It is interesting. It misses the important point though.<quoted text>
Interesting article about Nate.
The author stands corrected toward the end of the piece.
Nate didnt just call the states, he called the probability of the data being useful. That is how the bayesian models differ from the other models such as frequency models. Basically, he modeled certainty. That is an art and a math excercise.
Furthermore, ultimately, looking past certainties, what really mattered to the rest of us was the simple yes or no. I.E. Would Obama (or Romney Win)? To get there he had to aggregate and weight all the 50 states. Ultimately he called it more certain for Obama, and Obama did win. Thats all.
Therefore, unless you are prepared to really examine the accuracy of his certainties by testing his assumptions, as far as most people are concerned, Nate won.