How should voting work?

[chewy]

I have trouble picturing a ballot that implements Borda's method that is less confusing than the condorcet method. How do you suggest people go about doing the ranking? If they have to write numbers 1-10 down next to the names there will be an awful lot of duplicate and skipped numbers. Furthermore, ranking things like that is *hard* it would take me an hour to figure out how I want to rank the candidates. The condorcet method is tedious, but you're faced with a simple decision each time.

What is wrong with skipping candidates or having gaps in the numbers? There is no way a person could keep track of all the 3rd party candidates' positions anyway. Anyone who isn't "ranked" on a given ballot is given 0 points.

On a related note, what about using Borda for determining party candidates at the primaries? I'm not saying there should be a law forcing the issue (each party is entitled to chosing its own method), but I would have loved it for the 2000 primaries. There were several interesting candidates in both major parties. My problem is the guys I like the most in any given party never make it past the primaries.


cheers,

steve

[Bill Glasheen]

Justin

the SSD form of condorcet...

Precisely my point earlier. Condorcet doesn't work without modification.

Last night as I was cutting my grass, I came up with a very loose mathematical proof that Borda is a subset of the Condorcet method. Bottom line is that there's nothing you can do with Borda that you can't also do with Condorcet. So Borda isn't inherently flawed if you can find a way to modify Condorcet to make it work.

This is the same issue with Plurality voting, Justin. What we have isn't a pure Plurality voting system. Rather we have electoral college voting overlayed on the back end of Plurality. These hybrid methods tend to resolve most if not all of the practical problems in any voting method.

I can write the loose proof out if you wish... But I don't think it's necessary if you see my point already.

- Bill

[Valkenar]

I can write the loose proof out if you wish... But I don't think it's necessary if you see my point already.

No, that's all right, I do have a sense for why a ranked list is equivalent to a comprehensive pairwise matching. My (possibly incorrect) understanding was that the SSD method doesn't have a corollary in the Borda scheme because it essentially grabbing the numbers before they're condensed down to a ranked list. But I suppose it should be possible to go bawkwards from the ranked list, if the systems are really equivalent. And I'll accept that they are, though they do require either SSD or the equivalent to be at their best.

Can you think of a strategy-reducing methodology that might be applied to Borda? I haven't thought about it much so far, but nothing is immediately suggesting itself.

Personally, I would still favor Condorcet over Borda because the latter almost requires electronic voting, which isn't ready yet (and who knows when it will be), whereas condorcet could be implemented now.

Incidentally, the cyclic resolution system is a fundamental component of any Condorcet method, it's not just a modification.

Chewy:
The problem with gaps and repeats is that it really presents a different system and undermines the mathematics of the voting system. Basically allowing gaps and repeats turns it into a system where you rate each candidate from 1 to 10 (or 1 to 15 or however many candidates there are) instead of just ordering them. It's possible that there's a fine system based on that, or that it actually works as-is converted to a rating-based system, but I'm not sure there aren't properties of the system that rely on the fact that it's an ordered ranking. Could be wrong though. In fact, the more I think about it, the more it seems like I am.

As for the primaries I'd say sure. Really, any time you're taking a vote on more than two things, it's better to use condorcet, borda, approval or almost any other system than just a plain vote.

[Bill Glasheen]

Personally, I would still favor Condorcet over Borda because the latter almost requires electronic voting

Agreed.

Can you think of a strategy-reducing methodology that might be applied to Borda?

How about Instant Run-off Voting overlayed on top of Borda? (See description I posted on page 1 of thread.) An electoral college system overlayed on top of Borda?

I like Borda because it prevents paradoxical voting, and everything is right there in front of you in clear display before you "pull the lever." It seems to work really well for the Coaches Poll in the NCAA, where conferences could trash rival conferences so they end up better off at bowl time. The good teams still end up on top (more or less) with the lesser teams in lesser tiers.

And I believe Borda would result in more credit given to "third party candidates." Our politicians need to see and acknowledge minority opinions. Views are much, much more complex than can be expressed with a dichotomous choice.

Borda would have been a very practical method for the recent California recall vote for governor. The Condorcet ballot would have been prohibitively too long. The Governator still would have won, and there wouldn't have been as much of a risk for a lesser individual squeaking by when two good candidates were on the ballot.

And I really would have liked to see where the porn star ended up with respect to the competition...

- Bill