There's a problem you run into with any game wrapped around game or simulation concerns that isn't entirely compatible with some sorts of improvisational approach.
That is, the individual steps are assumed to matter. The various things Retreater had to do are things that may well be significant in their own right in other contexts, so you want to manage their resolution and time consumption. But in this particular case its really only the final result Retreater cared about, and all the steps made it an unattractive choice.
In addition, you sometimes don't want some declarations to be too easy, because the benefits they provide are stronger than the normal resolution.
This is why some more narrative-focused games are really only concerned about the final output--but may also not produce the degree of result a player expects.
So is it better in a game that a strong mechanics/simulation focus to:
A) Not allow improvisational actions on the part of the player
B) Apply what mechanics may be available but make the result sub par in order to bring them back to what the mechanics were designed for
C) Try to find a risk/reward ratio that gives the character the result they want for having to go through a number of mechanical hoops?
For my part, I try to do C, but am increasingly leaning towards just accepting A as the answer when these situations come up.