One More for the Dems
| Wed Nov. 12, 2008 10:24 PM PST
ONE MORE FOR THE DEMS....It looks like Mark Begich is going to beat Ted Stevens in the Alaska senatorial race after all. Which begs the question: Is there anything that Nate Silver isn't right about?
Continues Below
Continued From Above
Comments
Submitted by Keith M Ellis (not verified) on Wed Nov. 12, 2008 10:16 PM PST.
Apparently not. He's a treasure.
Submitted by Nicholas Beaudrot (not verified) on Wed Nov. 12, 2008 10:33 PM PST.
I think his baseball projection system has been wrong on a certain number of players, and that a couple of other systems are now slightly more accurate than his.
Submitted by Yellow Dog (not verified) on Thu Nov. 13, 2008 5:21 AM PST.
Tim and Nancy: Double-ditto.
C'mon, Kevin - I know you know better. Now that you're 50, you need to start hitting the Gingko Biloba. Otherwise, before you know it you'll be typing it's for its.
Then we'll have to put you out of our misery.
Submitted by Winston Smith (not verified) on Thu Nov. 13, 2008 5:55 AM PST.
Yeah, good luck with that Tim and Nancy. I've tried to get Kevin to change his wicked ways re: "begs the question" like ten times.
Does he listen? No, he does not. I can't believe his indifference on this matter vital to our national security.
Well, I made a bet with myself that 3 comments, of 11, would focus on the "begging the question" usage, but everyone exceeded my expectations and the total was 7, or 133% more than I expected.
I am sorry to see begging the question lost because it is shorthand for a logical error that is not terribly easy to describe otherwise. "Circular reasoning" and stuff just don't work. Although you have to be familiar with the concept to understand the logical fallacy, however it is described.
So now I'm down 150%.
Submitted by washerdreyer (not verified) on Thu Nov. 13, 2008 9:04 AM PST.
Is there anything that Nate Silver isn't right about?
As of their respective last pre-election updates Pollster.com's projections of margin's of victory in each state were significantly better than his.
This is somewhat obscured by the fact that pollster never moved from their state margins to a prediction of electoral votes, instead classifying states as Strong Dem, Lean Dem, Tossup, Lean Rep, and Strong Rep, but if you just project every state by their margin you got 353, which is better than Nate's final projection.
Which is to say, Nate's great, but he shouldn't get credit for being better than he is.
Submitted by fumphis (not verified) on Thu Nov. 13, 2008 10:00 AM PST.
I believe Sam Wang at Princeton did a better job, since he predicted Indiana and maybe Omaha correctly. I like Nate a lot, but several less public analysts beat him out a little bit this time. However, I'm betting this will spur him to redouble his efforts and adapt his model, which should be exciting to see.
Wang predicted 364 total E.V., he didn't predict Indiana correctly. If you don't believe me, ask him. As far as I've seen, he didn't make available state by state margins of victory, so I haven't been able to look at how well he did at predicting them.
Good for Nate, and sorry, Sarah. It appears you'll have to watch Alex Ovechkin from your satellite TV instead of getting comfy comp Senate seats at Verizon Center. (Given your record of jinxing home teams this fall when dropping the first puck, as a Caps fan I say it's just as well.)
Submitted by low-tech cyclist (not verified) on Thu Nov. 13, 2008 1:58 PM PST.
This gives me one more opportunity to quote the mouse-over text from last Wednesday's xkcd: Someday I'll be rich enough to hire Nate Silver to help make all my life decisions. 'Should I sleep with her?' 'Well, I'm showing a 35% chance it will end badly.'
Submitted by Robert Waldmann (not verified) on Thu Nov. 13, 2008 6:51 PM PST.
well he did predict that McCain would win Indiana and Omaha.
The spooky thing was that on the 6th 2 days after election day, it was unclear whether Silver was going to get one wrong, because, after over a full day of counting, the ones he wasn't sure of weren't decided.
Would it count if Coleman wins in Minnesota where Silver said Franken had a 52% chance of winning ? Too close to call for Silver and *still* too close to call on November 13th.
Submitted by Sam Wang (not verified) on Sat Nov. 15, 2008 12:58 AM PST.
Hi - Sam Wang here. Ironically, despite the hype, Nate Silver does worse than a model in which no fancy assumptions are made. He was very entertaining and had the good fortune to make his splash in an easy year for poll analysis.
