[EM] IRV splits majorities to empower everyone?

[Dave Ketchum]

Below I concentrate on understanding vote counting.

On Fri, 18 Nov 2005 18:02:29 -0800 Scott Ritchie wrote:
Subject: [EM] Fwd:  IRV splits majorities to empower everyone,	& find common	ground

> On Thu, 2005-11-17 at 21:54 +0000, IRV Alliance wrote:
> 
>>   The thing you have to understand about IRV is that it is not easy 
>>to exclude voters unless they want to be excluded.  You get 3 chances 
>>to pick a winner. (in SF)  If you cannot find any candidate that the 
>>other voters can agree on, you are probably too far out there to 
>>deserve a strong voice in public policy.  
>>
>>  If you care to be honest about how real-life politics works, it is 
>>not possible to have a dominant majority using IRV.  If one 
>>group/issue has more support than the others, let's say 40-60%+ 
>>support, then realistically  more than one candidate will focus on 
>>that majority group and split the 1st choice vote.   This is the 
>>effect of spoiler-proof multi-candidate elections: No one can count 
>>on a safe majority.  It raises the bar by allowing 2nd choices/more 
>>voters to be counted.

 >>
 >>  Out of all the candidates that split the 1st-choice majority using
 >>IRV, the one who reaches out to the most minority voters for 2nd/3rd
 >>choices will win.
 >>
Not quite true.  Can happen that multiple candidates split the majority 
vote and each get some help from 2nd and 3rd choice votes, but they have 
to get and stay on the list of possible winners by the total vote count - 
neither some first choice votes nor some 2nd choice votes tell the whole 
story.  BUT, candidates from a strong dominant majority CAN maintain control.

> 
> This isn't true.  Someone could be ranked second on literally every
> ballot and be the first loser.  Imagine a committee using IRV to elect
> its chair, where everyone ranks themselves first and then the ideal
> chair (who didn't vote for himself) second.  That's why the difference
> between Condorcet methods and IRV is so important.
> 

In fact, ranking second on EVERY ballot guarantees loss - see below. 
However, Scott does not really help, so I will go into more detail.

I ignore ties below - process such with your choice of method, such as 
lot, and continue as that decides.

Going into a bit more detail with my understanding of IRV:

Stack the original ballots into piles according to first choice on each.

1.  If biggest pile is a majority of ballots still being counted, that is 
winner.  Else, all piles are possible winners except smallest pile (and 
candidates not having a pile) have LOST.

2.  Note that if second choice for every ballot in smallest pile agreed as 
to next choice, and that next choice did not already have a pile, giving 
it a pile would leave it as smallest pile - so whole effect of discussing 
this step is to understand it accomplishes nothing.

3.  Discard remaining choices for each ballot in smallest pile until 
finding a choice that still has a pile, and place that ballot on indicated 
pile.  Discard any ballots with no pile to go to (they are no longer part 
of total for deciding majority).  Back to step 1.

> 
...

> 
>>  Also, you could check the "draft" write-up on this page
>>http://ElectionMethods.com and offer suggestions if you want. I think 
>>everyone might as well admit that Ranked Pairs(IRV+) is a little more 
>>logical than IRV, but IRV will do for now to address the real 
>>problems of minority disempowerment using 1-choice ballots.
>>

Agreed that Condorcet is better due to IRV's weakness, but I have not 
emphasized that topic above (they usually agree as to winner).


>> Thanks.

-- 
  davek at clarityconnect.com    people.clarityconnect.com/webpages3/davek
  Dave Ketchum   108 Halstead Ave, Owego, NY  13827-1708   607-687-5026
            Do to no one what you would not want done to you.
                  If you want peace, work for justice.