[EM] No. Condorcet and Hare do not share the same problem with computational complexity and process transparency.

[robert bristow-johnson]

> On 03/16/2024 5:52 PM EDT Michael Ossipoff <email9648742 at gmail.com> wrote:
> 
> 
> In 35 years of promotion of a ranked-method & heavy spending, what does FairVote have to show for it? Two states & some cities. That isn’t success, for 35 years.

It's better than nothing, but FV needs to be more honest about failures.

The repeal effort in Alaska in 2024 is grossly underestimated by the IRV happy-talkers just as the repeal effort in Burlington was underestimated in 2009 and 2010.  There is evidence of that in the postings I made to this list in those years.

I predict that RCV in Alaska will be repealed in November, we'll see.

> 
> If they’d been instead offering Approval there’d likely be a lot more success.

That we don't know.  Fargo and St. Louis are the only governments using Approval.  And it doesn't take away the pressure on voters to vote tactically whenever there are 3 or more candidates.

> …& don’t forget that Condorcet, too, has a very computationally-intensive & computationally-demanding count,

We talked about this before, Michael, when I posted.  Because of Precinct Summability and the decentralization of the tabulation (that does *not* exist with Hare IRV), there is *no* practical bottleneck of computational burden (like there is with Hare IRV that first requires the centralization of all individual ballot data).

When the ballot is inserted into the ballot machine, the ballot is scanned and if the ballot is accepted as not spoiled (Dominion is quite a pain in the ass about this), then tallies are immediately incremented, as they are now with FPTP.  That will take all of a millisecond.  When the election day ends and polls close, all there needs is to print out the tallies for that polling place.  Then they are securely published and no one is gonna be able to "... find, uh, 11780 .. votes."

If they're counted by hand, if C is the number of candidates, then IRV may require up to C-1 passes through the ballot pile.  Condorcet will require C(C-1)/2 passes with the ballot pile.  Laborious, but straight-forward.  And if the precinct is so large that this manual tabulation is unfeasible, then they just need tabulators, no matter what the method is.

It's not at all the same problem as with IRV, where, other than 1st-choice votes, no one knows what's inside the tabulators or ballot bags that we haul from each polling place to City Hall in the evening.  But with FPTP and Condorcet (and Approval and Score, I will concede), the only opaque section of the path lies solely in the tabulators (human or machine) at the polling place.

> with the consequent loss of transparency, & difficulty of security-auditing for count-fraud.

Nope, that's a falsehood.  Condorcet is Precinct Summable and those tallies add up and the original tallies come right outa the same machine that the ballots go into, just like FPTP.  There are no intermediate steps that are opaque.  It's not at all the same problem as with Hare RCV.

If we have Precinct Summability the "transparency, & difficulty of security-auditing for count-fraud" is the same as we have now with FPTP.

But without Precinct Summability, there is a serious loss and I am trying to persuade policy makers in our state of that fact.  So far I got two state reps and two state senators to understand this.  This bill (which is not going anywhere) was inspired by me: https://legislature.vermont.gov/bill/status/2024/H.424 

But, unfortunately, S.32 (Hare) is doing better, but *is* getting hung up a little in the House Gov Ops.

--

r b-j . _ . _ . _ . _ rbj at audioimagination.com

"Imagination is more important than knowledge."

.
.
.