Happy Birthday to Mal Reynolds himself, Nathan Fillion, 41. You know, Castle is reasonably good fun (although I haven't actually seen any of it this season, I don't think, so I have no idea whether it's slipped or not). Meanwhile, you all know about this, right?
I could go on, but instead I'll get to the good stuff:
1. I'm not going to link to everything about ACA and the Supremes, but I'll give you Jonathan Cohn's version of what conservatives have already won. Not sure I agree, but worth reading.
2. Dan Drezner is skeptical about the idea that Barack Obama would unveil a secret, radical foreign policy agenda if he serves a second term.
3. Greg Sargent on a GOP Senate candidate who has an innovative response to fact-checkers: just ignore them.
4. The House Democrats have a budget! Rosalind Helderman reports. I suppose I need to take a look at it to see if it's as phony as the GOP budget, although it's hard to get a lot of enthusiasm for reading through what the House minority wants. Maybe.
5. And Nerdwar is back. Nate Silver goes after academic general election prediction models; Andrew Gelman is first out of the box for the defense, or the other side, or whatever. More to come. I think what Ed Kilgore wrote here is pretty reasonable.
Kilgore may have been reasonable, but Silver is really just stirring the pot for no good reason here.
ReplyDeleteMy favorite model is Abramowitz's. That model is usually within 3 points, and it's very simple. It has both predictive and explanatory power. And yet, it's just casually treated as "one that isn't based on the fundamentals." Look at the models closely. You'll notice that, in the ones that include ANY kind of polling data (ie, all the ones Silver is only trashing by association and vaguely, as opposed to directly), the "fundamental" variables actually do a lot of heavy lifting. They do even more so once you step back and realize that polling results are, themselves, endogenous. They are CAUSED by the economy, war, etc.
Next, Silver is going to tell the congressional modelers that they need to look at polling more, because it's more accurate than the fundamentals of the economy and incumbent presence in races. This is similar to how I can look outside and tell you that it's sunny now, and then when I actually get outside, gee, what do you know, I was right! As opposed to making a prediction based onn weather models. Yes, the models are less accurate....but I didn't need to know what the weather was today to predict the weather today!