Nate Silver: Republicans set to gain between 17 and 78 seats

[Kurt_and_Hunter]

I am having a little fun at Nate's expense in the headline. His prediction is 47-48 seats but he acknowledges that the 30 seat MOE makes a wide range of results plausible.

_______________

October 10, 2010, 9:29 am
Number of Competitive House Races Doubles from Recent Years
By NATE SILVER

The House forecast that we released on Friday establishes an over-under line for Republican gains at a net of 47 or 48 seats. But, as I noted at the end of the article, the confidence interval on this forecast is very wide. Its margin of error is about ±30 seats — meaning that a gain of as few as 17 seats, or as many as 78, is entirely possible — and there is a small chance of even larger or smaller gains.

http://fivethirtyeight.blogs.nytimes.com/2010/10/10/number-of-competitive-house-races-doubles-from-recent-years/

[sharp_stick]

1. I'm surprised he didn't put it up

as between 17 to 78 in the title on his own, he seems to be the kind of guy that can poke fun at himself. I'll bet the copy editors at the NYT wouldn't let him do it.

[OHdem10]

2. My Gosh, what a spread. I could predict those odds.

17-78???

[DCBob]

3. That shows how very soft that 47-48 number really is.

Basically the headline should have been.. I have no idea what is going to happen except that the GOP will gain quite a few seats.

[oswaldactedalone]

4. If that doesn't motivate you...

to open up your wallet in these final weeks and contribute to GOTV efforts, to volunteer at your local headquarters, and to do whatever it takes to get our people to the polls, then I don't know what will.

The teabaggers have to be shut down and they have to be shut down now!!!

[grantcart]

5. Hard to argue with that. (between 17-78 lol)

Seriously I am surprised that Nate used his model for the total House outcome.

It doesn't seem to fit.

As I understand it the model he uses is based on using a significant number of polls over larger population samplings, like states.

The individual House seats don't get enough polling and the population sample is too small to really work with his system.

This becomes even more of a problem in the 2010 where there is no consensus whatsoever among pollsters on who is actually coming out to vote. In Nevada the Reid battle has polls showing a huge lead in registered voters and losing to Angle by 2 votes. Really? When was the last time that someone won a bitterly contested statewide election with only 32% of the electorate supporting them? Seems like a stretch to me.

[zipplewrath]

6. Standard deviation

Just eyeballing his house curve, it appears one standard deviation is between 190 seats for the pubs and 230, with 218 being the tipping point. That's a big variation here. What is missing is that there are far fewer positive outcomes for the republicans than there are for the democrats. Likewise on the Senate side. One standard deviation has way more chance for the dems to hold, than for the GOP to take over.

[rocktivity]

7. I was afraid this would happen

once the New York Times became his cyber-landlord.

:shrug:
rocktivity

[truebrit71]

8. So not a large spread then?

:evilgrin: