Hacker News new | past | comments | ask | show | jobs | submit login

Aw man, when first I read the actual algorithm for IRV, my first thought was "This has got to be the worst way of managing ranked-ballots that preserves some semblance of enforcing voter will.

[edit]

Okay it's not actually that bad, but it's poor, particularly since the mid-90s we've had the technology to cheaply run as many simulated elections as we want given machine readable ranked-ballots.




I would be highly in favor of a state adopting ranked ballots without changing their counting method - just continue to count the top choice as your vote, like before. But then we'd have some amazing data sets. That's the dream.


Wouldn't this nullify the primary advantage of range ballots, if I understand you correctly? Namely, you'd still have to rank one of the two main candidates as first so that the other doesn't win.


Yes, there wouldn't be any additional benefit in terms of vote-counting at first. Our current "vote for one candidate" approach is the same as picking your top choice in a ranked ballot.

But the benefit would be that we'd actually have data on being able to say, "Here's who would have won if we had used IRV, here's who would have won if we had used Condorcet", etc. And then later pick the actual counting method that proves effective.

I mostly like the idea is because I think the debate over counting method gets in the way of the debate over using ranked ballots.

A risk would be someone picking a different first-choice candidate than they otherwise would, out of improperly believing that their preferences would be counted.


That risk is sorta what I was bringing up: if it only counts the top person then people have to vote strategically and hence will not be voting in the same manner as in IRV so we still wouldn't know who would have won by IRV.


It would be best if they picked their plurality vote, and then could optionally rank candidates; we could then tally up how many people vote for their non-favorite as a big stick to wield.


If preferences beyond the first don't mean anything to election results, the data sets produced will be garbage because people won't spend any thought on them for the most part.


Yeah, I think you're right - your comment and aidenn0's below have convinced me you pretty much have to pick the counting method at the same time as introducing ranked ballots. Darn, because the IRV people have a clear marketing advantage at this point.




Join us for AI Startup School this June 16-17 in San Francisco!

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: