About the speaker: Persi Diaconis is an American mathematician of Greek descent and former professional magician. He is the Mary V. Sunseri Professor of Statistics and Mathematics at Stanford University.
He is particularly known for tackling mathematical problems involving randomness and randomization, such as coin flipping and shuffling playing cards.
Abstract:Â ‘Everybody knows’ that a vigorously flipped coin is equally likely to come up heads or tails AND that different flips are independent. But is it true? I will report work with Susan Holmes and Richard Montgommery showing that vigorously flipped coins are slightly biased (it’s about .51) to come up on the same side they started. The math involves classical mechanics and ‘the method of arbitrary functions as well as image analysis to look at what real coins do when flipped by real people. Similar analysis applies to most of our basic notions of random phenomena; rolling dice or roulette balls. This thinking hard about the underlying sources of randomness sheds some light on the variety of ‘made up models’ prevalent in much of applied mathematics (in particular, my field of statistics).
Further details are available on the website