Aidan Lyon (MCMP/LMU) gives a talk at the MCMP Colloquium (6 November, 2013) titled "Measuring Overconfidence with Imprecise Probabilities & The Wisdom of Collective Credences". Abstract: What explains the Wisdom of Crowds effect? Page (2008) has made some initial progress on this question with what he calls the Diversity Prediction Theorem. The upshot of the theorem is that if a collective has some diversity to it, then the collective’s estimate is guaranteed to be more accurate then the typical estimate in the collective. This appears to be a simple and very general account of the Wisdom of Crowds effect. However, for the theorem to have explanatory power, it needs to be supplemented with additional assumptions. In this paper, I analyse these assumptions, and discuss their drawbacks, and how we might overcome them. A consequence of this analysis is that the Wisdom of Crowds effect behaves very differently for probability estimates than it does for regular quantity estimates.