The same could not be said of the reverse. At time T you only have two choices: (1) do A now, or (2) postpone A for later (i.e. hopefully T+1 if all goes according to plan). You cannot decide at time T to do A at time T-1, nor can you decide at time T+1 to do A at time T.

Furthermore, at time T you only "know" your value stack at time T. Your action A at time T reveals to the world that action A resides at the top of your value stack *then *(and only then). At time T you can only predict/estimate (but not *know*): (1) the state the universe at time T+1, and (2) your value stack at time T+1 affected by it. These predictions (estimates about the state of affairs in T+1) performed at time T affect your value stack at time T.