Newcomb's Problem in Game Theory

Publication Date

Spring 2026

Presentation Length

15 minutes

College

College of Sciences & Mathematics

Department

Math and Computer Science, Department of

Student Level

Undergraduate

Faculty Mentor

Andy Miller

Presentation Type

Talk/Oral

Summary

Newcomb's Problem, also known as Newcomb's Paradox, involves a decision between taking one or two boxes, where the contents of the second box depend on a near-perfect predictor’s prediction beforehand. It is easy to overcomplicate Newcomb's Problem just as it is easy to oversimplify it, even after noticing the ambiguity that lies in the words "near perfect predictor". The presentation will begin by showing how the classic Newcomb's Problem pits two solutions against each other, revealing how differing implicit assumptions can lead to opposing conclusions while still following rational game theory logic. While one stands on the concept of strategical dominance, the other stands on the concept of expected payoffs. Then, a twist that alters Newcomb's Problem into a combinatorial game will be explored using similar tools in game theory such as nimbers, game trees, and classifying positions. In the end, the presentation aims to show that Newcomb's Problem is more than just a difference in one's preference of philosophical perspective, and serves as a good example of the logic behind some basic concepts in game theory.

This document is currently not available here.

Share

COinS