Bayes Theorem in “The Signal and the Noise”
In a recent article by Trevor Parsons he references a book he has been reading called “The Signal and the Noise” by Nate Silver. He states that in the Book Silver looks at predictions across a number of areas and then examines how successful those predictions have been. Trevor summarizes Silvers approach as an attitude based on Bayes Theorem. Silver uses the theorem as the basis for incorporating probability, uncertainty, and testing into analysis.
I thought the article was interesting because it described some ways in which prediction can fall apart. One example is the difference between risk and uncertainty. Risk is a gamble with odds you can put a price on; however uncertainty is a risk that is hard to measure. This is definitely applicable to class as we calculated payouts using the probabilities found in Bayes Theorem. When these probabilities are know then decisions like in our blue and red marble example come up then there can be justified decision made. This becomes increasingly difficult in more complex models, and as Parsons states this is key to issues the financial industry has with prediction. People may make incorrect assumptions on the level of uncertainty or walk in with bias. This can be seen in some homework examples as when we find out information the question makes certain to state that the information is true.
I realized that we are working with simplified models in class that set a good foundation for some of the more complex algorithm that people use everyday in some occupations.
http://www.sys-con.com/node/3203895