[EM] MMPO(IA>MPO) (was IA/MMPO)

[Forest Simmons]

Yes, seeing IBIFA as an improvement on Bucklin is the key,

I don't see any counting shortcut.  We would need the full n by n pairwise
matrix with another n by m matrix of approvals based on the m possible
cutoffs.  For three slots, m would be 2.

Perhaps Max Spread (between threshold approval and opposition) would be a
name. MAOS?

Keep those ideas coming!


On Sat, Oct 26, 2013 at 1:29 PM, Jameson Quinn <jameson.quinn at gmail.com>wrote:

> I like the relative simplicity of this method. It's also nice because it
> has a clear relationship to Bucklin/median methods. It seems as though it
> would do well against the chicken dilemma; though perhaps it would be even
> better if it used pure MPO as a tie breaker.
>
> I have two questions, then. What would you call it? And, how would you
> count it in practice? Would you need to keep a full Condorcet matrix, or
> are there tricks which would allow you to do a hand-count without that?
>
> As to naming, here are some ideas:
>
>    - Pairwise Opposition Bucklin
>    - Instant Runoff Bucklin
>    - Threshold Approval Minus Oppsition (TAMO)
>
>
>
> 2013/10/26 Forest Simmons <fsimmons at pcc.edu>
>
>> Here's an idea that is essentially a more seamless version of Chris
>> Benham's IBIFA:
>>
>> Each voter rates each candidate on a scale of zero to some maximum
>> allowable rating..
>>
>> Set the approval cutoff level at the highest level possible consistent
>> with at least one candidate having at least as much approval as max
>> pairwise opposition.
>>
>> Elect the candidate with the biggest difference between approval and max
>> pairwise opposition.
>>
>> That's it.
>>
>> It turns out that if no other approval cutoff yields a candidate with as
>> much approval as MPO, then setting the cutoff between zero and the next
>> higher level (thereby yielding "Implicit Approval") will work.  In fact
>> when approval is given by Implicit Approval, the candidate with the most
>> approval is one candidate (but not necessarily the only one) that will have
>> at least as much approval as MPO.
>>
>> This method satisfies the FBC and the mono-add-top version of
>> Participation, etc.  It also solves the same problem of ordinary Approval
>> and MAMPO that IBIFA was designed to solve, viz getting your Compromise out
>> of the way when your Favorite has sufficient high level support.
>>
>> What do you think?
>>
>> On Fri, Oct 25, 2013 at 7:10 PM, Chris Benham <cbenhamau at yahoo.com.au>wrote:
>>
>>> Forest,
>>>
>>> Another FBC-complying method (that I'd forgotten about) that I also like
>>> a lot is
>>> "ICT"  (as Mike O. renamed it).  It should only be used on 3-slot
>>> ratings ballots.
>>>
>>> It is just like Kevin V's  ICA  (Improved Condorcet//Approval) except if
>>> there is
>>> no "tied-at-the-top" rule pairwise beats-all candidate then instead of
>>> electing the
>>> most approved (rated Top or Middle) candidate we elect the most
>>> Top-rated candidate.
>>>
>>> http://nodesiege.tripod.com/elections/#methica
>>>
>>> Unlike  ICA, IBIFA, Shulze and other methods, it meets Mike O's
>>> "Chicken dilemma
>>> criterion".
>>>
>>> Chris
>>>
>>>
>>>
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.electorama.com/pipermail/election-methods-electorama.com/attachments/20131026/acb98b93/attachment-0004.htm>