Re: Understanding where we are with TAG/AB voting

Brian,

I think Chaals summarizes it well.

I'm not sure where you are coming from with your question - what you 
know and what you don't know.  A good place to start is [1] a 
description of the STV voting that we use.  To count the votes we use 
the Meek method [2].

Jeff

[1] https://en.wikipedia.org/wiki/Single_transferable_vote

[2] https://en.wikipedia.org/wiki/Counting_single_transferable_votes#Meek


On 12/1/2017 6:29 PM, Chaals McCathie Nevile wrote:
> TL;DR: Rank all candidates for the best value in your vote. If you 
> prefer two or more equally, you might as well use a random method to 
> choose between them...
>
> On Fri, 01 Dec 2017 22:25:49 +0100, Brian Kardell <bkardell@gmail.com> 
> wrote:
>
>> As someone interested in keeping the TAG and AB doing what I believe 
>> are good things, I have to >admit that despite the flurry of 
>> conversations and concerns and even a few TPAC conversations, I'm 
>> >still not entirely clear where this leaves things right now.  So, 
>> can someone help me clear things >up?
>>
>> There are 4 nominees for 2 seats.  Let's say, hypothetically, that my 
>> ideal outcome is that two of >them get reelected.  On a scale of 1 to 
>> 10, I'm putting them both at 10.   I very much desire >this.  But 
>> there are two other people too and, in the old system, I couldn't 
>> express anything about >that, despite the fact that I do, actually, 
>> have a strong preference of which I'd like should one >of my ideals 
>> not work out.
>
> This is pretty simple.
>
> Sadly we don't yet have "equal ranking" implemented, so you have to 
> choose which of your 2 desired candidates you put as "1", and which as 
> "2". Hopefully that will change in the near future, but not for this 
> election.
>
> You should continue, ranking the others 3 and 4 according to which you 
> prefer. If your preferred 2 candidates win, the rest is irrelevant, of 
> course. But if they don't, then the preferences you expressed as 3 and 
> 4 will be counted in deciding which of the other candidates gets elected.
>
> In more detail than you probably want...
>
> *How much* your further preferences count depends: If your successful 
> top-2 candidate won with barely enough votes then most of your vote 
> went to them, and the "leftover fraction" to determine the other place 
> is small. If the candidate got many more votes than needed, a large 
> fraction of your vote will be counted in working out who of the others 
> gets in.
>
> In all cases your preferences are counted along with everyone else's 
> votes, of course.
>
> So rank all candidates in the order you prefer - and where you prefer 
> two equally you might as well flip a coin or otherwise randomly choose 
> which goes first and which goes second.
>
> Where you don't have a preference between 2 candidates, it doesn't 
> matter much which one you rank first. Effectively you can only make it 
> more probable that *both* will be elected except by others to vote for 
> them, and however you rank them makes no difference to the chance that 
> at least one of them will be elected.
>
> If you care about an outcome that is not "your preferred candidates 
> all winning", you should rank your preferences. Unless you get 
> everything you want, those votes *will* be counted and influence the 
> end result.
>
> Note also: it is truly said that tossing a coin high in the air is a 
> good way to decide between two things you value equally. By the time 
> it lands, you may have realised that you *do* care which way it comes 
> out - and if not you have applied a sufficiently good decision-making 
> algorithm.
>
> cheers
>
> chaals
>

Received on Saturday, 2 December 2017 23:19:50 UTC