Richard Jeffrey said that Newcomb’s Problem may be seen “as a rock on which... Bayesianism... must founder” and the problem has been almost universally conceived as reconciling the science-fictional features of the decision problem with a plausible causal analysis. Later, Jeffrey renounced his earlier position that accepted Newcomb problems as genuine decision problems, suggesting “Newcomb problems are like Escher’s famous staircase”. We may interpret this to mean that we know there can be no such thing, though we see no local (...) flaw in the puzzle. In this paper, I develop the critique of Slezak to show that Jeffrey’s analogy is apt for a puzzle whose logical features can be precisely articulated and I propose an easily realizable experiment that convincingly supports this analysis. I suggest that these real-life analogs of Newcomb’s Problem have diverted attention from the essential function of the science-fiction of a predicting demon. Following Eells, similarities with realistic problems such as Prisoner’s Dilemma and common cause cases have suggested that Newcomb’s Problem may be given a coherent description and realized in some way consistent with classical causation. However, the peculiarity of the apparent link between one’s choice and the previously determined contents of the second box is the central, defining feature of Newcomb’s Problem. I argue that the predictor may not be merely a fiction providing insufficient “extra information” as both McKay and Levi suggest. Rather, I suggest that the decision problem permits a precise specification revealing the source of the notorious perplexity. Notwithstanding efforts to rescue the coherence of Newcomb’s Problem and a classical causal account, a simulation of the decision problem establishes the correctness of my analysis and, thereby, hopefully ending the debate. (shrink)