Your particular example does not suggest anything about strategic voting. If the business interests prop up an African American candidate, such a candidate would very likely receive little first-choice support. So even if African American voters ranked him or her second, they would eventually rank that candidate you spoke to third, presuming he is correct in his assessment that he can convince them of his qualifications. So the candidate you spoke to is right: he will be the ultimate beneficiary.

Lastly, it is false that IRV elections can’t be audited and make it harder to detect electoral fraud and errors. (You used the incorrect phrase “voter fraud,” fraud by a voter at the voting booth, when what you meant is “electoral fraud,” which is fraud embedded in the electoral system itself, such as the machines.) San Francisco audits all of its IRV elections, and it does so in a manner that is more likely to detect errors and fraud than traditional audits. For more information on how these audits are run, please this document about IRV and Election Integrity.

A minor footnote on the Albany ballot measure: the main issue is whether to have an elected mayor (like many California cities, Albany currently has a mostly ceremonial mayor chosen by the city council from its own ranks). Given that neighboring Berkeley and Oakland are committed to implementing IRV soon, it seems highly probable that Albany would choose IRV rather than an expensive runoff election.

This candidate thinks that he can educate the African-American voters to vote for him second, and since he feels he will be in the top-two IRV runoff, he will win.

But wait – what makes him think that the conservative business interests who recruited the African American candidate to siphon votes from the progressive candidate won’t work hard to educate voters who vote for the African American candidate to vote for the conservative business candidate second? The answer – nothing will prevent it, and strategic voting will still take place in IRV – it will just be harder to do, take more money and resources to accomplish, thus driving up the cost of campaigning and favor candidates with deep pockets.

Furthermore, the fact that IRV elections can’t be audited with any of the ease of traditioal single column elections will make it harder to detect voting fraud or even human errors. That is what happened in the Cary IRV election – many errors were made, and some were detected by observers (like me) but many were only found behind closed doors and the public was told it won’t happen again – not what we want to encourage for election integrity.

And I challenge you to produce any evidence that the same kind of rigorous audits were done of any RCV elections in San Francisco. People have tried without success to get an idea of what sort of audits have been done there.

Mandatory ranking in IRV is called a full preference ballot, and there are some colleges that require you to rank every candidate in every race – even the ones you don’t like – in order for ANY of your votes to count. Not being required to rank every candidate in every race – uin other words you might get some undervotes – is called a partial preference ballot. And with those you always run the risk that you will exhaust all the ballots and not have a majority winner – and then you need to have a traditional runoff. That is what happened in Cary – 1401 votes is not 50% plus one vote of the 3022 ballots that were cast in the District B race. There is a reason to get wound up about how some people claim that IRV ensures a majority winner in one election when it simply isn’t true.

I prefer optional-ranking IRV (where a voter is not required to rank all the candidates) over mandatory-ranking IRV. All IRV elections for public office in the US are optional-ranking and no IRV organization in the US is proposing anything different. There’s no need to get all wound up about mandatory-ranking when no is proposing it.

First, Cary did not replace traditional runoff elections with IRV. Cary participated in an IRV pilot, an experiment to see if IRV worked or did not work.

Yes, Cary participated in an IRV pilot program, a program which replaced the traditional runoff with an IRV election. As you later mention, that program was extended due to its success.

IRV advocates are fond of saying that IRV ensures majority winners in one election instead of two. That did not happen in Cary . . . By the time all ballots were exhausted, the leading candidate had only 1401 votes – not the 50% plus one vote majority.

An absurd argument. IRV does ensure that the winner have a majority of the non-exhausted ballots. Your example includes ballots on which the voter deliberately chose to abstain from the runoff by not choosing a second and third choice. In a traditional two-round system, this is equivalent to them staying home or casting a blank ballot in the runoff. Abstentions and no-shows are never counted in the denominator when determining the “majority”.

As for the rest of your misinformation about costs and transparency, they were soundly put to rest by the Wake Country Board of Elections, whose Chair and Secretary authored this letter soundly disputing your claims.

You’ve done an admirably job posting your misleading graffiti to every blog on the Internet. Fortunately, the NC legislature, the NC Board of Elections, editorial boards around NC, and the NC voters see IRV as a money-saving, turnout-boosting success that deserves to be continued.

First, college elections do not require the same election integrity measures that real elections require. At NCSU in Raleigh, NC, the IRV elections were done on a $1800 computer program written by a recent graduate. There was no way to audit the election and the student “Board of Elections” didn’t even count the votes. In the real world, our election software must be federally certified and should be auditable.

The NCSU election also required that each student rank each and every candidate in each and every election on the ballot, otherwise no votes would be counted. That can’t happen in the real world either.

And even with that requirement, NCSU still admitted that you might not get a majority winner even after all the ballots were exhausted – and a traditional runoff election might still be needed.

It does not bode well for elections if college students are getting used to voting on computers with no certified software, no ability to audit, and if most of them don’t really understand who they are really voting for. Just because you have more choices doesn’t mean that you always make intelligent choices. In Australia, many of the votes that voters are required to cast are “donkey votes” – votes that are just cast because they are forced to rank all candidates in all elections.

IRV advocates are fond of saying that IRV ensures majority winners in one election instead of two. That did not happen in Cary. 3022 votes were cast in the first round of Cary’s District B race, but no one got the 50% plus one vote majority (1512 votes). So they proceeded to count the 2nd and 3rd column votes. By the time all ballots were exhausted, the leading candidate had only 1401 votes – not the 50% plus one vote majority.

The Wake BOE should have announced that IRV did not deliver a majority winner and held a traditional runoff election. I am sure that contest would have delivered a different winner, since the second place finisher in the IRV race actually got more 2nd and 3rd column votes, but she could not overcome the number of voters that the 1st column leader brought with him. So I am not really sure that is fairer for the candidate who lost the election but who had more second and third column votes cast for her.

In other NC races, the second place finisher in the traditional race went onto win the runoff election. Both were African-American candidates. In one race in Rocky Mount, NC – turnout was greater in the runoff than in the main election. That does not always happen, but IRV doesn’t allow people to make a choice between the top two candidates and learn more about them after the main election – a flaw with “preferential voting” that even Robert’s Rules of Order recognizes.

And as to IRV being easier, I am not really sure you can say it was easier for voters or for election administrators unless you release all the raw voting data to make sure that voters really understood how to rank their choices so that all choices were counted. For example, one candidate told his supporters to rank him 1st, 2nd and 3rd. And one voter ranked all three of his choices in the same column – rendering his ballot uncountable. Who knows how many more were like that? And the Wake BOE didn’t follow proper procedures for tabulating IRV votes, and most likely UIRV tabulations procedures are illegal under NC election laws. The pilot was extended, but with the requirement that IRV only be used once goals, standards and criteria can be developed for implementation and evaluation that follow NC election laws.

And as far as IRV being cheaper, no records were kept of all the expenses and time spent promoting and educating voters about IRV. Other states have studied the issue – like Maryland. They found that IRV was much more expensive than traditional elections when you include the costs of software and hardware, training for election administrators and poll workers, extra documentation, higher ballot design costs (for the longer ballots), and voter education. Applying the MD costs per voter times the number of NC voters, you would end up with implementation costs between $17 and $20 million, and on-going voter education costs of between $2.8 and $4 million. Compare that with the cost of holding rarely-needed runoff elections even statewide, and you would find that IRV could cost NC voters $40 million or more over traditional runoff elections over the next 33 years.

We can study other ways to reduce or eliminate the need for runoff elections, but we shouldn’t lower the bar for election integrity. So I am not really sure that we can call IRV a success in North Carolina.