Posted 2012-November-02, 08:56
Somehow the question of who will be president for the next four years has taken a back seat to whether Nate Silver is a good forecaster. Probabilistic forecasting is difficult to evaluate, and surely neither Silver nor his model should be judged on the basis of one result. But of course Silver does now and has before predicted the results in a state by state way and that gives us far more data to look at.
There is the old joke that a weatherman who lives in an area that has rain on 30% of the days could simply predict a 30% chance of rain everyday and guess what, we would find that indeed it has rained on 30% of the days that he had said the chance of rain is 30%. Whatta guy.
It's easier to compare forecasters than to assign a specific rating to any one by himself. If A says that the chance of rain is 30% and B says that it is 50% you take the average, 40%. Then you give odds of 60 (no rain) to 40(rain) and have them each (happily I presume) place a bet at these odds, either money or rating points, You watch over a period of item and see who rakes in the cash.,
This could be applied if there was another person publishing odds on each of the states, and the results could be compared. It's more difficult to compare two prognosticators if one is posting odds on the outcome and another is publishing polling results. If not apples and oranges, it is at least winesaps and galas.
Anyway, I suppose Obama is going to win. If I were inclined to do some political work, it would be in support of democratic reps and sens. We really cannot afford four more years of paralysis.
Ken