Murder at the Asylum
Tanya Khovanova
TL;DR
The paper presents a Smullyan-inspired logic puzzle that integrates three axes of agent type—sanity (sane/partial/delusional), truthfulness (truth-teller/liar/alternator), and guilt (innocent/accomplice/guilty)—into a round-based interrogation on nine suspects. It develops a four-question scheme to distinguish types, then shows how a three-valued ($0$,$1$,$2$) extraction encodes the solution as a word, demonstrated by deriving ALTERNATE. The contribution lies in combining advanced belief dynamics with a structured extraction method to craft and solve a multi-layered mystery, illustrating novel puzzle design techniques for the MIT Mystery Hunt and expanding Smullyan’s tradition of islands of truth and deception.
Abstract
I describe a puzzle I wrote for the 2018 MIT Mystery Hunt which introduced new types of people in logic puzzles. I discuss the puzzle itself, the solution, and the mathematics behind it.