Newcomb's Problem

Newcomb's Problem is a thought experiment in decision theory in which an agent chooses between taking one opaque box (whose contents were determined by a near-perfect predictor based on a prediction of the agent's choice) and taking both the opaque box and a transparent box containing a smaller, certain sum. The predictor has already placed a large sum in the opaque box if and only if it predicted the agent would take only that box. Causal Decision Theory prescribes taking both boxes, since the predictor's action is already fixed and the choice cannot causally affect the contents. Evidential Decision Theory prescribes taking one box, since doing so provides strong evidence that the predictor foresaw this choice and placed the large sum inside. The problem is not a puzzle about free will; it is a demonstration that the classical framework of Decision theory cannot resolve a choice when causal and evidential structure diverge. The deeper lesson is that rational choice requires a theory of how the agent's decision procedure is represented in the environment — a question that blurs the boundary between decision theory and systems theory.\n\n\n