It is intuitively attractive to think that it makes a difference in Newcomb’s problem whether or not the predictor is infallible, in the sense of being certainly actually correct. This paper argues that that view is irrational and manifests a well-documented cognitive illusion.

The recent operationalization of the famous Newcomb's game by Schmidt (1998) offers an interesting and thought-provoking look at the plausibility of backward causation in a Newtonian universe. Hereby we investigate two details of the Schmidt's scenario which may, at least in principle, invalidate his conclusion in two different domains: one dealing with the issue of Newtonian predictability in specific instance of human actions, and the other stemming from a possible strategy aimed at obviating the anthropically oriented view of backward causation (...) as applied to a judicial and ethical problem posed by a version of the scenario. We conclude that the scenario is at least to be more complex than originally presented in order to remain viable. However, it points to a very deep and delicate question of compatibility of backward causation with the conventional ethical standards. (shrink)