Skip to main content
15 events
when toggle format what by license comment
Jun 24, 2021 at 22:12 comment added uhoh @JJJ oh! I don't really follow (apparently in more ways than one) so never noticed it was there on each answer as well. tnx!
Jun 24, 2021 at 21:01 comment added JJJ @uhoh I think you need to follow answers individually if you want to receive notifications.
Jun 24, 2021 at 4:29 comment added eps @ChrisH you can also have every district tally up individually using a pseudo-candidate rank ordering, which quickly becomes absurd with more than a handful of candidates: rangevoting.org/IrvNonAdd
Jun 24, 2021 at 2:18 comment added JJJ @phoog you're right, edited a bit to be more specific about the NY implementation. In my original answer I focused on the way these votes can be counted by hand, as that seemed to be a significant part of the question.
Jun 24, 2021 at 2:14 history edited JJJ CC BY-SA 4.0
added 1286 characters in body
Jun 24, 2021 at 1:46 comment added phoog The counting could be done by hand, but that's not an answer to the question, which is "how will NYC" count the votes. The ballots are computer scannable, so the obvious intention is to have the computer count the votes, not people.
Jun 23, 2021 at 17:48 comment added endolith @Douglas Yes, C would be the rightful winner, but most people who vote for RCV referendums just want to rank the candidates and don't care or understand how the ballots are actually counted/eliminated.
Jun 23, 2021 at 17:46 comment added endolith @ChrisH Well is that what they actually do? Calling each district and collecting results iteratively over a long period of time? Other jurisdictions transport them all to a central location: "If no candidate wins a majority of votes on Election Night in the races that have three or more candidates, the ballots and memory devices from each municipality are securely transported to a central tabulation site in Augusta."
Jun 23, 2021 at 16:07 comment added Kevin @Douglas: IRV is not perfect. What you describe is a great example of IRV failing to satisfy the Condorcet criterion (anyone who would win every pairwise election should win the whole election).
Jun 23, 2021 at 16:00 comment added Douglas @Bobson Suppose there are two really popular but controversial candidates, A and B, who split most of the #1 votes between them, but voters for each of them really hate the other one. There is also a well regarded compromise candidate C that everybody likes, but in most cases not as their #1 pick. C takes an overwhelming majority of the #2 votes. C seems the obvious best option, with by far the broadest support if you check #2 votes, but C gets eliminated for having only 5% of the #1 votes, and you end up with a winner who half the voters like somewhat more than C but the other half hate.
Jun 23, 2021 at 15:30 comment added Bobson @MichaelRichardson - That could be a really interesting new question to ask. But if you vote for one of the top candidates, why would it matter who else you wanted if they didn't get the job?
Jun 23, 2021 at 15:26 comment added Michael Richardson I hadn't put extensive thought into it, but I don't think it had occurred to me that for ballots that end up with their #1 vote being for one of the top candidates, the lower choices are completely disregarded. Or is there a similar method that does give weight to lower picks?
Jun 23, 2021 at 11:58 comment added Chris H @endolith you don't need to take them all to a single location, you can carry out the count once per district and call in the counts to a central location. Then recount the ballots where the first choice was for the lowest-ranked, and continue. What you do have to do, of course, is wait for all districts to report their results so you know who's last at any given stage
Jun 23, 2021 at 11:54 comment added endolith Is this what they're actually doing in NYC? Transporting all ballots to the same location and physically counting them?
Jun 23, 2021 at 3:14 history answered JJJ CC BY-SA 4.0