Secondly, the world changes: what works politically or economically in one decade might not work in another. A rational person may well have to change his ideas to say "well X would have worked, but in this new situation we need Y instead."[ Parent ]
There are examples of beliefs which turn out to be correct without needing any data at all. The classic example is the existence of atoms, which some greek philosophers believed in without ever being exposed to evidence for or against them.
Such a philosopher, if he had been exposed to evidence for atoms, would have seen his beliefs confirmed. The opposite, being exposed to evidence that the world is not composed of atoms, cannot occur of course. Therefore, such a philosopher could not have changed his mind based on data. What was wrong with him?
On your second point, where is the data which makes the rational person change his mind? If X would have putatively worked, but it was neither tested in time, nor is it relevant due to a changing world, then what grounds based on data are there to change one's mind to Y? -- $E(X_t|F_s) = X_s,\quad t > s$[ Parent ]
