# Re: Fwd: Hand count elections

From: Kaj Telenar <kaj_at_telenar_dot_com>
Date: Tue Nov 27 2007 - 15:24:14 CST

I have been in favor of IRV for quite a while (20 years or so, although
I called it something else back then). The advantage being that the
winner has been backed by over half the votes cast, without the added
expense of running another election (or several additional elections).

I think the cost is lower than Danny's estimate because each time
someone is eliminated, you are only counting the eliminated ballots. You
don't have to recount the ballots for the other candidates. You only
have to add the new ballots to the old total.

Attempting to put some worst-case numbers on here starts as follows:

1. round one is the same as any stack and count race. Each candidate
gets a stack of ballots for that candidate and the stack gets counted.
This should take the same amount of time as a standard race since no one
is eliminated yet.

2. I am assuming that all write-in candidates can be eliminated at this
point and redistributed. This is the wild-card, since it's unknown how
long this could take.

3. Take the lowest total from the rest of the candidates. Worst case,
that's still less than 100/n percent of the ballots. If we use Danny's
estimate of taking twice as long for each ballot, that comes to less
than 200/n percent of the time of round one.

4. The process could short-circuit at any point along the way if one
person ends up with over half the ballots. But in the worst case, it
would take T + 2T/n + 2T/(n-1) + ... + 2T/3. This reduces to T + 2T(4/3)
for a 9 person race or less than four times the cost of counting a non
IRV race. The actual fraction was 1 829/2520, but who's counting.

The problem I see with IRV isn't the cost, it's the higher chance of
fraud because it generally assumes a central counting area, which has
more chance of fraud than counting at the precinct level. If we want to
keep a record of how each precinct counted, then we could end up with a
long and complicated form listing all possible combinations of
candidates - that would speed up the recounts, but slow down the
original counts. Even if we take a picture of each ballot, we still have
to look at those pictures later if we want to check for fraud.

-Kaj

Danny Swarzman wrote:
> My rough guess is that counting an IRV contest costs about 20 times
> counting a normal election. This is supposing that one step in IRV
> counting takes about twice the time of one step in normal ballots
> because of the extra time to distinguish the exhausted candidates. I
> multiply that by 10 as a typical number of candidates.
>
>
> On Nov 27, 2007, at 8:58 AM, Teresa Hommel wrote:
>
>> I don't have that info. A discussion of different IRV types of math a
>> few years ago convinced me that IRV was a bad idea, and that it was
>> better to hold run-off elections where everyone knows who they are
>> voting for and how the numbers work.
>>
>> Teresa
>>
>> Brent Turner wrote:
>>>
>>> Teresa- Have you analyzed the Instant Run off / Rank Choice /
>>> FairVote aspect- ? I’d like to know how the math works-
>>> Volunteer hours etc - BT
>>>
>>>
>>>
> ------------------------------------------------------------------------
>
> _______________________________________________
> OVC-discuss mailing list
> OVC-discuss@listman.sonic.net
> http://lists.sonic.net/mailman/listinfo/ovc-discuss
> By sending email to the OVC-discuss list, you thereby agree to release the content of your posts to the Public Domain--with the exception of copyrighted material quoted according to fair use, including publicly archiving at http://gnosis.python-hosting.com/voting-project/
_______________________________________________
OVC-discuss mailing list
OVC-discuss@listman.sonic.net
http://lists.sonic.net/mailman/listinfo/ovc-discuss
By sending email to the OVC-discuss list, you thereby agree to release the content of your posts to the Public Domain--with the exception of copyrighted material quoted according to fair use, including publicly archiving at http://gnosis.python-hosting.com/voting-project/
==================================================================
= The content of this message, with the exception of any external
= quotations under fair use, are released to the Public Domain
==================================================================
Received on Fri Nov 30 23:17:28 2007

This archive was generated by hypermail 2.1.8 : Fri Nov 30 2007 - 23:17:32 CST