[EM] Condorcet, please

[Dave Ketchum]

Looking at Condorcet:
     Bullet voting, ala Plurality, is permitted.
     Approval, also permitted.
     Ranking, approving but showing difference in liking, permitted.
     Rating, being able to indicate strength of difference in liking,  
not permitted or required - meaning no need or ability to compensate  
for what other voters might do with such.
     Summary:  While the Condorcet voter can approve of more than one,  
there is no requirement for indicating difference of preference among  
such.

Summary:
      Condorcet is worth backing.
      Range is competitive - ranking and rating each have advantages,  
though I prefer ranking:
            Ranking lets voters indicate which candidates they prefer  
- candidates preferred by more voters win.
            Rating lets voters indicate how much they prefer A over B  
- means a minority of voters agreeing where to apply emphasis can win  
- truly great power - and GREAT duties to get it right.
      IRV, Plurality, Approval, Bucklin, and others have  
disadvantages, though I will say little on this this time.

If a majority of Plurality voters vote for A, A wins.  IRV and Bucklin  
are normally restricted to one top rank candidate - similar voting  
assures the same winner, it not mattering what is done by other voters  
or in other ranks.  If a majority of Condorcet voters choose to obey  
the same restriction as to A and top rank, they will also elect A.

Looking at Condorcet ranking:
     higher>A&B>lower>unranked:
           Each higher counts as above each of A & B.
           A & B count as above each lower or unranked.
           For A>B, A counts as above B.
           For A=B, neither counts as above the other.

Bullet voting?  NOTHING wrong with this for many voters much of the  
time - if they see a preferred candidate and see no other worth even a  
lower rank.  Condorcet offers other choices that are USEFUL when the  
selection of candidates and issues is seen as more complex.

Approval cutoff?  Why bother?  If a candidate would rank below this  
level why not leave unranked - what useful is there to do with ranking  
the candidate unless it means a better chance at winning?

Write-ins - should be permitted.  Usually not worth bothering, but  
there are exceptions, such as someone becoming much more or less  
electable after nomination period - one I remember was in a town near  
here when ZERO petitions got submitted by the deadline.  Turned out  
someone's error got believed - solution was to vote by write-ins only.

Primaries?  Plurality desperately needs them, for multiple candidates  
for a party split the available votes.  Condorcet and many other  
methods allow voters to vote for multiple candidates in such cases.   
Need to balance what remaining good primaries offer against their  
expense in money and time.

Runoffs?  Plurality desperately needs them any time the leading  
candidate fails to get a majority of the votes, for the voters have  
not had a chance to fully express their desires - and could agree on  
other than that leader.  Condorcet and many other methods allow voters  
to more completely express their desires.  Need to consider when the  
expectable good from runoffs justifies their expense in money and time  
- getting near to a majority could be the most it is worth investing  
effort on with these methods.

Each of the above needs careful thought, for doing without them means  
different strategy becomes appropriate:
      Primaries mean that the most generally acceptable candidate does  
not get to the general election if rejected by primary voters in own  
party.
      Making the general election become usually the final step can  
encourage more complete preparation for that step.
      Both affect what schedules are practical for  election timing.

How many explicit ranks to support, besides unranked?  Three seems  
common in IRV and Bucklin. Condorcet cares not.  More can mean  
complexity in voting - and pleasing some voters - have fun!

Why have rules about how voters assign ranks, beyond the permission to  
assign any rank number to as many or as few candidates as the voter  
chooses - the counters properly only care about comparing two  
candidates at any instant - every possible pair for which a voter  
ranked one or both members - omitting only pairs of unranked members  
(for this last, both members have the same rank).

Multi-winning to elect multiple members from a district?  This is  
considered for various reasons.  Here I only suggest design/ 
consideration here of a ranking method similar to Condorcet if  
Condorcet is used for other elections.

I read of election systems getting positive credit for producing  
results similar to other systems.  Need thought also to noticing  
differences and paying more attention to whether these represent  
improvements.

On Jan 16, 2010, at 3:28 AM, Dave Ketchum wrote
Re: [EM] IRV vs Plurality (Dave Ketchum):

> On Jan 15, 2010, at 10:51 PM, Kathy Dopp wrote:
>>
>>> Date: Fri, 15 Jan 2010 22:05:58 -0500
>>> From: Dave Ketchum <davek at clarityconnect.com>
>>> To: Juho <juho4880 at yahoo.co.uk>
>>>
>>> On Jan 15, 2010, at 7:46 PM, Juho wrote:
>>>> On Jan 14, 2010, at 2:13 AM, Dave Ketchum wrote:
>>>>>>>
>>>>> When there is a CW in Condorcet, the CW has won in comparison with
>>>>> each other candidate.  While a few may like X or Z enough better  
>>>>> to
>>>>> have given such top ranking, the fact that all the voters together
>>>>> prefer the CW over each other should count, and does with  
>>>>> Condorcet.
>>>>>
>>>>> Else there is a cycle in Condorcet.  Perhaps the following Minimum
>>>>> Margins Method Condorcet variant should be used to establish
>>>>> Condorcet's preferability over other methods.  Then let other
>>>>> variants compete with this one before finally deciding which to  
>>>>> use.
>>>>>
>>>>> Minimum Margins Method:  Consider the cycle, such as A>B>C>A, and
>>>>> the margins that create it, such as 60A>30B, 40B>20C, 21C>20A.
>>>>> Delete the weakest margins as many times as needed to destroy the
>>>>> cycle - in this case A becoming the CW (note that if one C>A voter
>>>>> had voted A>C in  this election, A would have become CW with no
>>>>> cycle).
>>
>> Great idea. Is this Dave Ketchum speaking above?  Very simple and
>> logically coherent plan.  Thanks for sharing.  Could it be possible
>> that this plan would ever not work? (I.e. same margins?)
>
> I did this, though suspecting the idea already has a variant name.   
> The adoption is intentionally two steps:
>     1.  Use this variant to easily prove there is a Condorcet  
> variant ready to compete against such as IRV.
>     2.  If there is a better variant, even though likely more  
> complex, let it compete against this one.
>
> Same margins is a possibility requiring a response be attended to  
> before actual use.  At proposal time the possibility needs  
> mentioning - I see nothing more needed at that time (probably delete  
> them one at a time in some specified order).
>>>>
>>> Note that this method breaks the cycle at the point where the  
>>> smallest
>>> number of ballots being voted differently would have broken the  
>>> cycle.
>>>
>>> Note that weaker candidates are unlikely to get enough votes to be
>>> part of a cycle - being weak they get few high rank votes.

Dave Ketchum
>
>>
>> Yes, one could certainly say that this allows the top cycle to  
>> prevail
>> by breaking the weakest link where weakest is defined as the smallest
>> margin in this case.  This minimum margins method is so logically
>> correct and fair.
>>>>>>
>> Condorcet is SO much easier and quicker and simpler to count than IRV
>> if one simply tallies one n x n matrix for each precinct and sums the
>> corresponding positions for all the precincts.  No need to wait for
>> all the absentee and provisional ballots, no need for centralized
>> counting, no need to sort and resort ballots into dozens of piles of
>> ballots, or even worse with STV keep track of which portion of which
>> ballot goes into which pile (tearing or cutting up the ballots would
>> help in that manual counting nightmare.)
>>
>> Truly, I cannot imagine a more insane method of counting rank choice
>> ballots than IRV/STV when one begins to consider the practicalities  
>> of
>> election administration.
>>
>> That minimum margin idea for resolving Condorcet cycles is neat, if
>> one wants to use rank choice ballots.
>> -- 
>> Kathy Dopp