Manbearcat
Legend
To me, the 4e table was a straight jacket, and serves primarily to make the DM's ruling predictable... which actions like those you mention above really shouldn't be.
<snip>
In my experience, it just added a whole level of, "Oh, I guess I have to look at this table and resolve everything the same way. Everything's got to be translated into a Power, I guess. Can't have things being unpredictable." It was too many rules for things that don't have them because they shouldn't.
See this is where things get really, really difficult to discuss because our preferences are literally diametrically opposed.
Here you actually opine that predictability in the resolution mechanics is damaging to gameplay generally and creativity specifically. I'm incapable of disagreeing more. As a GM, I want predictability for a few specific purposes:
1) When I'm composing challenges, I want to understand how each PC specifically (with respect to the mesh of their fictional archetype and [ii] the numbers that represent the tendency toward properly manifesting that protaganism during conflict) and the party generally will perform relative to that challenge. GM-side, this should be predictable so I can interpose the proper frequency and potency of antagonism between the PCs and their goals, thus making conflicts exciting, engaging and climactic. If this becomes unpredictable to me (the GM), then I lose climax or engagement or excitement (or perhaps all 3) and I may have to resort to player-agency subordinating GM force/illusionism to attain it. I NEVER EVER want to have to result to force or illusionism.
2) When the players are making action declarations on behalf of their PCs, I want them to have as much agency as possible, reflecting the OODA (observe, orient, decide, act) Loop that our brains undergo in real life. In the real world our brains perform an extraordinary number of high resolution subconscious computations. This creates a high level of confidence in the prospect of outcomes in most actions that we undertake. If the stakes are high (life or limb or loved one) and the margin of error of any particular course of action becomes something that is untenable, our OODA Loop pretty much drops that out of the "decide phase" of our loop.
Players need to have some (it doesn't have to be 100 % perfect, but the margin of error better be pretty ascertainable) analog to this in game, lest (i) their agency be short-shrifted and (ii) their field of potential action declarations NARROWS DRAMATICALLY. If you devise an opaque stunting system or an inconsistent stunting system (or both), then you should expect the field of potential action declarations to narrow accordingly (eg players do NOT stunt).
In my experience, players improvise when one player tries something unusual and it works. The framework doesn't make that happen, the DM does. When players see improvisation work, they start to expand their creativity. And what I want to reward is creativity, not encouraging the players to swing from every chandelier or topple every bookcase or carry around a mug of ale to cast into every thug's face.
In my experience (in life and in gaming), players include "improvise" amongst their field of potential action declarations when the prospective outcomes of resolution fall within a reasonable level of predictability/margin of error and the risk/reward is tenable.
I don't want "creativity" to become "find the stage prop to abuse." Or, worse, for the improvised actions to become the rote tactic. Predictable improvisation is... I mean, no, it's not particularly different than a Barbarian's rage or Rogue's sneak attack, but improvisation rules shouldn't encourage players to be creative once.
Why does
A) The GMing advice and action resolution mechanics that govern the stunting system are coherent, functional, and effectively reward "playing to archetype".
have to inextricably produce the dynamic of
B) creativity all of a sudden becomes "find a stage prop to abuse."
? If A then B? Is that the premise? Are you putting over before reward in A above and then implicitly asserting that there is no system paradigm that exists, or can exist, that does not adequately (neither over-reward nor under-reward) reward "playing to archetype." If so, I don't know what to say, because I've managed to run a lot of systems (not just 4e) with robust advice and resolution mechanics for stunting which adequately reward "playing to archetype."