Alright, burial might be prohibitive problem for wv & MMPO.
Chris was right about bottom-end burial being a likely big problem for wv &
MMPO.
In wv & MMPO, if people rate candidates as top-set & bottom-set, then,
contrary to what I said, they won't be deterred from burial among their
bottom-set. ...because they don't care which of them wins, if they have
at least some tiny chance of defeating them all...as they could if the
CWs's voters ever forget to plump, or misjudge who's CWs. (If the CWs is in
their bottom-set).
If so, then maybe burial could be so rampant that the CWs would rarely win.
..a mess.
Bucklin & Approval (including 3-Slot ICT as an Approval-version) have a
simplicity, solidity & reliability that pairwise-count unlimited-rank
methods don't have.
Michael Ossipoff
There are conflicting incentives, at least under the Schwartz methods. Contributing to the defeat strength of some candidate A>B is bad for B, but it can be good for A, because it can give A a stronger path against "top-set" candidates, and make them lose. There's a transitivity to defeats.
If you have an estimate of who is the worse of two frontrunners (and there are just two), you shouldn't rank that candidate over anybody. This is true even if you feel indifferent among your bottom-set.
Kevin
De : Michael Ossipoff <email9648742@gmail.com>
À : election-methods@electorama.com; C. Benham cbenham@adam.com.au
Envoyé le : Vendredi 30 septembre 2016 20h40
Objet : [EM] Alright, burial might be prohibitive problem for wv & MMPO.
Chris was right about bottom-end burial being a likely big problem for wv & MMPO.In wv & MMPO, if people rate candidates as top-set & bottom-set, then, contrary to what I said, they won't be deterred from burial among their bottom-set. ...because they don't care which of them wins, if they have at least some tiny chance of defeating them all...as they could if the CWs's voters ever forget to plump, or misjudge who's CWs. (If the CWs is in their bottom-set).If so, then maybe burial could be so rampant that the CWs would rarely win...a mess.Bucklin & Approval (including 3-Slot ICT as an Approval-version) have a simplicity, solidity & reliability that pairwise-count unlimited-rank methods don't have.Michael Ossipoff
Election-Methods mailing list - see http://electorama.com/em for list info
Because I like some properties of wv and MMPO, I could try to defend them
by saying that maybe the problem would be so well known that the
CWse-preferrers would never forget to plump, & so the burial would never
work, & so people wouldn't bother.
But no one can really be sure who will be a CWs. Even a little uncertainty
means that there will be mis-estimates, & that burial will sometimes be
rewarded.
...changing the win from buriers' bottom-set to their top-set, giving them
incentive for burial among their bottom-set.
So wv & MMPO remain doubtful proposals.
It could be argued that a CWs tends to be someone popular, & unlikely to be
in many people's bottom-set.
But we've always had to consider the possibility of a low-favoritness CWs.
And the CWs in the Burlington debacle was least favorite of the top 3.
So maybe a big faction could consider the CWs to be bottom-set.
Of course honest elections & honest media would be a completely different
world, & none of us have any idea what it would be like.
Maybe the current ridiculous disparity of candidate-merit, & the existence
of a bottom-set, is an artifact of our phony elections & media.
Maybe, but, for current conditions, maybe method-choice still has to assume
the unlikely worst.
Michael Ossipoff
On Sep 30, 2016 6:40 PM, "Michael Ossipoff" email9648742@gmail.com wrote:
Chris was right about bottom-end burial being a likely big problem for wv
& MMPO.
In wv & MMPO, if people rate candidates as top-set & bottom-set, then,
contrary to what I said, they won't be deterred from burial among their
bottom-set. ...because they don't care which of them wins, if they have
at least some tiny chance of defeating them all...as they could if the
CWs's voters ever forget to plump, or misjudge who's CWs. (If the CWs is in
their bottom-set).
If so, then maybe burial could be so rampant that the CWs would rarely win.
..a mess.
Bucklin & Approval (including 3-Slot ICT as an Approval-version) have a
simplicity, solidity & reliability that pairwise-count unlimited-rank
methods don't have.
Michael Ossipoff
You-lot were right about MMPO having a prohibitive problem.
Approvalists & IRVists were right about wv having too much of a burial
problem.
...& Chris was right all along about how problematic bottom-end burial is.
Michael Ossipoff
On Oct 1, 2016 8:08 AM, "Kevin Venzke" stepjak@yahoo.fr wrote:
There are conflicting incentives, at least under the Schwartz methods.
Contributing to the defeat strength of some candidate A>B is bad for B, but
it can be good for A, because it can give A a stronger path against
"top-set" candidates, and make them lose. There's a transitivity to
defeats.
If you have an estimate of who is the worse of two frontrunners (and there
are just two), you shouldn't rank that candidate over anybody. This is true
even if you feel indifferent among your bottom-set.
Kevin
De : Michael Ossipoff email9648742@gmail.com
À : election-methods@electorama.com; C. Benham cbenham@adam.com.au
Envoyé le : Vendredi 30 septembre 2016 20h40
Objet : [EM] Alright, burial might be prohibitive problem for wv & MMPO.
Chris was right about bottom-end burial being a likely big problem for wv
& MMPO.
In wv & MMPO, if people rate candidates as top-set & bottom-set, then,
contrary to what I said, they won't be deterred from burial among their
bottom-set. ...because they don't care which of them wins, if they have
at least some tiny chance of defeating them all...as they could if the
CWs's voters ever forget to plump, or misjudge who's CWs. (If the CWs is in
their bottom-set).
If so, then maybe burial could be so rampant that the CWs would rarely win.
..a mess.
Bucklin & Approval (including 3-Slot ICT as an Approval-version) have a
simplicity, solidity & reliability that pairwise-count unlimited-rank
methods don't have.
Michael Ossipoff
Election-Methods mailing list - see http://electorama.com/em for list info
Of course, the fancier a method is, & the more luxury advantages it offers,
then the more potential problems can result from its more complicated
mechanism.
So it isn't so surprising that MAM & MMPO (in both it's plain & Smith
versions) make you pay for their unique strategy advantages...
...with a potentially extreme problem, in the form of the perpetual-burial
fiasco problem that started this thread.
Of course I'm not saying for sure that will happen, but it led me to call
MAM & MMPO "doubtful" proposals.
But let me mention some mitigation:
- If that fiasco began, then maybe any voter whose favorite has any
likelihood of being CWs would use the plumping countermeasure. ...maybe to
the extent that the burial would reliably never gain anything.
If so, then maybe people wouldn't have any reason to take the trouble to
bother with the burial.
- The fiasco scenario depends on the CWs being in the buriers' bottom-set.
That's why they're indifferent to the punitive consequences of their burial
when countermeasures are used. They don't care which bottom-set candidate
wins.
But any remote possibility of the burial defeating the CWs & electing from
the buriers' top-set, depends on the CWs's voters liking the buriers'
preferred candidate(s) enough to be inclined to rank them well.
But aren't those 2 requirements a bit mutually contradictory?
Maybe that upgrades MAM & MMPO from "doubtful" to "maybe a bit
questionable".
Maybe they're worth a try, though there'd need to be some disclosure of the
potential perpetual burial fiasco.
Michael Ossipoff
On Oct 1, 2016 8:08 AM, "Kevin Venzke" stepjak@yahoo.fr wrote:
There are conflicting incentives, at least under the Schwartz methods.
Contributing to the defeat strength of some candidate A>B is bad for B, but
it can be good for A, because it can give A a stronger path against
"top-set" candidates, and make them lose. There's a transitivity to
defeats.
If you have an estimate of who is the worse of two frontrunners (and there
are just two), you shouldn't rank that candidate over anybody. This is true
even if you feel indifferent among your bottom-set.
Kevin
De : Michael Ossipoff email9648742@gmail.com
À : election-methods@electorama.com; C. Benham cbenham@adam.com.au
Envoyé le : Vendredi 30 septembre 2016 20h40
Objet : [EM] Alright, burial might be prohibitive problem for wv & MMPO.
Chris was right about bottom-end burial being a likely big problem for wv
& MMPO.
In wv & MMPO, if people rate candidates as top-set & bottom-set, then,
contrary to what I said, they won't be deterred from burial among their
bottom-set. ...because they don't care which of them wins, if they have
at least some tiny chance of defeating them all...as they could if the
CWs's voters ever forget to plump, or misjudge who's CWs. (If the CWs is in
their bottom-set).
If so, then maybe burial could be so rampant that the CWs would rarely win.
..a mess.
Bucklin & Approval (including 3-Slot ICT as an Approval-version) have a
simplicity, solidity & reliability that pairwise-count unlimited-rank
methods don't have.
Michael Ossipoff
Election-Methods mailing list - see http://electorama.com/em for list info