Expected Utility Maximization directs us to weigh the value associated with each possible outcome by its probability, yielding an overall expected value. This essay explains Expected Utility Maximization as a decision theory and defends its most distinctive feature: that it can advise us to choose options that will predictably lose.
This essay was adapted (with permission) from two of Joe Carlsmith's essays on expected utility theory—"Skyscrapers and madmen" and "Why it can be OK to predictably lose"—and edited and rewritten for utilitarianism.net by Vikram Balasubramanian.
---
Outline:
(00:31) On Expected Utility
(03:30) Maximal Skyscraper
(05:59) Only One Shot
(09:37) Why it is OK to predictably lose
(10:28) Conditional Probabilities
(15:29) Nothing special about getting saved by heads versus tails
(18:54) What would everyone prefer?
(20:24) Taking Responsibility
(24:16) About the Authors
The original text contained 7 footnotes which were omitted from this narration.
---