• Security incident: ISF was recently accessed by intruders. Please change your password, and change it anywhere else you used it. Read more

If the Republican primaries had used Instant Runoff Voting

lpetrich

Muse
L
Joined
Feb 14, 2007
Messages
762
That is an alternative to first-past-the-post and delayed runoffs. One ranks the candidates by preference, and the ballots are counted by removing the candidate with the fewest highest preferences and then recounting with the preferences for that candidate being absent.

Let's look at this very simplified version:
4 people vote Cruz, Christie, Trump
3 people vote Christie, Cruz, Trump
3 people vote Trump, Cruz, Christie
2 people vote Trump, Christie, Cruz

The top-preference votes are Trump 5, Cruz 4, Christie 3. Under first-past-the-post rules, Trump would win, since he got the most votes. But let's see how it works under Instant Runoff Voting (IRV), a.k.a. Alternative Voting (AV).

Remove the candidate who got the least votes: Christie. The ballots become
7 people vote Cruz, Trump
5 people vote Trump, Cruz

So Cruz wins. This is because a majority of voters find Cruz OK even if not the best.

Ranked Choice Poll of GOP Voters Yields Insights - FairVote -- FairVote's name for IRV.
To address the plurality problem, we recently conducted a ranked choice poll in partnership with the College of William and Mary as well as Yougov, to demonstrate how ranking candidates provides voters with more meaningful choices and pollsters with more accurate information about the shape of the race. It allows us to simulate a national GOP primary using ranked choice voting (RCV).
Click to expand...
This was early this year. It had some interesting results.

Trump got 37% of the first choices, but only 10% of the second choices and 6% of the third choices. This decline slowly continued to about 2% or 3% until the last one, where it jumped up to 22%. So Republican voters either love him or hate him.

Ted Cruz won in second choices at 20% and Ben Carson in third choices at 16%.

The page also has an instant-runoff vote count on the poll. Here is who dropped out with how much of the vote at each stage:
0. Santorum 0.93%
1. Huckabee 2.39%
2. Fiorina 3.22%
3. Christie 3.53%
4. Kasich 4.05%
5. Paul 7.50%
6. Bush 8.66%
7. Carson 12.46%
8. Rubio 25.45%
9. Trump 49.32%
Winner: Cruz 50.68%

So Ted Cruz barely wins.
 
....which IMO is worse than Trump being the candidate.

The issue IMO isn't the method that was used by the GOP to select their candidate, it's that someone like Trump can garner as much support as he did. If his support had been 5% or 10% of the GOP then he would have gone down as an amusing Cheeto-coloured footnote in the GOP history.

Instead he managed to generate enough support and media attention to make him a viable candidate and the rest is history....
 
The Don said:
....which IMO is worse than Trump being the candidate.
Click to expand...

Maybe so, but you're not a Republican. Ted Cruz would have probably been more broadly acceptable to the GOP.

The best system would probably be range voting, which does a pretty good job satisfying preferences and reducing the effects of spoiler candidacies/gaming.
 
It would appear with that method that if you voted for the candidate with the lowest number of votes then your second and subsequent preference would no longer be taken into account because your ballot gets removed. Correct? If so, you have effectively disenfranchised those voters.

Instead, why not redistribute those ballots according to their second choice, and so on if more redistribution rounds are required?

Or have I missed something in the original explanation?
 
Last edited:
Norman Alexander said:
It would appear with that method that if you voted for the candidate with the lowest number of votes then your second and subsequent preference would no longer be taken into account because your ballot gets removed. Correct? If so, you have effectively disenfranchised those voters.

Instead, why not redistribute those ballots according to their second choice, and so on if more redistribution rounds are required?

Or have I missed something in the original explanation?
Click to expand...
In the OP, they were redistributed. In the first round, nobody had a majority of votes. So the lowest performer (Christie) was eliminated, and whoever voted for him then had their #2 vote used. In the case of the op, all the Christie voters had Cruz as their second option. So for irv, Cruz wins 7 to 5 whereas in fptp, Trump wins 5-4-3
 
TubbaBlubba said:
For what reason?
Click to expand...

Because it lets you vote for a candidate without having to worry about it essentially being a vote for the candidate you hate. Consider the 2000 election. You think Nader voters would have rather had Bush instead of Gore? If IRV or some other alternative vote system was in place, Nader voters could have comfortably voted for him and Gore as a second choice. So their real preference would be seen in the actual election results, but they wouldn't have essentially cast a ballot for Bush.

First past the post is a horrible way to vote. Instead of electing people the largest number of people would be ok with, you end electing people that large percentages of people HATE and even worse than that, sometime the majority of people hate the winner. If 60 percent of the population strongly agree with position A and there are two candidates that support A, then the vote can be split, meaning an anti-A candidate wins despite having minority support.
 
As a crude approximation, the recent Republican Presidential primaries were divided among supporters of Donald Trump and supporters of anyone but him. That put the "anyone but Trump" camp in a disadvantage, because that camp was rather badly split.


I'll now calculate something called a "Condorcet matrix", a matrix of which candidate beats which other, as calculated from the preferences. A sort of virtual round robin.

Rows: winner, columns: losers
Order in rows and columns: Christie, Cruz, Trump
0 5 7
7 0 7
5 5 0

Cruz > Christie
Cruz > Trump
Trump > Christie
Since Cruz beats all the others in this toy election, he is thus the winner according to the Condorcet criterionWP. So IRV and Condorcet agree here. The two also agree that Christie lost the biggest.
 
Another algorithm: the Borda count. For n candidates, each ballot's top candidate gets n, the next one gets n-1, and so on, with the last one getting 1.

A modified version is for partially filling out a ballot: n becomes the number of candidates actually ranked, even if some were left unranked.

Trying the Borda count on my toy election, I get Christie 24, Cruz 26, and Trump 22. Though Borda agrees with IRV and Condorcet on the winner, Borda disagrees with IRV and Condorcet about the biggest loser.
 
I checked the calculations of my previous two postings, and the Condorcet victories ought to be
Cruz > Christie
Cruz > Trump
Christie > Trump
Thus making an overall order of Cruz > Christie > Trump

However, my Borda-count calculation is correct, and both Borda and Condorcet agree on the order Cruz > Christie > Trump.
 
lpetrich said:
I checked the calculations of my previous two postings, and the Condorcet victories ought to be
Cruz > Christie
Cruz > Trump
Christie > Trump
Thus making an overall order of Cruz > Christie > Trump

However, my Borda-count calculation is correct, and both Borda and Condorcet agree on the order Cruz > Christie > Trump.
Click to expand...

It really doesn't matter what the calculations show because the rules were not in place to count in that fashion. If your theoretical "even playing field" has all candidates knowing that placing second or third is going to add up to a final victory, then they handle the campaigns and expenditures differently. They didn't have that knowledge going in, so were running in this primary in a Super First Past the Post.... You win the state - in many states - you get ALL the delegates. Jeb Bush famously didn't dedicate any portion of his considerable resources to Iowa, for instance. It was never in the cards for him to do well, there. No one spent a lot of money in Texas, 'cuz it was going to go Ted Cruz. If they had a system where second place and third place carried over to their advantage, they'd have all run different campaigns.

Notwithstanding whether or not the idea might be workable as a system, making these post-dictions based on this year's race is sort of like having a basketball game played by the rules everyone knows and then adjusting the score afterwards to award points for achievements that presently do not get you points. Say 1 point for a steal and 3 points for a blocked shot. Those teams might enjoy playing such a game; they'd like to know before it starts, though, that that's what they're playing.
 
Foolmewunz said:
...like having a ... game played by the rules everyone knows and then adjusting the score afterwards to award points for achievements that presently do not get you points.
Click to expand...

We shall call this approach Dumbledoring.

On a more serious note, while I like alternative voting systems for primaries, I am not sure they can overcome all disadvantages from have such a lopsided and irrational primary schedule. Why should Iowa and New Hampshire be major players and California be almost always left out?
 
Ladewig said:
We shall call this approach Dumbledoring.

On a more serious note, while I like alternative voting systems for primaries, I am not sure they can overcome all disadvantages from have such a lopsided and irrational primary schedule. Why should Iowa and New Hampshire be major players and California be almost always left out?
Click to expand...

I like "Dumbledoring". A lot of the Bernie Sanders supporters would go for that. "That's 3444 for House BillandHill to 1188 for Sanderbern...... And four million votes for Sanderbern for being so gosh-darn likable. Sanderbern wins the nomination."

The problem is that the states put themselves in these strange positions wanting more attention. If the parties hadn't stepped in, we'd have a spiral and a time warp, with states vying to be first to have their primary. Florida has its peak tourist season from about Dec. 19 to Jan 15. They'd probably love to fill all their hotels with reporters and campaign workers from Dec 1 to 18. The first caucus and primary, Iowa and NH respectively, are huge money-makers for their states. On direct expenses for the campaign and media buying Iowa got fifty million bucks from the Dems and GOP combined. That's a lot of money for professional politicos, newspapers, radio, tv.... and that's not including all the meals and hotels that the journalists and pundits spend.

The parties just let it happen. The Feds have nothing to say on the subject. It would probably require a constitutional amendment to get control of the situation, and I don't see the votes to do that. The parties like a system, however convoluted, they think they've figured out how to game. And when they can't game it, like the GOP, they change the rules.

The question, of course, is "Where to start?"
 
I would love to find an online poll where people could vote and compare different systems for this upcoming election . e.g. vote in a FPTP "one person one vote" then an IRV, and then an approval method, etc

Use the top three candidates in each primary, and include significant third party candidates like Stein and Johnson.


If I tried to set it up myself, i would
a) screw up the math somehow
b) not have a representative sample size since my friends tend to lean the same way.

Does the JREF polling allow ranking or only picking multiple options?
 
You must log in or register to reply here.

Share this page

ISF - Join now!

Every member here is approved by hand. No bots, no spam, just people who care about evidence and honest debate.

Membership is free!

Create your free account

Back
Top Bottom