IRV complaint vs. FPTP: "your entire vote is not counted"
-
@rob said in IRV complaint vs. FPTP: "your entire vote is not counted":
unless the election comes down to a near tie, your vote isn't going to change the outcome.
That's a semantic trick right there. In almost all seriously-proposed systems, unless it comes to breaking a tie by casting lots, the outcome depends only on the votes. The whole semantics of a vote is in its effect on the outcome. If one person slaps you, you probably survive. If 10,000 people line up and slap you, you might not survive. Which one of them killed you? Would you argue that unless you were already almost dead when person x slapped you, the slap by x isn't going to change the outcome?
-
@jack-waugh If it is a semantic trick, that's kind of the point. There are so many different places in the process you can choose to say "your vote doesn't count", this is just one more of them.
You don't know going in that your vote isn't going to count in one way or another. In many cases, your vote is "there if it is needed," but in some cases it isn't needed. And that's ok and not worth fretting over.
@jack-waugh said in IRV complaint vs. FPTP: "your entire vote is not counted":
Would you argue that unless you were already almost dead when person x slapped you, the slap by x isn't going to change the outcome?
I might argue that, sure, but my overarching argument is that none of that matters. Overall, your vote is counted. Likewise your slap, or your potential to slap, does matter.
And in your analogy, even the person that never got a chance to get a slap in, helped defeat me. Why? Because they were there just in case. That's my whole point.
-
@rob said in IRV complaint vs. FPTP: "your entire vote is not counted":
Remember, under any method, unless the election comes down to a near tie, your vote isn't going to change the outcome. We all know that, it's the nature of elections.
Not only did you say that, you emphasized it with boldface. I still don't understand what you mean by it, and what is its relevance. Are you arguing from it that certain kinds of argument are invalid? Is it falsifiable/verifiable?
FPtP uses a strictly additive tally. The winner depends on total scores received by the candidates by adding up the votes. Even though the form of the vote is unduly restricted, at least, the vote is certain to increment the tally. You could argue that this also applies to your top choice in IRV and that it is guaranteed to be added in to the tally for the first round, which determines to eliminate one candidate. And I think you'd be right. IRV is at least as powerful for the voter as FPtP, just on the basis of the top choice. However, the intermediate rankings don't necessarily count. And when we offer the voter ranking as the abstract grammar of the ballot, i. e. we allow the freedom of movement to rank some candidates, our message to the voter carries the insinuation that those rankings will be counted toward the outcome. If we were doing Copeland, they would. But IRV wants to throw a lot of them away. So, it is as powerful as FPtP, but by choosing IRV, we, the election designers, are lying to the voters about how much power we are giving them.
@Sass, does that about cover the "counts all the votes" argument that you make in favor of FPtP over IRV? Do Rob's points suffice to knock down your argument?
-
This post is deleted! -
@jack-waugh Not sure what you don't understand.
I'm saying that there are a huge number of ways you could say "your vote doesn't count." The one I mentioned being the most obvious one.
You have concentrated on one very specific way of saying that to attack IRV, and I don't think it makes sense. If you can't predict ahead of time whether or not your vote will count, I'm saying it counts.
All of these are like saying that the soldier at the back of the phalanx that never got a chance to use his sword didn't help win the battle. That's simply not true. Since no one knew going in which soldiers would ultimately be needed, whether it would be a close battle, etc, having lots of soldiers on hand and ready was important.
All of this applies to your complaint that "your entire vote isn't counted." If it is there, ready to be counted if needed, that's good enough.
Especially when you know that other parts of your vote indeed are counted.
-
@jack-waugh Almost. Yes, under IRV, your first choice is always counted in the first round, but who cares about the first round? The final tally is the most important one that will get reported, and the final tally is the round that throws out the most ballots. As an example, in the 2021 New York City Democratic Mayoral Primary, Eric Adams was reported to have received 50.5% of the vote to Kathryn Garcia's 49.5%, but that's only because that tally ignored over 140,000 ballots. In reality, Adams only received 43% and Garcia 42%. That matters. Electability is rooted in perception. The voters of NYC were tricked into believing that Eric Adams had majority support when actually there's a clear majority that didn't vote for him. Under Choose-one Voting, that would have been much clearer. Under Choose-one Voting, I know that my vote will always send a message, even if it doesn't affect the outcome of the election.
-
@sass said in IRV complaint vs. FPTP: "your entire vote is not counted":
The final tally is the most important one that will get reported,
and
Under Choose-one Voting, I know that my vote will always send a message,
Are you worried that people -- i.e. the recipients of your message -- only look at the final round in RCV-Hare? Why are you assuming this?
One thing I've been working on is bar charts that can provide meaningful (but easy to take in) results information for both RCV-Hare and RCV-BTR (the latter being a condorcet method).
That said, those people who are recipients of your message are probably perfectly capable at looking at the data when presented in more complex ways (such as a Sankey diagram, a table of numbers showing each round, etc). I don't understand your implication that they simply ignore everything that didn't get to the final round.
Looking at the above bar chart, any reasonable person can see exactly the same kind of information you can see from a choose-one election. They can see, for instance, that Jane Kim actually did quite well, even though she got 0% in the final round (since she had already been eliminated).
And you can see this from the ways it is presented today. For instance, they can know that Mark Leno almost beat London Breed, as well as that Jane Kim almost beat Mark Leno. This is extremely easy-to-know information.
The difference is that choose-one is highly distorted by strategy and party nominations in ways that are less of an issue in better voting systems (including Hare, Condorcet, approval etc). The other thing is that it has a lot less information per ballot. So if you had preferred Jane Kim to Mark Leno, but voted for Leno under choose-one because you thought he had a better chance.... that information (i.e. that message) is completely lost. Not just because of where it was or wasn't considered in the tabulation, but you weren't even able to express it on your ballot.
Under any ranked or rated method, that information is available to anyone who is looking to know how the populace felt about the candidates, how close any candidate was to winning, and so on.
Anyway, I just don't see why you seem to say that the "message" your vote sends is ignored simply because you assume they are looking at the results through an extremely simplistic and naive perspective.
-
@rob How did you get the lengths for those bars?
-
@jack-waugh said in IRV complaint vs. FPTP: "your entire vote is not counted":
@rob How did you get the lengths for those bars?
I explained it in some detail in a previous post, although I have actually simplified it a bit after my initial approach.
The lengths (i.e. "scores") are calculated after the whole IRV logic has been performed. (although you'll need to play it out to the end, rather than stopping early when and if one candidate gets over 50%)
Start by giving the winner "maximum score", which can be any arbitrary number. I used 100. Then you go down the list, calculating the relative score of each candidate to the candidate just above them. So, for example, after giving London Breed a 100, then calculating Mark Leno's score relative to that, here is how you get Jane Kim's score:
-
It's like this, then, to restate it?
In each round, the effective "votes" for the round are the ballots where a given candidate is the top-ranked on that ballot, from among the candidates still in the running. The round is going to eliminate someone from further consideration and you are going to find also the one candidate from among the remaining candidates in the round who received the second-lowest count of "votes" in the round. You will take the ratio of "vote" counts of those two candidates and remember that as the ratio between the scores you will report for those candidates. Eventually this chain will lead to the winner, and you assign the score 100 to the winner and use the remembered ratios to calculate the rest of the scores. Since every score is a product of 100 with some numbers that are strictly greater than zero and less than or equal to one, every score will be positive.
