But I want to talk about it! So, I guess the Pit it will be.
The bracket organizer used the article Which 2020 Democrats Already Have A Fan Base — And Which Don’t on 538 for the seeding. In the article there is a table listing potential candidates, sorted by “net favorability difference”. This is the difference between a candidate’s actual and projected net favorability, based on polling data and a 538 model.
To me, this seems like a poor quantity to base seeding on. It’s a measure of how well a candidate is doing versus how well 538 thinks they should be doing. A better seeding metric would the actual net favorability polling itself or some other sort of polling data. So, from the same 538 table, we’d have
1. Joe Biden +69
2. Bernie Sanders +52
3. Elizabeth Warren +44
4. Kamala Harris +38
5. Cory Booker +33
6. Beto O’Rourke +32
7. Julian Castro +23
8. Kirsten Gillibrand +21
9. Eric Holder +21
10. Amy Klobuchar +15
11. Sherrod Brown +13
12. Michael Bloomberg +11
13. Terry McAuliffe +10
14. Pete Buttigieg +6
15. John Hickenlooper +6
16. John Delaney +5
– cut line –
17. Tulsi Gabbard +4
18. Jay Inslee +3 This seems like an obviously better seeding to me.
As the originator of the poll: I approve and am happy to join in the discussion.
When i conceived the idea, I started to look for a single source that would give both a comprehensive list of candidates, and a metric to rank them by. When i saw the 538 article I thought my work was done; but in hindsight, perhaps the net favorability was too odd a metric to go by.
Ideally, we’d work from a national poll of some sort, indicating favorability ratings or voting preference. But I don’t think such a poll exists yet.
Early polling is going to be quirky. I’m not sure how to get around that except waiting for better polling.
I’m thinking a World-Cup-style bracket would be fun in June, to go along with the Womens matches. Have 4 groups of 4, with every candidate up 1v1 against each other candidate in their group. Top 2 from each group go into single-elimination bracket. Use percentage-point differential as the tie-breaker.
I’d seed the groups thusly,
Group A: 1, 8, 9, 16.
Group B: 2, 7, 10, 15.
Group C: 3, 6, 11, 14.
Group D: 4, 5, 12, 13.
And then bracket as
A1 vs B2 _
D1 vs C2 /
B1 vs A2 _
C1 vs D2 /
There’d be clear favorites in each group, but a round robin can mix things up in unexpected ways. (X beats Y, Y beats Z, Z beats X, etc.) More chance for interesting match-ups.
It may be a hijack of this thread, but the cast of interested participants may be similar so …
For me, far more interesting than the March Madness tournament would be a** Karachi Auction Betting Game** (in Thread Games forum) where Dopers could win pretend money by guessing who the D nominee would be. Any interest?
Oh, I’ll likely have some sort of official contest come December or January for this. We’ll be close to the outcome by the end of March, given the way the primary calendar is loaded. Prize will be the same, some sort of custom thread title and bragging rights.
I admire your passion (which is a polite way of saying you’re crazy. )
A World Cup style round-robin bracketing is certainly a more thorough and fair way of getting to an outcome – but managing that in a series of thread polls would be REALLY unwieldy.
My only regret is that I do think that Bernie deserved better than a #4 seed. If I did it over, I’d override the 538 ranking and give him a #2 somewhere.
I think the ranking in the 538 article is more useful at this point in time. Mainly because it accounts for national renown. In general candidates who more people have heard of are more favorable but that just means they are loudmouths on the national scene. The 538 metric is basically predicting where people’s opinion will go once they know more about these candidates which is useful this early in the campaign since we will have serious candidates that haven’t had a Meet the Press appearance yet.