Ezra Klein

How do you know she's the same person? Have you ever seen Rachel Maddow with and without her glasses at the same time?

Perhaps I've said too much.

Comic books.

Terrified. Though I was pretty careful in the column. The point of this model, as I write, is to help us think more clearly about the election, and give us a guide for how it'll work if it follows past patterns. Nate's critique is heavily about pure-fundamentals models that make ad-hoc adjustments to achieve a better fit. We're not pure fundamentals and we don't make ad hoc adjustments. He also makes the point that these models sometimes perform worse out of sample, which I mention in the piece. But when we run an out of sample test on our model, it actually performs better, with a lower average error.

Soon! And Annie does nothing irritating at home. She is an angel and I adore her.

Oh, except for that one thing. She knows what I'm talking about.

I'm going to cheat and answer a few easy questions early because I have a moment. So! When I moved here, all the young wonks -- by which I mean, young political folks having trouble paying rent -- lived in Columbia Heights. But that got expensive. There's even a Target. Now it's Bloomingdale, which is where I live.

In terms of places to hang out, Big Bear Cafe. Boundary Stone. Red Derby (which is up in Columbia heights, I should say). I'm a huge partisan of Duffy's on 9th and V, which is around the corner from the much cooler American Ice Co.

Not sure how simple it will be.

Excited about? Technological change. The mass elevation of hundreds of millions and maybe even billions of people to the point where they can be producing innovations that make our lives better.

Concerned? Well, everyone worries about deficits. But for reasons I explain here, they're not my primary concern:

http://www.washingtonpost.com/blogs/ezra-klein/post/why-i-dont-stress-about-deficits/2012/04/06/gIQAvVEczS_blog.html

I worry a lot about a broken political system heavily biased towards inaction. ANd under that model, what scares me are issues where you need lots of congressional action and energy to do anything. Global warming is the obvious one -- and even more difficult due to its international component -- but pandemic flus, loose nukes, etc, all keep me up, if only because they're the sort of questions that don't have much of a constituency and thus can easily be ignored until something goes terribly, catastrophically wrong.

Sorry -- I'd forgotten to go back and give a non proof-of-existence answer.

The model has a 2.8 percent error rate at predicting the vote share on elections since 1948. If you run out of sample tests -- that is, take an election out, then run the model against it fresh -- the error is even lower. It's not, as someone suggested in this thread, only looking at second terms. In the column I've linked to, I go through the major missed calls. It should be predictive this year -- the question is, as I say in the piece, whether you believe this year is, for some reason or another, different.

I only eat soup that comes in cartons.

Done: http://twitter.com/ezraklein/status/194802290484449281

I do. There's another book, by Phillip Tetlock, I think, that runs a fairly rigorous test on political prognosticators, and finds that we suck. I agree entirely, in large part because I think Washington focuses on the absolute wrong things. in my view, politics is still waiting for its Moneyball revolution.