You can only know your “present” value stack, and you disclose to the outside world that action A resides at its top by perfoming it then. (i.e. in the “present”).
You are missing the point. People “prefer” performing action A now (vs. not-now, i.e. later) exactly because it is then that it resides at the top of their value stack, by definition. The “present” value stack is special that way, when compared to the (unknown) value stacks at any point in the future (be it “sooner”, or “later”).
But how can anything general be derived from this? When I prefer to do A now, I also prefer to not do B now. Can we infer that I always prefer to do A sooner rather than later? Certainly not. We can only infer that I prefer to do A now. In the next moment, I may prefer to do B. If all we are considering is what I prefer now, then “time preference” isn’t present in the analysis at all.
But Mises says: “Satisfaction of a want in the nearer future is, other things being equal, preferred to that in the farther distant future.” Now given what you said, that we can only know what we prefer now, then how can we reach Mises conclusion that we prefer sooner satisfaction to later satisfaction? If we don’t know what we will want in the future, then how is it possible to prefer something sooner to something later? If we only know our present value stack, then how can we conclude anything about our future preferences a priori?