Josh Kraushaar: "Why I Don't Agree With Nate Silver"

[Never]

The National Journal's Josh Kraushaar wrote a thought-provoking article about why Nate Silver's FiveThirtyEight and Upshot are underestimating Republican chances to win Senate seats in some states, while they are overestimating the Democratic odds in a few Senate races.

Among other things, Kraushaar argues that Arkansas, New Hampshire, Virginia, and Iowa are more favorable to the Republicans than what the current polls suggest, while Michigan is going to end up being a relatively easy state for Democrats to retain a Senate seat in, especially in light of Terri Lynn Land not appearing to be as strong of a Republican Senate nominee as prior candidates the GOP has chosen in the past in this state.

Overall, I think this was a very objective article by Kraushaar, as it seemed to rightly point out that polls of Senate races fail to capture the individual strength of Senate candidates and how the political dynamics of a state relates to the success of those candidates. He also pointed out that there are some strong indicators that 2014 could be as good for the Republicans as 2010, and considering that the GOP picked up 6 seats in 2010, it might be plausible to claim that Republicans are favored to win the Senate, especially since the Senate map is tilted in their favor much more than last midterm. On the other hand, I had some qualms about his claim that Warner might not be heavily favored to win reelection to his Senate seat. I personally hold the opinion that Democrats have proven themselves very capable of negatively defining lesser-known opponents in the past and making the election about the record of the Republican rather than the issues, and an unknown like Gillespie in VA is a good target for this line of attack. 

[RogueBeaver]

I agree with 1 & 4, certainly not 2&3. Also in before non-blue avatars start posting Kraushaar's Romney wins Tweet.

[7,052,770]

Interesting. I was expecting an article with this title to be arguing for a Republican pickup of only 3 or 4 seats.

[publicunofficial]

Kraushaar is the rich man's Dick Morris.

I mean c'mon:

You must be logged in to read this quote.

Most hacks will use the "UNDER 50%? TOSS-UP" logic, but Kraushaar won't be happy until Merkley is pushing 60% apparently, in a state that re-elected their last Senator in a landslide in 2010.

I agree with 1 & 4, certainly not 2&3. Also in before non-blue avatars start posting Kraushaar's Romney wins Tweet.

Saying "inb4" doesn't make it not exist

It's hard to believe Kraushaar is a "Numbers Guy" when he had Romney taking Wisconsin, a state he never had the lead in at any point (Rasmussen had Obama/Romney tied, that was the best it ever got)

[7,052,770]

Wow, anyone who honestly thought Romney was going to win on Election Day (!!) should not be trusted to make predictions ever again...

[IceSpear]

When will Dick Morris weigh in?

[MurrayBannerman]

Gonna side with Nate Silver cuz the man is good at what he does.

[Mr. Morden]

Jonathan Bernstein critiques Kraushaar's argument here:

http://www.bloombergview.com/articles/2014-04-30/senate-forecasting-how-to-beat-nate-silver

Short version: There very likely is information that Silver's model doesn't use, such as information that helps one predict what the economic conditions are going to be in November, as opposed to what they are right now, that may in fact be used by an expert economist or expert political observer to improve on Silver's predictions.  But it's not clear that Kraushaar has actually done that.

You must be logged in to read this quote.

[ElectionsGuy]

When will Dick Morris weigh in?

I've actually seen a piece from Dick Morris on the Senate outlook (my grandparents wanted me to look at it to see if its biased). And yes, he thinks the GOP will pick up 11 seats and that Iowa and New Hampshire are "toss-ups", so he clearly hasn't learned from 2012 when the actual result mirrored what he predicted.

[free my dawg]

When will Dick Morris weigh in?

I've actually seen a piece from Dick Morris on the Senate outlook (my grandparents wanted me to look at it to see if its biased). And yes, he thinks the GOP will pick up 11 seats and that Iowa and New Hampshire are "toss-ups", so he clearly hasn't learned from 2012 when the actual result mirrored what he predicted.

oh my god

i'm reading it right now, this is gold

"It is also worth following are... New Hampshire, where Jeanne Shaheen manages only 50 percent of the vote against former Massachusetts Sen. Scott Brown’s 38."

actually laughed out loud.

[Nichlemn]

You must be logged in to read this quote.

Kraushaar talks about Nate's "model" as if it's algorithmic, when it's (almost) just numbers he's made up that seem right.

[The world will shine with light in our nightmare]

I agree with Nate Silver; I have learned over the last three election cycles (2008, 2010, 2012) not to take anything for granted. Silver has been wrong before however. He said Feingold had a better chance of winning 2010 than did Reid.

At what point in the cycle did he say that?  I thought Reid was toast the day after Angle was nominated, too.

[7,052,770]

I agree with Nate Silver; I have learned over the last three election cycles (2008, 2010, 2012) not to take anything for granted. Silver has been wrong before however. He said Feingold had a better chance of winning 2010 than did Reid.

Jesus Christ,  again??!?

Why can't people understand that Silver deals in probabilities,  not predictions. He is never "right." He is never "wrong."

You can multiply the numbers and find that he certainly didn't give a 0% chance of Reid winning and Feingold losing.

[Never]

I agree with Nate Silver; I have learned over the last three election cycles (2008, 2010, 2012) not to take anything for granted. Silver has been wrong before however. He said Feingold had a better chance of winning 2010 than did Reid.

At what point in the cycle did he say that?  I thought Reid was toast the day after Angle was nominated, too.

For a little while after her nomination, I thought that Angle could defeat Reid, but as the election neared, I thought it was very weird how she managed to hold a lead in the polls, considering how extereme she was. I thought that Sue Lowden was a much better option for the Republicans, and that if the polls were correct about Angle, Lowden would have had a double-digit lead.

By the day of the 2010 midterms, I thought that the Republicans would get to 48 Senate seats, since I felt Angle would end up losing, while Buck would win, along with the six GOP pickups that actually materialized. I didn't pay enough attention to the CO Senate race, and I failed to realize that Buck was just as extreme as Angle. Maybe if I had known that, I would have had a better prediction that year. Oh well. Since both Angle and Buck led their opponents in most polls in 2010 despite ultimately losing, those races showed me to be careful about using relying on polls to predict the victor of competitive contests.

[IceSpear]

When will Dick Morris weigh in?

I've actually seen a piece from Dick Morris on the Senate outlook (my grandparents wanted me to look at it to see if its biased). And yes, he thinks the GOP will pick up 11 seats and that Iowa and New Hampshire are "toss-ups", so he clearly hasn't learned from 2012 when the actual result mirrored what he predicted.

LOL!

What about the unskewed polls guy? Is he back in business?

But yeah, pundits need to be held accountable for their awful predictions and hackishness so we can remember not to take them seriously again. I know I'll never read anything Peggy Noonan ever writes, for instance. As for the guy in the OP, it's one thing to suggest Romney would win in a squeaker, it's quite another to suggest he'd carry every single competitive state, including many in which he never led in the polls during the entire campaign.

[badgate]

Badgate: "Why I Don't Agree With Josh Kraushaar About Why He Doesn't Agree With Nate Silver."

[Maxwell]

Michigan is more likely than any of the states he mentioned. What an ass.