FiveThirtyEight: Biden's chances to win the Primaries moved from 15% to 88% in less than a week!

[beastie boy]

https://projects.fivethirtyeight.com/2020-primary-forecast/

I am speechless...

[TwilightZone]

1. From the text, it sounds like the model is a little intentionally conservative.

So the odds are probably even higher. MI is going to tell us a lot.

[dubyadiprecession]

2. But, but, but....

Obama had nothing but great things to say about him Four years ago. I know this because I saw it in a Bernie approved campaign commercial.
Does this mean Bernie’s losing “The Establishment” vote, he is now seeking?

[musicblind]

11. Don't rub it in too hard.

Some of us like Bernie, too.

[empedocles]

3. Any guesses what Biden's chances will be next week?

[beastie boy]

4. With Biden projected to win 5 of 7 primaries on March 10, it's safe to assume

he is not going to drop out of the race.

[empedocles]

5. Guess at 538 numbers

[sweetloukillbot]

7. I think Washington and Idaho are the only two Sanders is project to win out of the rest

The model is real bullish towards Biden - more than I would be, I think.

[musicblind]

14. I don't think it is likely

But don't be shocked if Sanders squeaks out a surprise win in Michigan.

He's fighting for his life in that state and crazier things have happened this cycle.

[musicblind]

12. Biden's chance is significant.

Sanders wouldn't be moving so many resources to Michigan if it weren't.

That said, take nothing for granted.

[tandem5]

6. Let's be thankful that 538 isn't modeling hurricane paths!

[Awsi Dooger]

8. This is the reason I never wager on primaries

Likeminded people sense the same thing at the same time so they can switch on a dime. I've seen it countless times. The models and percentages make sense until it happens. Then zoom. I can't figure it out. It's like exhibition season in the NFL.

The mistake is to assume that general elections are even remotely similar. For example, when Lamont overcame the massive deficit to win the primary against Lieberman, there were assertions here and elsewhere that there would be no problem doing the same thing in the general election, once cowardly Lieberman announced he could not allow the primary result to stand.

Polling had Lieberman well ahead in the general election but nobody wanted to believe it, or that it would last. Keep in mind this was pre-Nate Silver. My life was much simpler pre-Nate Silver. I know darn well that if Nate had already been making political forecasts based on numbers then Lieberman would have properly been an overwhelming favorite. The percentage would have been 80+. The ideological numbers and independent numbers were going to thwart Lamont at every turn. The Republican was a farce candidate so no chance he would steal anything significant.

Since we were still in a fragile primitive subjective political wagering world, the odds on Lieberman opened at a comical 5-7 (140 to win 100) and briefly actually dropped to 5/6 (120 to win 100).

It was impossible not to laugh. I was posting the betting link on MyDD while describing it as free money on Lieberman. As so often is the case, what I want to happen politically is exactly the opposite of the proper way to wager. I cleaned up on that race. I'm not sure it isn't the single greatest political wager I have ever seen. It felt like zero chance to lose. The margin ended up being roughly 10 points, which is where it figured all along based on rigid general election realities.

[tandem5]

9. I think you can model a primary.

It's just that 538 did an extremely poor job. I mean consider the hurricane path projection analogy. A week out it could make landfall practically anywhere, the uncertainty cone is huge. As it gets closer and you get more information the cone narrows. That's how it should work here. 538 can't control the speed at which the cone narrows because it can't control when meaningful information comes in so it's not a failure on their part to create a model that keeps a wide cone over a seemingly long period. People seeking instant gratification may not like that, but you cant make something out of nothing -- It's arbitrary to say after three or four primaries that they had enough information to make specific probability claim when the information just wasn't there. They operated their model like a static band and not a narrowing cone. In other words they traded accuracy for precision and ended up with neither.

[Cha]

10. That's some kind of jump.. possibly

historic! Thanks, beastie

[Bucky]

13. I think that's way too optimistic.

What fool would think Bernie Sanders doesn't have a comeback or two in him? One of the great things about Sanders is he doesn't know what quit is. He just goes and goes and fights for a beautiful cause that he believes in down to his marrow.

This ain't over. Not by a long shot. I will vote for Statler or Waldorf, whoever we nominate. But don't any of you kid yourself that either of the two men left standing (from a field of two dozen candidates!) is a pushover.

[beastie boy]

16. Surely you dont mean the 538 algorythm is too optimistic.

Their prediction is a moving target. A week ago it showed Bernie miles ahead. What I was bringing attention to was the unprecedented turnaround Biden accomplished in less than a week. And frankly, I don't think Bernie can match it. That doesn't mean Bernie is finished though, not by any means.

[comradebillyboy]

15. An amazing turnaround. I had almost reconciled myself to

having BS as our nominee and was planning to tune out all politics until the general election. And then everything changed. What a relief.

[borgesian]

17. Fivethirtyeight predicted a super tuesday sweep for Sanders.

So I guess this is actually bad for Biden.

[TexasBushwhacker]

18. I just hope we have a clear winner before the convention

I don't want to go to a second ballot and involve the super delegates. I realize they serve a purpose, but being part of the "establishment" they will lean towards Biden and the Sanders supporters will erupt.