Manbearcat
Legend
Thanks @Manbearcat, no pressure or worries about the timeframe of your replies.
Thanks. Bit of a moment so I'll try to get more in here.
If I'm understanding you clearly from your post, when you speak of adding tension/drama it is mostly because of the the subjective use of the skill DC.
I'm not sure I could put a percentage on it, but It is surely a part of it.
It perpetuates the focus being exclusively on the conflict-charged scene. The "on-screen action" will always contain threats and imminent danger of a certain difficulty whereby the antagonism will (a) never be relatively trivial nor (b) be wholly beyond the PCs' means. The design is metagame and outcome-based (* synched with PC build #s, usually a conflict resolution framework, and GMing techniques and principles), engineered to consistently deliver drama and climax. You don't want unclimbable mountains and you don't want doors that are trivially bypassed.
Contrast that with serial world exploration where the expectation for gameplay is that the "on-screen action" will invariably have the PCs sometimes encounter and interact with stuff that is relatively trivial or stuff that is wholly beyond their means....or spend a stretch of gameplay on conflict-neutral color. The design is based around the "exploration of a living, breathing world" premise...not to push play toward conflict and focus the experience exclusively on dramatic scenes. Now and again, you do want unclimbable mountains so PCs have to figure out other ways around (what spell to deploy, etc). Now and again, you do want doors that are trivially bypassed...or that at least look like doors. The door may be animate, a mimic, be trapped, or have an earworm (or whatever the creature was). Part of that sort of play is meant to sow paranoia in players and to test PC skill in exploration procedures.
In 5e, the higher one's level, the easier the easy/medium/hard tasks become to the point where the DM allows one to auto-succeed in tasks which lack 'dramatic pressure' so to speak (mechanically chance of failure is low). So 5e similar to 4e one enters the paragon/epic tier where simple tasks are immediately ignored and only when tasks appropriate to one's power level can challenge one does one generally roll to ensure success, so there is always that 'dramatic pressure' and in that way higher DCs are set which are line with the in-game fiction which as you say in more cases than not, IMO, are objectively set.
So whether one uses a scaling system like in 4e or 5e's bounded accuracy, the DM always makes one roll in times of uncertainty or when failure is a real possibility as you say so as to ensure dramatic pressure exists through the mechanics.
However, I do not agree with you that 'dramatic tension' is lessened because one uses the 5e mechanic instead of the 4e. There is no basis for that - to reiterate, as one rises in levels the challengers one faces are greater, stakes increase due to the in-game fiction and so the DCs for those remarkable tasks.
Example: the Jesters skill challenge of summoning Charon would not be performed by lower level PCs, the DCs in a skill challenge for such a task would always be at least 25+ therefore ensuring mechanics aid in the dramatic pressure.
I've said before (here, I believe, and elsewhere), that 5e's noncombat resolution mechanics and GMing techniques advocated for reminds me something of a mash-up of AD&D, 3.x, and 13th Age.
In AD&D and 3.x, the primary locus of play is "the adventuring day" within the framework of serial "exploration of a living, breathing world". In terms of system tech required and GMing impetus, it contrasts rather strongly with the primary locus of play being "the conflict-charged scene or situation."
One thing the latter really requires is (as outlined above) system math (protagonist build #s synched with antagonist DCs synched with the conflict resolution framework) that funnels play toward an outcome-paradigm which, in turn, produces the most drama or the most "something interesting happens" on the most consistent basis. I think the main problem I see in attempting to jury-rig 5e's objective DCs toward the 13th Age's subjective DCs (picking that system because its noncombat resolution has much more in common with it than 4e) is that it appears that the variance in baseline competency can be somewhat severe, skill to skill and PC to PC, as PCs gain in level. 13th Age has a few things working for it to keep PC competency relatively equilibrated:
1a) All classes have 8 (? don't have my books with me) Background points to spend and the pyramid expenditure system keeps them pretty closely together.
1b) The Background system is a very open-descriptor system, putting the player in a proactive position of narrative justification when deploying Backgrounds to the Ability Check system.
2) The way defenses work in the game pushes all PCs toward a MAD paradigm so disparity of bonuses, ability modifier to ability modifier, is fairly muted by proxy.
3) + 1/level across the board that somewhat attenuates any possible numerical outlier.
As such, competency is pretty bounded toward the baseline and the subjective (easy, medium, hard) DCs are pretty intuitive math-wise as the game was engineered around that paradigm.
Conversely, 5e has:
1) A proficiency system that protects niche, but in doing so it stretches characters across the competency base-line further and further as play progresses.
2) 5e very much rewards compounding your base competency by spending your ability score bonuses in your niche (eg Str for Fighters, Dex for Rogues, Int for Wizards), exacerbating the (presumably intentional) system feature of 1.
Given these things, I think deriving explicit baseline DCs from the objective DCs, that are pretty well bounded to general PC competency throughout the levels, is going to get more and more difficult as play progresses.
If, however, you could do that, I'd still have the same problems that I have with 13th Age's scheme (despite its subjective DCs). I think Jonathon Tweet and Rob Heinsoo are brilliant game designers and I think 13th Age is a great game. However, the noncombat resolution system suffers in the same way that I feel 5e's does (and I'm pretty sure the 5e devs just basically settled on going the 13th Age route in the end). It advocates for a coherent, narrative focus but the lack of clearly defined play procedures (just saying "fail forward" isn't good enough), resolution framework (that cements results in as success, success with complications, failure and/or stipulates campaign win/loss conditions) and attendant GMing principles to propel the game forward in interesting, non-GM-force-driven directions, absolutely hurts.
I will add this however, over the course of 4e's lifespan either through published material as well as the general 4e community the mechanics behind the skill challenge became more sophisticated/mature which I find a shame was not directly included within the 5e core books. The system would have certainly been richer with an optional inclusion of the complexity skill challenge mechanics, IMO.
What I think could have been interesting (could be?) for 5e is if they would have:
1) Grounded the proficiency and ability bonus inflation fairly tightly through the levels so base competency wouldn't stretch too terribly far apart at all.
2) Implement subjective DCs based on the above baseline competency.
3) Implement a unified conflict resolution framework based off the advantage/disadvantage mechanic.
4) Rewarded xp only for failures in noncombat conflict resolution.
On 3, I think it would have been cool to see something based off Tennis's "Deuce." Basically the track would look like this:
Win Disadvantage Start Advantage Lose
All conflicts would obviously begin at "Start". Stakes would be laid out at the inception of the conflict. The result of player action declarations would be narratively rendered based off of (a) intent with respect to the stakes of the conflict and (b) where they are currently on the track. Disadvantage next to Win and Advantage next to Lose would have the tendency to keep play in the "success w/ complications" and "fail forward" arena for an extended period. From an outcome-based perspective, those being the most interesting consequences of resolution and creating the most dramatic tension (rallying from the jaws of defeat and having a setback while on the cusp of victory) is precisely what you're looking for.
From an "in-world justification" perspective, you can easily just have the the mechanical setup working as the weird tendency for the world to push events toward a balance (perhaps the mojo that inexplicably keeps duece games going...and going...and going).
This is my little homebrew conflict resolution framework for 13th Age and it works pretty great. I think 5e would have potentially drawn in more gamers (of a similar disposition to me) if they would have provided such a robust noncombat conflict resolution module in the DMG (including robust GMing advice on techniques and the principles that underwrite their usage).