9 responses to “Pundarts!”

1. 

   Herb

   November 8, 2012 at 11:35 am

   Huh……I coulda swore I saw an earlier version that had Nate Silver somewhere else with an under in the double digits…..

   Maybe it was a different version altogether.
2. 

   Stormy Dragon

   November 8, 2012 at 11:52 am

   @Herb:

   It says in their update that they originally incorrectly listed Silver’s prediction as his “average of all outcomes”, and have since corrected it to his “most likely outcome”.
3. 

   Stormy Dragon

   November 8, 2012 at 11:57 am

   @Stormy Dragon:

   As an explanation of the difference, suppose I had a die with sides reading 1, 1, 1, 2, 3, 4. The average outcome of rolling it is 2 (if we roll it over and over, in the long term we expect the average value across all rolls is 2). The most likely outcome is 1 (if we roll it once, the mostly likely outcome is 1).
4. 

   Ben

   November 8, 2012 at 1:01 pm

   @Stormy Dragon:

   In other words, they originally listed the mean, and they’ve since updated it to the mode.
5. 

   James Joyner

   November 8, 2012 at 1:37 pm

   The OTB gang did pretty well here. Doug missed one state (FL), you and I missed two states (VA and FL) ; Dave missed three (VA, FL, and NH). Dodd brought the average down a mite, though.

   Really, the only thing really separating Nate Silver and the other stat gurus here was correctly calling Florida. The rest of us just relied on the raw polling averages and did quite well. Doug correctly saw the VA trend, which RCP picked up the next day; had I picked Monday instead of Saturday, I’d have given VA to Obama, too, since the RCP average had flipped.
6. 

   Stormy Dragon

   November 8, 2012 at 2:24 pm

   @James Joyner:

   If I may toot my own horns, my prediction that Ohio, Virginia, and Florida were all going the same way ended up being correct.
7. 

   James Joyner

   November 8, 2012 at 2:41 pm

   @Stormy Dragon: Yup, it sure looks that way. I thought Ohio was like 95% sure to go to Obama, flip-flopped on Virginia as the poll numbers moved, and thought Romney had a 75% chance of taking Florida. Some of the averages had Florida as a tie or leaning Obama, but most had it leaning Romney.
8. 

   David M

   November 8, 2012 at 3:11 pm

   @James Joyner:

   I missed FL as well, the movement towards Obama was pretty late, and as it actually turned out, most of the final polls underestimated Obama’s margin. Still I think that VA/FL were so close that getting those wrong shouldn’t really be held against anyone, assuming they still predicted an Obama win, which wasn’t in as much doubt.
9. 

   mattb

   November 8, 2012 at 10:31 pm

   Hell, I thought he’d take one of the three: Colorado

   I had assumed that FL and VA were just out of reach. Plus I didn’t want to jinx anything.