Sneaky Angel and Devil Game
Cameron Bernstein, Logan Crone, Alexandria Kwon, Sydney Le

TL;DR
This paper introduces an imperfect information variant of the Angel and Devil game, proving that the Devil still has a winning strategy, thus extending previous results to a more complex game setting.
Contribution
It proves that the Devil wins in the imperfect information variant, generalizing the original game's known outcome to a new, more complex version.
Findings
Devil has a winning strategy in the imperfect information game
Generalizes previous results from the original game
Extends understanding of game strategies under imperfect information
Abstract
In this paper we introduce an imperfect information variant of the Angel and Devil game, which was first introduced in 1982 by Berlekamp, Conway, and Guy. The Devil player has a winning strategy in this game, but the main problem they pose is whether this changes if the Angel player is allowed two moves for every one of the Devil. This and many other variants of the game have been considered, and the original question was only solved in 2007, when it was shown that the so-called power 2 Angel could win. Our main theorem, Theorem 4, is that the Devil player wins the imperfect information variant of the game. This generalizes the result that the Devil player has a winning strategy in the original Angel and Devil game.
Peer Reviews
No public reviews on file for this paper yet. If you reviewed it on a platform where reviews are public (OpenReview, ICLR, NeurIPS, ICML), you can paste yours below so the community can read it here.
Videos
No videos yet. Explain this paper in a talk, walkthrough, or lecture? Add one.
Taxonomy
TopicsArtificial Intelligence in Games · Digital Games and Media · Gambling Behavior and Treatments
