Measure changes with extinction (Q1012118): Difference between revisions

The authors prove that the inverse of the density of a change of probability measure is only a supermartingale (and not a martingale, as some earlier papers claim...)