Where's your god now?
No normal graphs have dependent variables on the X axis. Including, most importantly, actual learning curves in psychological literature.
Why would you ever flip a graph in a non-conventional way and then coin a term, when you could vastly less confusingly just coin the opposite term in the normal direction?
I think the graphs usually put "skill needed" in the Y axis, so a shallow curve means the game difficulty increases slowly.
Wat?
That doesn't make sense either. The game doesn't change itself to be harder when you play it longer (especially considering most new players will die repeatedly before getting far enough for any sieges at all)... Skill needed for DF is always the same and would be a flat line. So if that's what you mean, it in fact would have the shallowest POSSIBLE curve of anything (horizontal line)
Also, in general, that's not what anybody is talking about when they talk about steep learning curves colloquially anyway, since people use it to refer to things like real life skills, pole vaulting, bike riding, etc., which don't have levels.
You're forgetting to factor in total learning required.
If it takes you a hundred hours to get a good grasp of Dwarf Fortress, and an hour for Angry Birds, and Dwarf Fortress is a few hundred times more complicated, the learning curve for DF is a few times steeper.
No, that would just be this:
Angry brids being red. You just max out and there's nothing left to learn, but the learning portion is still a steeper curve.
(the slope is still shallower in DF due to poorer tutorials and UI etc. Even learning the first bit equivalent to angry birds takes longer than in angry birds)