[EM] STV and weighted positional methods

[Jonathan Lundell]

On Feb 1, 2009, at 12:26 AM, Kathy Dopp wrote:

> On Sun, Feb 1, 2009 at 12:06 AM, Jonathan Lundell  
> <jlundell at pobox.com> wrote:
>
>> I know what you mean. I tried to automate STV with two tin cans and a
>> string, and got nowhere at all. STV sucks.
>>
>
> Jonathan,
>
> Don't know what you're trying to say.  If you mean that a spreadsheet
> won't work to automatically count STV, then that just shows how
> nontransparent the STV counting process is that an ordinary citizen
> who doesn't do computer programming and have programming tools and
> training cannot with out huge effort and time check the accuracy of
> any STV election by simply checking the tally during the canvass
> period, even if all the individual ballot choices of every voter are
> publicly published.
>
> Other methods, such as Condorcet and certainly the easy range or
> approval are trivially easy to count with a spreadsheet.

If technology-avoidance is your goal, to the point of counting by  
"ordinary citizens", I don't know why you want to bring spreadsheets  
into the picture. There are several forms of STV in use that are  
amenable to hand counting. However, the actual counting algorithm for  
any STV method (with a minor exception for a couple of proposed  
composite methods that are not in actual use) is quite  
straightforward, and it's not difficult to implement the count in a  
scripting language that ends up being considerably more readable than  
a complex spreadsheet.

What's more important, it seems to me, is that the counting software  
be open to inspection, and/or that the ballots be available for  
independent counting. There are multiple open-source counters  
available that could be used for either purpose (primary counting or  
verification).

I have seen spreadsheet-based STV counts, but spreadsheets simply  
don't express iterative algorithms very well. (For that matter, I  
wonder whether a complete Condorcet count by spreadsheet can fairly be  
called "trivial", depending on the method employed to deal with cycles).

This is all something of a red herring, though, isn't it?

>
>
> BTW, I am against using any method where voters can only cast one vote
> for filling two or more at-large seats because this takes away votes
> from the voter - especially when using a single STV vote method to
> fill two or more at-large seats where sometimes your second choice
> will never be counted, even though your first choice ends up losing
> and even though you would have, under the plurality method, been
> allowed to cast two votes to fill the two at-large seats as it should
> be.

By that standard, at-large plurality disregards all votes save those  
cast for the winners, since everyone else is eliminated and their  
voters are left without representation.

PR, including PR/STV, seeks to produce a body that is as fairly  
representative of the electorate as possible, a goal at which  
plurality at-large elections fail miserably.

>
> The more I've learned about STV and IRV, the more amazed I am that
> anyone would consider using such an unfair method in any election,
> especially to cast one vote for a multi-seat contest where your second
> choice may never even be counted.

I'm more concerned that my vote actually contribute to the election of  
a candidate. STV significantly improves the chances of that. If you  
can't recognize the core flaw of plurality-take-all elections, I'm  
disinclined to attach much weight to your STV nit-picking.