Isn’t the claim in 6 (that there is a planning-optimal choice, but no action-optimal choice) inconsistent with 4 (a choice that is planning optimal is also action optimal)?
4 is true if randomized strategies are allowed?
Yes, that’s right. I’ve made an edit to clarify that.
Isn’t the claim in 6 (that there is a planning-optimal choice, but no action-optimal choice) inconsistent with 4 (a choice that is planning optimal is also action optimal)?
4 is true if randomized strategies are allowed?
Yes, that’s right. I’ve made an edit to clarify that.