20

Nov

Teaser image to How Game-Theoretic Probability Makes Sense of Cournot's Principle

Colloquium

How Game-Theoretic Probability Makes Sense of Cournot's Principle

Glenn Shafer, Rutgers University

   20.11.2024

   4:00 pm - 6:00 pm

   LMU Department of Statistics and via zoom

Sufficiently high probability is practical certainty. This maxim, now called Cournot’s principle, was repeated by scholars for centuries before Jacob Bernoulli made probability numerical, and it has been essential to statistical inference ever since. Cournot taught us that it is the only way probability theory makes contact with phenomena. We test hypotheses by checking whether events of small probability happen. We make predictions using the law of large numbers and the central limit theorem. In the 20th century, when mathematicians decided that probabilities come in the form of complete probability measures, Counot’s principle was destabilized. How can it be reconciled with the “lottery paradox”, the obvious fact that an event of small probability always happens? Game- theoretic probability’s notion of a testing strategy rescues Cournot’s principle from the lottery paradox by showing that practical applications of Cournot’s principle are a form of inductive reasoning.

Organized by:

Department of Statistics LMU Munich


Related

Link to Additive Density-on-Scalar Regression in Bayes Hilbert Spaces with an Application to Gender Economics

Colloquium  •  15.01.2025  •  LMU Department of Statistics and via zoom

Additive Density-on-Scalar Regression in Bayes Hilbert Spaces with an Application to Gender Economics

15.01.25, 4-6 pm: LMU Statistics Colloquium with Sonja Greven (HU Berlin).


Link to Experimental Designs for A/B Testing in Marketplaces

AI Keynote Series  •  09.01.2025  •  Online via Zoom

Experimental Designs for A/B Testing in Marketplaces

09.01.25, 12-1:30 pm: AI Keynote Series with Chengchun Shi from London School of Economics and Political Science.


Link to Part 2: Learning, reasoning and optimisation: Adversarial robustness of neural networks

Munich AI Lectures  •  17.12.2024  •  Senatssaal, LMU Munich Geschwister-Scholl-Platz 1 Munich

Part 2: Learning, reasoning and optimisation: Adversarial robustness of neural networks

17.12.24, 5:30-7 pm: Munich AI Lectures with Holger Hoos from RWTH Aachen on neural network robustness.


Link to Part 1: Dynamics of Strategic Agents and Algorithms as PDEs

Munich AI Lectures  •  17.12.2024  •  Senatssaal, LMU Munich Geschwister-Scholl-Platz 1 Munich

Part 1: Dynamics of Strategic Agents and Algorithms as PDEs

17.12.24, 4-5:30 pm: Munich AI Lectures with Franca Hoffmann from Caltech on predicting complex dynamics in algorithm-population interactions.