Jason McCullough
06-10-2008, 03:07 PM
Got this from a TNR blog today; polls (http://www.fivethirtyeight.com/) weighted by past error distribution of the pollstere and with error-bounds included on outcome estimates to give a probability distribution - so you get "McCain 55% likely to win NH based on current polls", which is really interesting.
Then a whole ton of scenario outcomes and probabilities on ways to win the electoral college. Fun stuff.
Then a whole ton of scenario outcomes and probabilities on ways to win the electoral college. Fun stuff.