What's not clear on the player side is why sometimes things are resolved with a single check and other times they are resolved in an entirely different way. There's no proper reason for why the skill check mechanic was used in one instance and the skill challenge in the other.
To the best of my knowledge, the first system to toggle between simple and extended resolution, depending on mood, stakes, pacing etc, was Prince Valiant - in 1989, so 20 years before 4e D&D.
It's a feature of HeroWars (2000), Burning Wheel (revised is 2005) and then 4e. Probably other systems too that I'm just not familiar with.
And I don't think it's unclear in 4e. The DMG states the relevant principle (pp 72-3):
A skill challenge can serve as an encounter in and of itself, or it can be combined with monsters as part of a combat encounter. . . . Set the complexity based on how significant you want the challenge to be. If you expect it to carry the same weight as a combat encounter, a complexity of 5 makes sense. . . . For quicker, less significant challenges, or for challenges that work as part of a combat encounter, set the complexity lower.
Is This a Challenge?
It’s not a skill challenge every time you call for a skill check. When an obstacle takes only one roll to resolve, it’s not a challenge. One Diplomacy check to haggle with the merchant, one Athletics check to climb out of the pit trap, one Religion check to figure out whose sacred tome contains the parable—none of these constitutes a skill challenge. . . .
In other words, just like in other systems that toggle between simple and extended resolution, it's about pacing and stakes -
how significant is this thing,
I realise that's controversial for those who think it's the job of the system to tell the participants, without any need for decision, how significant something is. But by the time 4e was published it was not a new piece of game technology. This is also part of the reason that 4e is a useful vehicle for narrativist/"story now" play.
The fictional explanations for why an attack hasn't killed an orc yet are trivial, you missed, orc deflected, orc was barely grazed, etc. The fictional explanations involved in any kind of interesting skill challenge require a much greater amount of fictional 'change' - otherwise you just end up with the thief picking lock on the chest, honoring that and yet finding some reason the contents of the chest are still out of reach - most of which aren't going to be particularly interesting fictional reasons - The most obvious example would be a locked chest inside the chest- but that kind of on the fly complication is the antithesis to most D&D play - where an impartial referee is supposed to know the obstacle or be able to extrapolate details for it and then make rulings about what the players are doing to overcome it.
Clearly using skill challenges
isn't antithetical to the play of 4e D&D, given that they are one of its two fundamental action resolution components (the other being its combat rules).
If you're complaining that the DMG doesn't do as much as it might to advise GMs on how to engage in those fictional "explanations", I agree and have posted as much for over a decade. Here's a post of mine complaining about it, from January 2011 (I found it by searching for my earliest mention of the Rules Compendium, but it's quite likely I posted the complaint prior to the Rules Compendium reproducing the problem):
Skill challenges are mentioned several times in the PHB, but without much explanation. Even in Essentials, explaining skill challenges continues to be the weakest part of the rules. For example, it is pretty central to running a skill challenge that the GM be prepared to have ingame events unfold not according to ingame causal logic, but according to a metagame logic driven by (i) skill check results and (ii) narrative imperatives. But no where do the rulebooks mention it - the only place you can see the idea at work is in the example skill challenge in the Rules Compendium (because a Streetwise check fails in inspecting a building, the GM in the example has some toughs turn up to hassle the PCs, although the toughs weren't themselves implicated in the scene with the building) and you have to extract it by osmosis.
But as you can see there, the examples
model it even though they don't explain what it is that they're modelling.
I know some people insist that adjudication
must be by way of an impartial referee knowing the obstacle in advance, and then complain that skill challenges are broken because when you adjudicate them in that fashion they break down. I put these people in the same corner as those who turn up to play Rolemaster, insist that rolling to attack
must be done with a d20, and then complain that they always miss (on most RM combat charts, the first significant damage is found at results of 80+).