[EM] IRV vs Plurality (Dave Ketchum)
robert bristow-johnson
rbj at audioimagination.com
Sat Jan 16 11:31:46 PST 2010
On Jan 16, 2010, at 12:05 PM, Abd ul-Rahman Lomax wrote:
> At 11:22 AM 1/15/2010, robert bristow-johnson wrote:
> (about a voting security expert)
>> you are in the rabid anti-IRV party.
>
> Robert, your slip is showing.
what slip? i don't have nuttin' under me kilt.
> Experts in various fields tend to be strongly against IRV.
> Political activists who are working for IRV tend to see strong
> opposition as "rabid." You want to see rabid opposition, you'll
> have to look elsewhere, though.
and FairVote.org will point to experts that strongly advocate IRV.
big fat hairy deel.
whether it's here or on the Burlington blog or longer ago at the
Fairvote site (that i have since gotten tired of), i have never
appealed to authority in evaluating or advocating any method.
> I'm thinking over the opposition material. It might look rabid to
> someone who isn't aware of the problems,
but i think i *am* aware of the salient problems. but despite that,
i support the goals that we had in adopting IRV. i still don't
support "correcting" IRV by reverting back to FPTP, given that is the
choice presented (the reverted rules would include a delayed-runoff
for less than 40%). and "voting security" concerns have been
persuasive. even with IRV, with a reasonably small number of
credible candidates (and assuming the worse case, that "write-in" is
always the same person, without yet checking), there are a finite
number of ranking permutations, and there can be a ballot for each.
continued below...
On Jan 16, 2010, at 10:41 AM, Kathy Dopp wrote:
> On Sat, Jan 16, 2010 at 2:06 AM, robert bristow-johnson
> <rbj at audioimagination.com> wrote:
>>
>> On Jan 15, 2010, at 11:34 PM, Jonathan Lundell wrote:
>>
>>> On Jan 15, 2010, at 7:51 PM, Kathy Dopp wrote:
>>>
>>>> Imagine sending all your ballots nationwide to DC for manual
>>>> counting
>>>> to check the outcome of a Presidential election. We'll simply
>>>> let the
>>>> GW administration, for instance, count the results in his own IRV
>>>> election!
>>>
>>> That's something of a non sequitur. Anyone with all the ballot files
>>> (every state, for example, or anyone else) could do the count.
>>
>> and, in fact, it can be decentralized to the extent it is now.
>> each state
>> could have their central place, and in turn, each county, each
>> precinct.
>> the entire tree could be a public record on the internet that has
>> links to
>> child nodes or parent node. with 3 credible candidates there are
>> 9 piles to
>> have to maintain. each precinct sorts the ballots into one of 9
>> piles and
>> counts it and puts the 9 numbers up in this public place on the web.
>> everyone can check their own node to see that it isn't
>> misreported. i do
>> not see why, physically, it would be more vulnerable to attack by the
>> government in power that what is presently the case. it's a
>> factor of 9/2
a typo, i meant to say 9/3. IRV has 9 piles, FPTP has 3.
>> more numbers to keep secure with that ranked ballot.
>
> I was talking about IRV voting. Where do you get "9" piles from?
it's 3!/0! + 3!/2! = 6 + 3
> (9 would be the number of Condorcet tallies for 3 candidates,
no, that would be 6. for N candidates, i think there would be
N-1
SUM{ N!/n! } - N!/1!
n=0
... piles if only relative ranking is salient.
the second term of the summation (in the case of N=3, it's the number
of permutations of ranking 2 candidates out of a pool of 3), counts a
superfluous permutation because when only one candidate is unranked,
it's equivalent to ranking him last. but we have to account for the
case where 2 or more candidates are unranked (and tied for last).
> *not* the
> number of ballot piles for either Condorcet - which does not require
> ballot sorting to hand count - and *not* the number of ballot piles
> for IRV voting.
doesn't matter if you don't worry about someone ranking their
favorite candidate "2" and no one a "1". there are a finite number
of meaningfully different ranking permutations and you need only 1
pile for each. if N is 3, that number is 9.
> If you want to make IRV precinct-summable for 3 candidates, it
> requires 3*2 + 3*2 + 3 = 15 separate tallies.
the middle term of 3*2, you don't need.
> To count IRV by sorting piles of ballots requires far fewer piles
> than 9 but also to do decentralized as you suggest would require
> everyone in the entire country in all precincts sitting around waiting
> for all the late-counted ballots to be ready and waiting for the total
> results to be tabulated centrally somewhere so they could sort the
> ballots for the next round - totally undoable practically.
if it makes a difference to the election outcome (like it's damn
close), isn't sorting and counting late ballots going to be necessary
anyway?
> Of course the number of tallies to make IRV/STV precinct-summable
> grows exponentially as the number of candidates grows and is equal to
> more than the total number of voters who vote in each precinct most of
> the time with a larger number of candidates.
it depends on who we expect are candidates we need to count. for a
very large (nationwide) election, there might be minor candidates
that we can sweep in with "write-in".
>>
>> each state, each little government would be responsible to confirm
>> their
>> precinct totals on the map and everybody gets to look at it. what's
>> particularly insecure about that?
>
> I don't think you yet understand the counting process for IRV/STV.
i *do* understand it. at least for how it's done in Burlington, i
understand exactly what information is used, how much information it
is, and how it is used in tabulation. not completely happy about the
tabulation algorithm, but i'm not worried about "precinct summable"
security issues. not unless there are 10 or 20 candidates. it's a
finite number of piles. computers can handle the numbers for a lot
of piles.
> Why
> not create a set of 200 ballots for one precinct with a mixture of all
> 15 unique ballot combinations on them for three candidates and try
> counting them so you can fully understand the process.
for 3 candidates, the number of piles needed is 3! + 3, not 3! + 3! + 3
but, admittedly, that number gets bigger fast. for N=4 the number of
piles becomes 40. more of a headache. but that's why we have
computers and optical scan machines. each precinct need not (and in
Burlington, they *don't*) physically sort the piles, the paper
ballots are in the order that they were inserted into the machine
(that issue could be a violation of voter's privacy if there is a
video camera recording each person inserting their ballot. if there
is a questionable ballot (for manual inspection) that are in the
original voter tally files, they can map from the file to the
physical ballot for a look see.
but, in IRV, it's nice to use a centralized computer to manipulate
the data of the ballots, but several different computers, residing
with several different interests, can duplicate the task and results
compared. and the physical instruments can remain as decentralized
as the states and counties deem is efficient. probably, precinct
ballots would go to city or county storage and stay there. there is
no need for all of these paper ballots to go to Washington in case
something manual has to be done with them.
but, IRV can pick the wrong winner from the data that it's given.
and it can suffer the spoiler problem. and it can punish a class of
voters for voting sincerely (thus encouraging strategic voting in
future elections). those are violating the first 3 principles (which
want wanted to prevail with IRV, it's why we voted for it in 2005) in
that little paper i wrote. and it happened in Burlington Vermont in
March 2009. out of 4 principles it only passed the "avoid delayed-
runoff because fewer voters show up" test. if flunked the other 3 in
2009. but passed them in 2006.
> Cheers
why, thank you.
--
r b-j rbj at audioimagination.com
"Imagination is more important than knowledge."
More information about the Election-Methods
mailing list