Re: Numbers of ranked preference votes

From: Arthur Keller <arthur_at_kellers_dot_org>
Date: Thu Apr 29 2004 - 14:56:47 CDT

At 3:15 PM -0400 4/29/04, David Mertz wrote:
>On Apr 29, 2004, at 2:58 PM, Arthur Keller wrote:
>>From your data, it appears infeasible to generate all the
>>combination in advance.
>Looking at the numbers, it is extremely likely that with a moderate
>number of candidates and slots, no two voters at a precinct will
>vote the same way. In fact, it is possible for no two voters at a
>state level to vote the same way (e.g. 7 ranks among 16 candidates
>allows 63M different votes).
>What this tends to suggest to me is that any sort of simple
>incrementing of identical ranked-preference votes is not really of
>significant purpose. Basically, the tallying algorithm almost
>treats all the individual ranks as raw data for an aggregate
>Btw. Doug obviously knows this, but some other subscribers might
>have missed it. Many ranked-preference tabulation styles make
>precinct-level canvassing impossible, or at least incomplete. For
>example, if IRV is used, you don't have the option of reassigning
>the Ralph Nader votes based on who "wins" the first round within a
>precinct. You have to gather together all the votes for the whole
>State before you can determine who makes it to round two, and how
>votes are reassigned. Precincts are largely constrained to
>collecting raw-data rather than reporting local tallies.

For IRV, let's separate out these two canvassing steps: (1)
tabulating the vote lists to create counts of people who vote the
same way, and (2) doing the "run off" algorithm based on the

(1) can be done incrementally, but (2) can only be done at the end.
If you separate (1) and (2) you can rerun (2) if (1) changes, by
using the revised (1) result. This approach works because (1) is a
monotonic function (mathematically), while (2) isn't.

Best regards,

Arthur M. Keller, Ph.D., 3881 Corina Way, Palo Alto, CA  94303-4507
tel +1(650)424-0202, fax +1(650)424-0424
= The content of this message, with the exception of any external 
= quotations under fair use, are released to the Public Domain    
Received on Fri Apr 30 23:17:24 2004

This archive was generated by hypermail 2.1.8 : Fri Apr 30 2004 - 23:17:29 CDT