Stress testing incorporating causal dependency [7]

Go to: Summary | Previous | Next   
Bullet points include: E.g. suppose some years ago we were asked to estimate probability of Barack Obama being elected US President (before he was actually elected) How would we have estimated the probability? A purely frequentist approach might answer zero because no-one of his ethnic background had previously been elected US president Or maybe 1 in c. 300m (i.e. approximate population of USA) or 1 in X thousand (i.e. approx. size of political class) or 1 in 2 once he was Democratic candidate What ‘good’ ways exist to estimate the probability? And to allow for the probability to change through time?

Contents | Prev | Next | ERM Lecture Series

Desktop view | Switch to Mobile