Sat 11 Mar 2017 8:14PM

NOTA (none-of-the-above)

AW Aaron Wolf Public Seen by 9

I like the idea. It amounts to rejection of the whole pool of candidates.

In IRV, if exhausted ballots add up to the majority (and people understand this so they can choose to not rank candidates they disapprove of), the election is canceled.

In SRV, perhaps it would be that to reach the runoff, a candidate must have a majority of ballots giving them a non-zero score.

If a voter abstains from a race entirely, they wouldn't count toward what constitutes a majority, so there needs to be a way to explicitly vote NOTA in IRV (score some 0s in SRV works).

What happens if NOTA wins?



Alan Zundel Sat 11 Mar 2017 10:10PM

It would have to be NOTO (none of the others) to be clear when to use it (unless it was your first choice, which would be another confusing usage). As I understand it, this insures majority support for the winner because what were exhausted ballots would become NOTO ballots and count for the total number of votes from which the winner needs 50%+1.

I also like the variation of IRV in which the candidate with the most last place votes is eliminated, rather than the candidate with the fewest first choice votes. (I forget the name of this.) Used along with NOTO (NOTO not qualifying as a "last place vote" for this purpose) how would that work? It appeals to me at first pass, but I confess I have not thought thru the ramifications.

In the event NOTO wins, I suppose there would be a temporary appointment as under vacancy rules until another election can be held. That is what Seth indicated elsewhere.


Aaron Wolf Sat 11 Mar 2017 11:19PM



insures majority support

I think "ensures" is what you meant, but yes! Not ONLY does it do that by addressing exhausted ballots, it also does it by encouraging non-supportive voters to leave candidates off, i.e. it gives extra weight to people intentionally allowing exhaustion.

That said, I now see a downside: if I disapprove of both A and B but prefer A to B, I make wish to still rank A in order to stop B. There could be a case where A and B are both approved by the majority of voters, so my NOTO for them won't block them and it's more valuable to me to get my preference.

It would be worse to have my preference of A be interpreted as support under a NOTO system than the non-NOTO version where you it's more obvious that you don't know if I actually approve.

So, maybe the ideal is a system that would let me disapprove of both A and B while still expressing my preference. If they get to majority approval, my preference is counted. But if enough of us disapprove, the NOTO result wins. I'm not sure what that ballot would look like, but one possibility could be a score ballot where numbers in the lower-half of the range are considered disapproval for the sake of NOTO and explicitly marked that way. Yes, this is getting more complicated.

variation of IRV in which the candidate with the most last place votes is eliminated

Yeah, I like that far better myself. It reminds me of 3-2-1 voting (but isn't identical). In a most-last-place-eliminated situation, is last place the lowest ranking allowed (maybe 6th of 6) rather than just lowest-from-each-ballot? The concern I'd have there is ensuring that voters actually express their last place. If most voters only marked top 3 of 6, then eliminating-most-last-place wouldn't work out so well.


Alan Zundel Sun 12 Mar 2017 2:25AM

Any candidate not in ranked on your ballot is regarded as a last place candidate.


Aaron Wolf Sun 12 Mar 2017 8:46PM

does it do anything about IRV spoilers?

NOTO certainly doesn't stop IRV spoilers. Consider a case where 100% of candidates are still approved by everyone overall. But NOTO does provide a measure to stop the worst possible cases of spoilers where the winner would be truly a lesser-evil, seen as clearly evil by the majority.


Sara Wolf Sun 12 Mar 2017 1:55AM

The key criteria for a NOTO option is to make sure that voters don't somehow sacrifice their voting power in the process of trying to convey a NOTO vote. I.e. DO NOT skip filling out the rest of your ballot just because you checked a box for NOTO!

In SRV's first round there are 2 ways to look at the total. SRV already counts the total scores, but counting an average score for each candidate gives a different bit of key info. (The average would include scores of 0, but would ignore blanks as a neutral vote) If a candidate has an average lower than some standard that equals a vote of no confidence, 1 for example, then they would be disqualified. This isn't exactly NOTO but similar. I'm not sure if this meets my criteria above actually. Hmm...

Another idea would be to just add a NOTO box at the end of each seats section on the ballot. The box would clearly state to still fill out your preferences even if you check the NOTO box. This would work for SRV, IRV, or Plurality and could be passed as it's own ballot measure regardless of voting system.

If NOTO wins the candidates are all eliminated, the election starts over with new blood and the standing politician stays in office till a successor is elected.


Aaron Wolf Sun 12 Mar 2017 5:38AM

A box at the end (not connected to any candidate) would be NOTA, not NOTO. Yes, there could be a NOTA box like that which would cancel the election if it got the majority, but that's too blunt for me. I like the NOTO approach.

I imagine one way to do NOTO with IRV or Ranked Pairs (i.e. an ordinal ballot) would be to just have it as a spot in the order as in the order A, B, C, NOTO, D, E, F where everything above NOTO is approved, everything below is disapproved. That ballot would go for D over E if both are in the running but would count as a cancel-the-election vote if D or E or F work out to be the winner (and NOTO would be the result if that's the case for a majority of ballots).

Off the top of my head, a way to do that with score ballot is to give a nobody-below-this score. Basically: score everyone just in terms of relative level of support, as makes sense for SRV. Then give a score for cancel-the-election. Basically: "I vote FOR canceling the election if any of the candidates I scored X or lower (such as 4 or 2 or 1 )"

Then, SRV results work out as usual, but if the winner is a candidate where the majority disapproved of them, the vote is cancelled.

That approach would allow voters full expressivity, meeting your (Sara's) criterion. With a 0-9 ballot, you could say that even your top-supported candidate of 9 is disapproved if you think they are relatively that great compared to the rest of the field but are still bad.


Mark Frohnmayer Sun 12 Mar 2017 6:55AM

NOTA fan. What's the impetus? How about we institute a voting system that lets any number of candidates compete equally for the job? What's the real win here?


Aaron Wolf Sun 12 Mar 2017 7:29AM


  • express disapproval in absolute terms (not just relative to the pool)
    • and, of course, approval too by not disapproving…
  • actually have ability to say "no, this is unacceptable, bring us better candidates"
  • the threat of that as potential incentive for parties to really care about putting forward approvable candidates rather than lesser-evil ones

I'm not convinced this is necessarily good. I want to see the researchers study it further. But I really like the first point personally. I want to express absolute approval/disapproval, and I think lots of others do too. What we do with that information I'm still trying to understand, but I think doing something with an actual potential effect is needed to be more than just a connect opinion poll. But mostly wanting to discuss the idea.


Alan Zundel Sun 12 Mar 2017 5:01PM

I'm not sure how much NOTO with IRV solves either. I would like to hear Seth's view of it, but he's not here. It would prevent a minority winner as measured by total ballots cast, but does it do anything about IRV spoilers? I might be missing something, but I can't see how.

I like the "eliminate candidate with most last place votes" variation of IRV. That seems to me to prevent IRV spoilers. We should have a separate thread on that. I will try to find the name of this and start one.


Alan Zundel Sun 12 Mar 2017 5:04PM

Coomb's method:


Apparently vulnerable to strategic voting.

Load More