The Department of Finance & Risk Engineering welcomes Dr. Glenn Shafer, Professor at Rutgers University, to present his talk on "Let’s Replace P-Values with Betting Outcomes!" as part of the Brooklyn Quant Experience (BQE) Lecture Series*.
How can we test the constantly fluctuating probabilities that Nate Silver offers for the outcomes of elections and sporting events? The natural (and perhaps only) way is to interpret Silver’s probabilities as betting offers and to bet against him. He fails our test if we multiply our money by a large factor. We can test a statistical hypothesis, as well as the efficiency of a financial market, in the same way. In the case of statistical hypotheses, this leads to a new understanding of likelihood ratios and to an alternative to the notion of power. See Working Paper 54 at www.probabilityandfinance.com and Game-Theoretic Foundations for Probability and Finance (Glenn Shafer and Vladimir Vovk, Wiley, 2019).
Glenn Shafer is best known for his work on the Dempster-Shafer theory of belief functions, especially his 1976 book A Mathematical Theory of Evidence. Beginning in the 1980s, Glenn has studied the mathematical, philosophical, and historical foundations of standard probability theory and on its limitations as a theory of evidence. Since the 1990s, he has collaborated with Vladimir Vovk on understanding the benefits of using betting games (as opposed to measure theory) as a mathematical foundation for the standard theory. Glenn began his career as an educator by teaching geometry in Afghanistan in 1968; he subsequently taught at Princeton, the University of Kansas, and Rutgers. From January 2011 to December 2014, he served as dean of the Rutgers Business School.
*The BQE Lecture series is held every Thursday at 6 p.m.
Attendance is free and highly encouraged. When possible, slides will be made available after the talk.