The proposed charter says: "In any subsequent rounds, to the extent that the votes counted for elected candidates exceeds the threshold, the votes exceeding the threshold are called the surplus and are counted for the continuing candidates who are ranked next on the ballots that had been counted for the elected candidates, with the counting being in proportion to the total numbers of such next ranking." That reference to proportion may very well mean that they will use fractions, which is also what FairVote recommends. I used a basic explanation of both RCV and STV. There can definitely be variations. I haven't heard or read yet exactly how the method would be executed in Portland and perhaps we can't know for sure until the County knows which technology would be used.
Based on discussions in the Charter Commission meetings, I understand the intent is to use fractions. Meanwhile, Cambridge, which is often used by proponents as a U.S. example of STV, does NOT use fractions. Instead, they randomize which ballots' votes transfer. So the choices for surplus transfers are apparently fractional votes from all "Candidate A" voters, or random votes from random "Candidate A" voters. In fact, the two different STV counting methods could provide two different results.
Good explanation overall. But let's understand this "fractions" issue in regard to RCV/STV. When candidates receive more that 25%+1 of the vote in a council election, it is a near mathematical certainty that the percentages to be derived from the second choice votes of those candidates' voters and then applied to the subset of votes that are transferred as surplus, will not be whole numbers initially. More likely you'll see results like 35.24% for Candidate C, 26.17% for Candidate D = 212.44 votes for Candidate C and 157.02 votes for Candidate D. etc. The Council, which has the authority to enact the final voting rules, will have to decide whether to calculate these percentages and the resulting vote transfers by rounding up or down to whoel numbers, tenths or hundredths. I doubt this is a technological problem, since computers can calculate any of those numbers very easily. Nor is it a math problem. Rather, it's just a choice to be made -- by the Council. Will they want fractions of votes to determine an election (imagine a candidate winning by one-third of one vote)? Will they want voters to think that votes will be chopped up into fractions of votes as well as transferred to other candidates? And if they want to avoid these oddities, will they trigger more contention over whether votes should be rounded up (awarding more votes than cast) or down (resulting in fewer votes than cast)? Apparently, Fair Vote recommends using fractions. So what fractions? I'm not making these points to be snarky, but to express my concern that to let the math wonks do this to an election system that in Oregon is working just fine, is to ask for confusion and contention where none now exists.
BTW, the instant runoff version of RCV has none of these problems. Votes are votes, 1=1, no problem there.
An STV expert once warned me that the “more you dig, the more complicated it gets” and that’s certainly true as you start going down the rabbit hole of voting methods and their internal variations! Once I learn more about the method in Portland I’ll try to write a story for my loyal geek readers ;-).
If the measure passes it will be fascinating to watch a council where the majority have come out in opposition to either portions of, or the whole package, have to work out the implementation plan!
Great job explaining the 2 versions under discussion in the Charter proposal.
A couple comments:
1. It is important for the voters to understand that, even though there are 3 council seats to be filled in each of the proposed districts, as you noted "voters cast only one vote." This means that even if they rank all candidates, at most they will have 1 vote counted for only 1 of the 3 seats- not 1 vote for each of the 3 seats.
2. In addition to incentivizing slates of candidates, STV's low bar of 25%+1 means it will be very difficult to ever dislodge an incumbent. And to make matters worse, it takes 25% +1 to get elected, but 50%+1 to get recalled.
3. It is not clear that technology will exist in the near future to allow voters to rank as many candidates as they want and have all the votes tallied. The Commission appears to have made no provision for this if turns out to be the case, by, for example, limiting the number of rankings that will be counted based upon certified voting systems being available.
The Commission report states that “Clear Ballot, the major vendor for Multnomah County, is preparing for a 2023 ranked choice voting election in Colorado that is very similar to the Charter Commission’s proposal.” After first reading the report on ranked choice by the research arm of the Colorado Legislature, I contacted election officials in the 2 Colorado jurisdictions scheduled to begin using RCV in 2023.
• Boulder: Boulder will be using instant run off RCV for mayor only. The 8 council members continue to be elected as before. The voting systems vendor: Dominion. The elections official had never heard of Clear Ballot.
• Broomfield City and County: Broomfield will move to instant runoff RCV for mayor and council. Council members are elected by ward, 2 each in 5 wards. The voting system vendor: Dominion.
My point is that: (1) neither entity uses Clear Ballot, so it likely that it will be a complex and expensive undertaking to change voting systems in Multnomah and the other 2 counties to a certified system that accurately tallies a first-in-the-nation voting method. (2) The new 2023 RCV elections in Colorado are not at all “very similar” to the single transferable vote method proposed for council elections in Portland. I did communicate with Clear Ballot directly. They wrote: "We are actively developing ranked choice voting capabilities within our system, but we have not conducted any work for the city of Boulder or their upcoming Mayoral primary."
Thanks again for your efforts- even on such a "boring" subject.
Bob, I’m glad you point out the common misunderstanding that STV means voting more than once. I’ve come to realize that many people make this erroneous assumption and I think it’s because in America the voting method is relatively new and we focus so much on explaining the technicalities, like the vote transfer process, etc, that we forget to explain the fundamental democratic principles.
You inspired me to want to write a story on vote counting technology. It’s another fascinating aspect of election reform!
"Those additional votes – called surplus votes – transfer to the second choice on the respective voters’ ballots. Let’s say you checked Candidate A as your first choice and Candidate B as your second choice. Candidate A wins a seat and gets a few surplus votes. If your vote is one of them, it now counts for Candidate B instead." Well, no, not exactly...
*Everyone* who voted for Candidate A, will have a *fraction* of a vote transferred to their second choice candidate, so that the total of the surplus is distributed proportionally to the second choices of Candidate A voters.
When it's too confusing for the Charter School instructor, maybe it IS too confusing.
While every voter gets to cast only one vote, some voters will get more than one vote in the tallying process. Those who vote for a winner in Round One will see some of their 2nd choice votes enter the tally in Round Two, while those who voted for a non-winner in Round One will still have only one vote in Round Two (either their first choice votes if their candidates are still viable or their second choice votes if their candidates are eliminated). Of course, Round 2 winners can generate surplus votes as well, so then, rinse and repeat.
Another puzzling aspect of all this comes from the question: 25% of what number? In Round Two and subsequent rounds of tallying, the vote total changes, both because of the addition of fractional votes and the deletion of so-called exhausted ballots (which don't contain a 2nd or 3rd or later choice). So is it 25% of the total votes cast (as in Round One) or 25% of the vote tallied in subsequents rounds?
I feel like we're engaging in one of those brainstorming sessions where every difficult question is relegated to a white board, labeled the "parking lot," which in this case will be left for the City Council to figure out later on. Meanwhile, let's adjourn to the bar and celebrate our cool new idea.
The proposed charter says: "In any subsequent rounds, to the extent that the votes counted for elected candidates exceeds the threshold, the votes exceeding the threshold are called the surplus and are counted for the continuing candidates who are ranked next on the ballots that had been counted for the elected candidates, with the counting being in proportion to the total numbers of such next ranking." That reference to proportion may very well mean that they will use fractions, which is also what FairVote recommends. I used a basic explanation of both RCV and STV. There can definitely be variations. I haven't heard or read yet exactly how the method would be executed in Portland and perhaps we can't know for sure until the County knows which technology would be used.
Based on discussions in the Charter Commission meetings, I understand the intent is to use fractions. Meanwhile, Cambridge, which is often used by proponents as a U.S. example of STV, does NOT use fractions. Instead, they randomize which ballots' votes transfer. So the choices for surplus transfers are apparently fractional votes from all "Candidate A" voters, or random votes from random "Candidate A" voters. In fact, the two different STV counting methods could provide two different results.
Good explanation overall. But let's understand this "fractions" issue in regard to RCV/STV. When candidates receive more that 25%+1 of the vote in a council election, it is a near mathematical certainty that the percentages to be derived from the second choice votes of those candidates' voters and then applied to the subset of votes that are transferred as surplus, will not be whole numbers initially. More likely you'll see results like 35.24% for Candidate C, 26.17% for Candidate D = 212.44 votes for Candidate C and 157.02 votes for Candidate D. etc. The Council, which has the authority to enact the final voting rules, will have to decide whether to calculate these percentages and the resulting vote transfers by rounding up or down to whoel numbers, tenths or hundredths. I doubt this is a technological problem, since computers can calculate any of those numbers very easily. Nor is it a math problem. Rather, it's just a choice to be made -- by the Council. Will they want fractions of votes to determine an election (imagine a candidate winning by one-third of one vote)? Will they want voters to think that votes will be chopped up into fractions of votes as well as transferred to other candidates? And if they want to avoid these oddities, will they trigger more contention over whether votes should be rounded up (awarding more votes than cast) or down (resulting in fewer votes than cast)? Apparently, Fair Vote recommends using fractions. So what fractions? I'm not making these points to be snarky, but to express my concern that to let the math wonks do this to an election system that in Oregon is working just fine, is to ask for confusion and contention where none now exists.
BTW, the instant runoff version of RCV has none of these problems. Votes are votes, 1=1, no problem there.
An STV expert once warned me that the “more you dig, the more complicated it gets” and that’s certainly true as you start going down the rabbit hole of voting methods and their internal variations! Once I learn more about the method in Portland I’ll try to write a story for my loyal geek readers ;-).
If the measure passes it will be fascinating to watch a council where the majority have come out in opposition to either portions of, or the whole package, have to work out the implementation plan!
This paper gives a good overview of various methods for allocating the surplus to continuing candidates, including the three varieties of the Gregory Method: https://www.elections.wa.gov.au/sites/default/files/content/documents/Determining_the_result.pdf
Great job explaining the 2 versions under discussion in the Charter proposal.
A couple comments:
1. It is important for the voters to understand that, even though there are 3 council seats to be filled in each of the proposed districts, as you noted "voters cast only one vote." This means that even if they rank all candidates, at most they will have 1 vote counted for only 1 of the 3 seats- not 1 vote for each of the 3 seats.
2. In addition to incentivizing slates of candidates, STV's low bar of 25%+1 means it will be very difficult to ever dislodge an incumbent. And to make matters worse, it takes 25% +1 to get elected, but 50%+1 to get recalled.
3. It is not clear that technology will exist in the near future to allow voters to rank as many candidates as they want and have all the votes tallied. The Commission appears to have made no provision for this if turns out to be the case, by, for example, limiting the number of rankings that will be counted based upon certified voting systems being available.
The Commission report states that “Clear Ballot, the major vendor for Multnomah County, is preparing for a 2023 ranked choice voting election in Colorado that is very similar to the Charter Commission’s proposal.” After first reading the report on ranked choice by the research arm of the Colorado Legislature, I contacted election officials in the 2 Colorado jurisdictions scheduled to begin using RCV in 2023.
• Boulder: Boulder will be using instant run off RCV for mayor only. The 8 council members continue to be elected as before. The voting systems vendor: Dominion. The elections official had never heard of Clear Ballot.
• Broomfield City and County: Broomfield will move to instant runoff RCV for mayor and council. Council members are elected by ward, 2 each in 5 wards. The voting system vendor: Dominion.
My point is that: (1) neither entity uses Clear Ballot, so it likely that it will be a complex and expensive undertaking to change voting systems in Multnomah and the other 2 counties to a certified system that accurately tallies a first-in-the-nation voting method. (2) The new 2023 RCV elections in Colorado are not at all “very similar” to the single transferable vote method proposed for council elections in Portland. I did communicate with Clear Ballot directly. They wrote: "We are actively developing ranked choice voting capabilities within our system, but we have not conducted any work for the city of Boulder or their upcoming Mayoral primary."
Thanks again for your efforts- even on such a "boring" subject.
Bob, I’m glad you point out the common misunderstanding that STV means voting more than once. I’ve come to realize that many people make this erroneous assumption and I think it’s because in America the voting method is relatively new and we focus so much on explaining the technicalities, like the vote transfer process, etc, that we forget to explain the fundamental democratic principles.
You inspired me to want to write a story on vote counting technology. It’s another fascinating aspect of election reform!
"Those additional votes – called surplus votes – transfer to the second choice on the respective voters’ ballots. Let’s say you checked Candidate A as your first choice and Candidate B as your second choice. Candidate A wins a seat and gets a few surplus votes. If your vote is one of them, it now counts for Candidate B instead." Well, no, not exactly...
*Everyone* who voted for Candidate A, will have a *fraction* of a vote transferred to their second choice candidate, so that the total of the surplus is distributed proportionally to the second choices of Candidate A voters.
When it's too confusing for the Charter School instructor, maybe it IS too confusing.
Good point.
While every voter gets to cast only one vote, some voters will get more than one vote in the tallying process. Those who vote for a winner in Round One will see some of their 2nd choice votes enter the tally in Round Two, while those who voted for a non-winner in Round One will still have only one vote in Round Two (either their first choice votes if their candidates are still viable or their second choice votes if their candidates are eliminated). Of course, Round 2 winners can generate surplus votes as well, so then, rinse and repeat.
Another puzzling aspect of all this comes from the question: 25% of what number? In Round Two and subsequent rounds of tallying, the vote total changes, both because of the addition of fractional votes and the deletion of so-called exhausted ballots (which don't contain a 2nd or 3rd or later choice). So is it 25% of the total votes cast (as in Round One) or 25% of the vote tallied in subsequents rounds?
I feel like we're engaging in one of those brainstorming sessions where every difficult question is relegated to a white board, labeled the "parking lot," which in this case will be left for the City Council to figure out later on. Meanwhile, let's adjourn to the bar and celebrate our cool new idea.