'Instant runoff results not so instant'

From WRAL.com

Raleigh, N.C. — State elections officials said Wednesday that it will take at least a month to decide a state Court of Appeals race in which an instant runoff was used.

Thirteen candidates ran for Judge James Wynn's vacant seat on the appellate court, and voters were asked to pick their top three choices to avoid a separate runoff election.

Candidates Cressie Thigpen and Doug McCullough collected the most first-place choices on ballots cast Tuesday. They now move to the runoff, where second and third choices for each will be tallied to determine the winner.

The State Board of Elections said it would handle recounts and challenges in other races before taking up the runoff on Nov. 29. Some ballots will need to be hand-sorted to consider the second and third choice selections, which could take several days, officials said.

North Carolina was the only state to use the instant runoff process in a statewide election on Tuesday.


nothiong wrong with a slow recount...

...as long as it's accurate...and as long as you're not al franken.

"...i feel that if a person can't communicate, the very least he can do is to shut up." --tom lehrer, january 1965

how do you make sure the count is accurate?

In the 100% op-scan counties, the count will be somewhat transparent. But wherever they cast some votes on DRE touchscreen machines, they will be using a special algorithm written by an SBOE employee to run on a Microsoft Excel spreadsheet to count the votes. Problem is, the best experts on Microsoft Excel spreadsheets do not advise using them for counting votes because the app was not written to count votes.

Our state election law was amended in 2005 to require testing at a federally-approved lab and federal EAC certification for us to use it. Then the vendor needs to post a performance bond that would pay for any costs associated with fixing problems caused by their election system (including a re-do of the election). Then the vendor has to provide the source code for the system. Then sign a letter stating that the system that they run elections on in our state is the same system that was tested and certified (meaning no uncertified patches) - otherwise they could go to jail.

Did the SBOE worker who wrote the algorithm or Bill Gates himself do any of the things we require from the vendor ES&S that jumped through all the hoops in 2005 after the law was passed? The answer is "no".

Why does that matter? Well, the DRE machines are used in 30 counties that record up to 40% of the votes. Mecklenburg County voters cast 227,648 votes - please tell me how you are going to audit a 227,648 rows spreadsheet?

Of course, it's kind of ironic that it was the two Republican members of the SBOE who pushed for this uncertified IRV vote counting on DRE machines - backed up all the way by Democrat Anita Earls - who is also on the Board of Directors of FairVoteNC, a group that pushed IRV in our state. Ms. Earls personally testified before the legislature twice - once in 2004 and again in 2008 - to promote IRV. Why isn't her work for IRV considered a conflict of interest (or the appearance of same) when it comes to a vote on IRV on the SBOE?

Can you feel it, smell it, taste it in the air.....I sense a lawsuit is not far away!

Chris Telesca
Wake County Verified Voting

what we did in washington...

...is move to vote-by-mail, which solves the whole recount problem rather neatly.

in '04 we had the legendary gregoire/rossi recounts, and that was a punch card election.

so what's my point? maybe this is the kind of situation that could get voters interested in vote-by-mail in nc...because if there's one thing that is certain about vote-by-mail, it's that you can perform a credible recount.

also a question: is there a legal impediment that would prevent mass use of absentee ballots by nc voters?

if not, maybe that's the way to move past this problem without having to enact any new programs; "absentee votes count...and they recount, too!" could be a decent sales pitch.

"...i feel that if a person can't communicate, the very least he can do is to shut up." --tom lehrer, january 1965

vote by mail has its own problems

at home you have no one to ask about the candidates and how to mark the ballot if you have questions.

The simpler way to perform a credible recount is to use 100% paper ballots. But a 100% recount of all the IRV votes would never get done - the Election Directors would claim it's not done, then pressure the county and State Board of Elections into not doing it.

So election integrity and the public confidence in elections has to suffer because it costs too much money to use the same standards for IRV that you use for single-column elections.

Chris Telesca
Wake County Verified Voting

Mail voting makes me nervous

I'm aware that more states are moving in that direction, but as one who routinely works with large mailing lists for a nonprofit, I see how many bad addresses, address changes, etc there are in any given mailout, and it makes me nervous about what happens to ballots that get mailed to the wrong place, or don't arrive, or whatever. And then there's the mailing BACK process--how many unscrupulous postal workers might "lose" a box of ballots from certain regions of a city (yes, I know that the vast majority of postal workers are very honest...but elections bring out funny traits in people), or even simply something like an unsecure mailbox get broken into, someone maliciously pour paint or other liquid into a public mailbox expected to be full of ballots, and just plain "it got lost in the mail".

At least at a BOE, you see where your ballot is going, and only election officials will physically have custody of it. There seem to be far too many unknowns with dropping ballots into mailboxes.

And that doesn't even addess things like domineering husbands dictating to their wives how to vote (though I guess that happens now with absentees) or just plain voting for them. There is no verification (such as it is) that the person filling out the ballot is the one who's supposed to?

thank you for giving Chris another thread

We will not be spending $5 million having a runoff for a Court of Appeals seat which would draw maybe 4.5% of the voters. We can wait a month.

"Man is free at the moment he wishes to be." -Voltaire

What difference will one month make?

If you have been following my posts (and not reading them) you would understand that IRV is a waste of time and money. All over the country where IRV has been used, the candidate who won the first round (in this case Thigpen) goes onto win the instant runoff rounds 98% of the time with a slightly larger plurality win.

We could have just as well declared Cressie Thigpen the winner here and now then wait to start the complicated tabulation done with illegal (uncertified) software).

In other words, you wouldn't need to spend $5 million on a runoff! Thigpen would be the winner today!

But your claims that IRV is better than a runoff is bogus anyway. I've cited studies by other states where they figure out how much it would cost to do IRV right - meaning buy machines that would do IRV without breaking the laws and in what I assume would provide for meaningful audits.

In 2006 and 2008, the MD Legislature claimed it would cost them $3.10 to $3.50 per registered voter to implement IRV - and $0.48 per voter for voter education. If you applied those figures to NC's over 6 million voters, it would cost $20 million to implement and $3 million every year for voter education.

Just the other day, Maine just figured out it would cost them $6 million to implement IRV in their state - but they have less than 1/6th the number of voters we have in NC - for a cost of $6.60 per registered voter. Using their numbers, that would cost NC $40 million to implement IRV now. No figures for voter education.

I do recall getting a google alert for either IRV or Ranked Choice Voting, where one California community spent over $3.00 per registered voter on voter education. With our 6 million voters, that would work out to $18 million. Still think IRV is cheaper?

Communities all over the country have already found out that IRV is more expensive than traditional runoff elections, and deliver results that are no better than plurality wins. Pierce County WA found that IRV doubled their election costs - and dumped IRV. San Francisco costs have skyrocketed and voter turnout is lower since implementing IRV in 2004. Aspen had to hire an independent consultant to implement IRV at roughly $2.50 per vote - and they just voted this Tuesday 2 to 1 to dump IRV.

In Minneapolis, their one IRV election in 2009 cost them $365K more than two elections in 2005 cost them - adjusted for inflation! Voter turnout was the lowest in over 100 years, and the spoiled ballot rate (a good indicator of voter confusion) was 3 times the rate in 2005!

You weren't at my precinct where the majority of voters were vocally unhappy with IRV. I worked there as a poll greeter for most of the day, and as a Democratic Party poll observer.

We had 26 spoiled ballots, a much higher than normal number. There were 8 over-votes in the 26 non-IRV races on the ballot, but 28 over-votes in the one single IRV race. We won't be able to get the rates of spoiled ballots and overvotes for the entire county or state, because the SBOE and the Wake BOE don't want to give us those results. They are contained on the paper summary tape that comes from every voting machine and should be posted for public viewing at every precinct, but very few of us look for it or ask for the spoiled ballot rate.

I didn't worry about the Republican voters, but I spent well over half my time explaining IRV to Democratic voters and to those UNA voters who wanted to vote for Democratic candidates. That is because over half the voters who showed up at the polls didn't know they'd be expected to rank their choices.

The Republicans were claiming this was a Democratic plot to keep a Republican judge from winning. Everyone complained that they spent way too much time on that part of the ballot - one voter felt she spent half her time on IRV alone.

Oh - and if you think that these voters were upset about having to cast their votes in an IRV contest - they were disgusted that the SBOE has known since 2006 that they might have to do county-wide or state-wide judicial IRV elections, and chose to do nothing about it until the last minute. The Republicans were especially infuriated that Republican members of the State Board of Elections were the ones who pushed the "bending" of state and federal election law to use uncertified software to count the IRV.

They all hoped that this would be the last time any of us would have to waste so much time explaining how to vote in a contest - instead of asking voters to vote for our candidates. Having to be told how to vote in some tricky contest insulted their intelligence!

Trust me - there will be lawsuits over this one, and when one of the losers figures out that the SBOE was using uncertified software to tabulate the IRV votes in violation of state and federal law - things will get very interesting. Not sure what will dump IRV first - the lawsuit or the legislature!

Chris Telesca
Wake County Verified Voting

What value do you put on election integrity?

And the public's confidence in the electoral process? You obviously haven't been paying attention to how IRV has been getting pushed all over the country with many promises (saves money, increases turnout, is more democratic, lets you vote your hopes and dreams, etc.).

Do you know how many communities that have done IRV have dumped it? Why can't we learn from their mistakes?

Why do some many Dems like John continue to bury their heads in the sand about IRV and OFA?

Chris Telesca
Wake County Verified Voting

Here is another analysis of the fullity of IRV


Battle for the Board: Waiting for Absentees and Ranked Choice Voting
by Paul Hogarth‚ Nov. 05‚ 2010

With Gavin Newsom on his way to Sacramento, political junkies are fixated on what machinations will lead to appointing the next Mayor. Meanwhile, we still don’t know who has won four contested races to the Board of Supervisors (i.e., the folks who will be making this decision.) There are still around 68,000 late absentees and provisionals citywide (or 25% of all votes cast), but we should expect similar results from the 75% of ballots that have already been counted. The real action will come this afternoon at 4:00 p.m. – when the Elections Department runs its first Ranked Choice Voting tabulation. By the end of today, we should have a clear winner in Districts 2, 6 and 8. As for District 10, the sheer number of candidates and the razor-thin margins suggest today’s count will tell us little – and it may take weeks before we know Sophie Maxwell’s successor.

Much has been reported about the uncounted late absentee ballots – and how their votes may shift some close races. Around 52,000 voters dropped off their absentee ballots at a polling place on Election Day, and more are trickling through the mail. But the Elections Department is not counting them consistently by District – making the citywide numbers misleading. For example, on Wednesday they counted twice as many ballots in District 6 than District 8 – but yesterday counted three times as many in District 8 than in District 6.

And while 68,000 sounds like a huge number, keep in mind that over 190,000 votes – or approximately 75% of the total – have already been counted. Late absentees votes have been known to shift the outcome in close elections, but only by a few percentage points.

The real action will happen this afternoon, when the Elections Department runs the first tabulation of Ranked Choice Voting – where candidates who came in last are eliminated, and their second-choice votes are re-distributed to others until someone gets over 50%.

In District 2, Janet Reilly currently has a 400-vote lead over Mark Farrell – and voters who preferred Abraham Simmons and Kat Anderson will pick the winner. In District 6, Jane Kim is ahead of Debra Walker by 539 votes – and the outcome will be decided by Theresa Sparks voters. In District 8, Scott Wiener now has a 1,500 vote advantage over Rafael Mandelman – and Rebecca Prozan votes will be transferred to each candidate.

But in each of these races, odds are that the current front-runner will prevail. Why? In places that have had Ranked Choice Voting for years (e.g., Australia, New Zealand and Cambridge, Massachusetts), the candidate who was in first place before the tabulations were run ended up winning 95% of the time. Only in rare instances where the outcome is very close (by a couple percentage points) can the second-place finisher end up winning.

San Francisco has only had Ranked Choice Voting for six years – and so far has never had a candidate come from behind in the tabulation to win. Many political junkies (including myself) have erroneously reported in the past that Ed Jew did, but the first Ranked Choice Voting tabulation was an unofficial count that did not include all late absentees. Once the late absentee ballots were counted, Jew actually had more first-place votes than Ron Dudum.

Moreover, recent San Francisco history shows that few voters cast “strategic” ballots when voting for their second-choice candidates. As David Latterman explained yesterday at SPUR’s post-election analysis, second-choice picks tend to mimic the initial election results. Of the three underdogs, Mark Farrell is most likely to pull off an upset – due to the smaller voter margin, and the fact that District 2 has far more uncounted late absentee ballots.

But it’s a totally different story in District 10 – where Tony Kelly currently clings to a 75-vote lead over Lynnette Sweet, with Malia Cohen 46 votes behind. At this point, less than 200 votes separate the top five candidates – and late absentees can shift the order. Although the Elections Department will run its initial (and unofficial) Ranked Choice Voting tabulation today, the results are likely to be incomplete. A District 10 candidate will turn out on top, only to have the order change once more late absentees are counted.

Alex Clemens made by far the best quip at the SPUR presentation yesterday, by quoting an anonymous City Hall insider who had made the following predictions: “District 2 – Janet Reilly; District 6 – Jane Kim; District 8 – Scott Wiener and District 10 – lawsuit.”

And you can see why. Today’s Ranked Choice Voting will declare a “winner” in District 10, who can just as easily fall by the wayside once all the late absentees are counted. So while everyone’s anxious to know who won District 10, don’t take today’s announcement as the final word – because it can very easily change once the official count is determined.

Chris Telesca
Wake County Verified Voting