We all love predictions, and Nate Silver has proven himself adept at making them, so it’s understandable attention turns his way on Election Day.
What is also clear is that we have a hard time understanding the nature of a prediction, which is why Silver not only says what he thinks is going to happen but also offers the odds that he’ll be wrong. To determine these odds Silver turns to Bayes’ Theorum and the more modern Bayesians, who have developed a way to measure uncertainty in a prediction based on the work of the English statistician and minister, Thomas Bayes (pictured).
This truth in packaging is what makes even Silver’s miscalls informational.
The mathematician Jordan Ellenburg takes a look at how many of Silver’s predictions will be wrong today in Slate, if Silver’s self-claimed odds of being wrong are correct. It won’t spoil the fun of reading the piece for me to tell you that Silver should be wrong about 2.5 senate races.