
about redefining the term. Regardless, I would really appreciate some
help coming up with good definitions.
Thanks
Rob
Election-Methods mailing list - see https://electorama.com/em for list
info
Election-Methods mailing list - see https://electorama.com/em for list info
I don't think the terminology is settled, but it makes sense to use the
term 'completion' for methods of the form "if M has a unique winner,
elect that candidate, otherwise elect the winner of Mdash", whereas a
tiebreak is limited in some way to the subset of candidates who tie
under M. Condorcet cycles are not normally called ties, so Condorcet
resolutions are called completions. It would be possible to have Llull
completions, eg. elect the Llull winner if it is unique, else elect the
Borda winner. Certainly Llull's method doesn't require a completion
method in this sense, but it's an alternative to the two natural ways of
breaking ties, and someone will one day prove that it has unique
robustness to a certain form of tactical voting.
CJC
On 04/05/2022 15:00, Hahn, Paul wrote:
I don't know of any methods besides Condorcet that result in Condorcet-style ties requiring a completion method.
--pH