Barking at shovels

I'll settle for interesting. I'm not sure myself how well it would work -- I can see a couple of potential flaws, and it's not clear to me whether or not the average player would buy into the necessary mindset -- but I've been trying to figure out good scaling mechanics for a while. How does one design a game that scales from street level to Nobilis, with consistent mechanics?

The obvious advantage is, like you say, that every combat system has a sweet spot, where the probabilities for all your likely actions are sensible, and none of your stats get jammed up against the tail of bell curves -- GURPS combat, for instance, works sensibly with Parry and Dodge ratings below 12, say, but becomes increasingly tedious above that. But I'm not sure how satisfying it would feel to have character advancement consist primarily of buying new enemies. If what is 'really' happening is the character is getting more competent, it seems strange not to change the character's stats. For instance, how do you handle all the random thugs you may have to fight, who aren't specific Enemies? It seems like it might end up playing like a level system, where when you save up enough XP you can buy up your Studliness Quotient, and now all Average opponents count as scrubs. But that wouldn't be very satisfying for distinguishing between the individual strengths of PCs.

I don't know. There's something appealing about the idea -- maybe just of conscious recognition of Opponents Not Worth Fighting -- but I'm not sure how I would play it in practice.

The question of random thugs is good, yeah. I think this scheme almost requires you to abstract combat into scene-based combat a la Everway or other such games -- Trollbabe comes to mind again.

Sigh. I need to run a campaign using this model and see if it works.