Nate Silver is pretty famous right now for successfully predicting the winner of all 50 states in yesterday’s election (thus predicting on November 5th, at 91% confidence, that Obama would win). A summary of his book, The Signal and the Noise: Why Most Predictions Fail – But Some Don’t (again from Wikipedia):
Silver rejects much ideology taught with statistical method in colleges and universities today, specifically the ‘frequentist’ approach of Ronald Fisher, originator of many classical statistical tests and methods. The problem Silver finds is a belief in perfect experimental, survey, or other designs, when data often comes from a variety of sources and idealized modeling assumptions rarely hold true. Often such models reduce complex questions to overly simple ‘hypothesis tests’ using arbitrary ‘significance levels’ to ‘accept or reject’ a single parameter value. In contrast, the practical statistician first needs a sound understanding of how baseball, poker, elections or other uncertain processes work, what measures are reliable and which not, what scales of aggregation are useful, and then to utilize the statistical tool kit as well as possible. Silver believes in the need for extensive data sets, preferably collected over long periods of time, from which one can then use statistical techniques to incrementally change probabilities up or down relative to prior data. This ‘Bayesian’ approach is named for the 18th century minister Thomas Bayes who discovered a simple formula for updating probabilities using new data.
Sales of his book went up 850% after the election.
While amazing in and of itself (what’s more amazing is how people seem to think that BT is somehow inapplicable when dealing with uncertainty) it seems as though we all start out as natural Bayesians.