Grasping Reality on TypePad, by Brad DeLong

Across the Wide Missouri: I would like Nate Silver and company to give an explanation for this:

I suspect that three things went on:

1. Around 8 PM EST the model took the behavior of white southerners as indicative of the behavior of whites elsewhere in the country—and that was a mistake, for white southerners really are a different ethnicity.

2. It really was a knife-edge situation: if the Democrats had only won the popular vote by 7 percentage points instead of 9, they would not now control the House.

3. The left-hand graph is miscalibrated: an 85% probability should not swing up to a 95% and then down to 40% before settling at 60% and then converging to 100% with the actual Democraic seat gain being equal to the original expected value...

#acrossthewidemissouri #politics #statistics