Don't sell Nate Silver short!

[Towlie]

Nate Silver is getting some criticism, but consider this:

It's currently 253 to 213 with Biden leading. Biden is favored in Nevada and Arizona for 6 + 11 = 17 votes, and there's a compelling argument that he'll win Pennsylvania for 20 more votes. If we assume that Trump will win North Carolina (15 votes) and Georgia (16 votes) it'll be Biden, 290 to 248.

Go to the fivethirtyeight.com map and put your mouse cursor on the right side of Joe Biden's chin (Joe's right, not yours).


Compare that little map to the one at CNN's website.

Also, Georgia is very close and could raise the win to 306 to 232. Amusingly, that's on the left side of Biden's chin!

So it seems to me that fivethirtyeight.com did a pretty good job!

[BusyBeingBest]

1. He can only work with the data and information he's given.

I don't fault him.

[BannonsLiver]

2. I think people saw the 89 and thought it was in the bag.

But if you throw 9 blue balls (forgive me) and one red ball into a bingo hopper, you’ve got some chance of the red ball coming up. About 1 in 10. So 89 is not 100.

[Drahthaardogs]

6. Polls said Biden was 10% ahead

No way he does that. Republicans turned out huge for their Orange leader

[BannonsLiver]

12. Biden still might end up +6 nationally when all is said and done

So that would be within the MOE of most polls showing Biden with a 9-10 point lead.

[a kennedy]

13. yah, all the closet repubs.......

[yardwork]

14. Nate Silver isn't a pollster. He analyses polls.

[Mosby]

23. He's a wannabe statistician.

Thats why he "studied" with Gelman at Columbia, all he has is a BA.

[a kennedy]

16. yah, all the closet repubs.......

[Stallion]

19. I Think 538.com ended up Biden +8.4

another factor that needs further evaluation is how many ballots were voided due the UPSO's failure to deliver ballots--538.com doesn't factor fraud. I wouldn't be surprised that when all votes are counted-especially from the West Coast, that the numbers are too far off from Biden +8.4. Its up to Biden +4 and it will end up around Biden +6. There are millions of votes from West Coast yet to be counted just like in 2016 and those will be strong pro-Biden

[bullimiami]

22. sorry. i think polls are right. have been. will be. they track voter intentions. voter suppression

and the other assorted tactics have been red-shifting our elections for many years.

if there is a flaw in the polling its the failure to account for the trickery.

the first time i really noted it was cleland/chambliss in ga.

tossing people off voter rolls, closing polling stations, making it easier for R districts to vote than D districts.
gerrymandering and the power it gives them to do this.
just some of whats in their trick bag.


when the results make no sense there is a reason.

[yardwork]

31. Ross Feingold in 2016. I'll never believe it.

[PubliusEnigma]

8. People are confusing certainty with speed.

[Mike 03]

17. +1

[Amishman]

9. And that 89 includes scenarios like the one playing out

89% to win doesn't mean 89% chance of a landslide win.

I like Nate. His is honest and does the best with the information he can get. We have a nasty way of shooting the messenger when they are telling us something we don't want to hear.

[TwilightZone]

3. On 11/1, he also wrote an article: "I'm Here To Remind You That Trump Can Still Win"

https://fivethirtyeight.com/features/im-here-to-remind-you-that-trump-can-still-win/

[OhZone]

4. I think it was mostly

the pollsters who were off.

Or the fact that people cannot be polled well any more.

There are so many scam calls on cells AND landlines, who wants to even answer the phone any more?

[edhopper]

5. He is not a pollster

he is a statistician, and the all polls he had to work with were terribly off.

[lagomorph777]

10. Or the suppression was terribly effective.

[yellowcanine]

7. Yeah I thought about that too. But the polling error (not Nate's fault, he just reports) was at

2016 levels.

[Amishman]

11. Even larger this time

The adjustments to account for education levels didn't fix it

[clutterbox1830]

15. I don't blame him at all.

He is not a pollster. He only list the probability from the data he's given.

In fact, the 538 blog is probably the most accurate and informative source of information with all of these close ballot counts.

[Towlie]

20. You're missing my point! I'm saying there's nothing to blame him FOR because he wasn't wrong.

[Imperialism Inc.]

18. If we are talking about Nate specifically then fine.

He said Biden could survive a 2016 level polling error, and even a slightly larger one. We got the larger one and he was barely right... probably. But for the polling industry this should be a huge embarrassment.

[LizBeth]

21. I think we will get Georgia too. And all the mail in held in post office is going to have an effect

we will find out later. It really has been a huge turn out. And bring to senate run offs, could be a blue wave. Didn't add to the house, we lost a few last I heard. But this works. Especially if we can tie the senate.

[meadowlander]

24. So he ran 40,000 simulations and one of the 100 most likely simulations is similar

to what is happening (but still not the same)?

Talk about hedging your bets.

[Bonn1997]

26. He's not betting; he's reporting probabilities

[meadowlander]

28. I appreciate that,

but the OP is framed as "look! Nate Silver got it pretty close to right!" which is significantly more impressive than "look! Nate Silver said here's 100 things that could happen and one of those things is pretty close to right!"

[Bonn1997]

30. Fair enough :-)

[Towlie]

29. What "bet"??? The point is that the presumed outcome is well within his probability distribution.

It's not like picking lottery numbers where you have to get the numbers just right.

[standingtall]

25. Could be that polling was off because the turnout was badly underestimated

The most I heard anybody say turnout would be before the election was that of the 1960's. Now I'm hearing this is the biggest turnout sense 1908. Generally true higher turnout usually favors Democrats and it did in this election too that's why Biden is going to win, but it's not as overwhelming as we like to think. We Democratic turnout increases significantly so does republican turnout and there are lots of right wing idiots that don't usually vote too.

[Bonn1997]

27. We should THANK pollsters when polls are off

How else would we know that we need to look into what happened?