Speaker:Prof David Spiegelhalter, FRS (Statslab)
Venue: Winstanley Lecture Theatre, Trinity College
Time: 15/02/2010 20:30, drinks from 20:15
A Bayesian perspective allows probability theory to be used as a formalism for epistemic uncertainty: ie a measure of our confidence in a current or past state of the world about which we are ignorant (although someone else might know the truth). This allows probability to be used to quantify our uncertainty about a suspect’s guilt, the image on the Turin shroud, the average effect of a medical treatment, or whether Jane Austen died before Napoleon. I will look at how these ideas have developed and are being used in practice, and how scoring rules can be used to assess how well you can quantify your doubt. A test will be given at the end, and a small prize awarded for the person who best knows what they don’t know.