sorcerer can be in ignorance about the source of it's power, but it still should be there to a degree,
I agree with
@James Gasik in the next post as proficiencies(weapons, skills, armor, tools) and bonus spells should be added from 1st level.
Like Scout rogue,
Proficiency in Nature and Survival as 1st level then expertise at 3rd level, with reaction movement.
It's much smoother than going from knowing nothing(Jon Snow) to being an expert in the field overnight.
same way for sorcerer, even if we say that new players struggle with new features), it's better then to get 2 spell at 1st level, actually use them for few sessions, rather then getting piled on with 4 spells, 2 of which you will miss out most of the time.
I mean how much do you use burning hands at 3rd level?
We're trying to square a circle with compass and straightedge.
On the one hand, in the old way, going back to 3e, since that's when
à la carte multiclassing became a thing and still somewhat present in 5.0, we have the problem of frontloading, in two directions. First and more important long-term, we have "front-loaded features become dippable power" and other, similar balance-related issues. Such things will only get worse over time, especially with 5e's extreme reluctance to issue errata to core class features (that's why we didn't get PHB1 errata, we got a brand-new "revised but totally not revised but it's way better but it's exactly the same" edition). Remember, for example, how they chose to
weaken Storm Sorcerer, because people complained that it would power-creep Dragon and Chaos due to them not getting bonus spells but the playtest Storm Sorcerer did. The "lesson" they took from that wasn't "fix the Dragon and Chaos Sorcerers to have bonus spells", it was "nerf all future subclasses so they don't automatically outshine the original ones". Second and more important short-term, we have the "don't scare the new players with a crapton of features all at once" problem.
On the other, we have the comments that have just come up in this thread, and which are the big complaint with 5.5e vs 5.0: the two directions that
delayed features cause problems, narrative and gamist. Which, again I would classify as "short term" and "long term" respectively; narrative concerns about what does or doesn't make sense for two levels fade once you're 5 levels down the line, more or less, but making those early levels suck or forcing weird early problems will be a perennial issue.
All of this would be solved by ditching the new-with-5e "the first level or two are training wheels, experienced players should skip them in most cases" style that D&D has adopted, and instead replacing this with robust Novice Levels.
By "robust", I mean that they enable a wide variety of starting points and growth rates, and can be extended more or less uniformly to telescoping out higher levels as well. So a bare-minimum 0th level character has....nothing. No proficiencies, presumably all 8s in ability scores, no spells, no saving throws, zip, zero, nada. Maybe one (spoken-only) language, since that's relatively important. That's about the absolute bare minimum you can have and still have "a character" in D&D terms. Then, a truly robust "Novice Levels" system would allow you to parcel out the
pieces of becoming a 1st-level character, preferably with the ability to "try before you buy" as it were. So maybe you pick up divine spellcasting and get a single spell, but that could be the first step on the road to becoming a whole host of characters, not just specifically a Cleric--you could become a Paladin, or a Divine Soul Sorcerer, or a Celestial Warlock, or a Druid, or maybe even a Bard (if the spell is reasonably compatible with being one, e.g.
cure wounds), or a character who started with Magic Initiate: Cleric (or equivalents) but
actually studied arcane magic.
With a system like this, it would be a
snap to develop two kinds of "0th-level adventures": "intros", designed to get brand-new players up to speed as they
learn by doing rather than having to read through a zillion manual pages before they can even start
to start to play; and "grinders", designed to give old-school fans that brutal, deadly, no-holds-barred challenge they crave. (For some old-school fans, I presume all and only grinders would be
their preferred "intros" for new players, but most of them are aware that that isn't gonna fly for most players.)
Further, we'd have the ability to directly
support old-school players who want that feeling of very very slow progression, AND 3e-style fans who love slow, methodical,
organic growth of their characters over time. You could, for example, employ "Novice Level" rules as a form of downtime activity training. Spend X number of days (and appropriate training costs/materials) in downtime, get half proficiency (or half Training or whatever) in a new skill. Spend another X (+costs/mats), you get full proficiency. Or maybe it's X to get +1 bonus, and you need to repeat that enough times to match your current proficiency(/training/etc.) bonus--so training is harder for higher-level characters because they're trying to achieve greater mastery that WOULD have come from using the skill in the field, but now they must learn that through accelerated education instead.
It really is one of those extraordinarily rare design elements that manages to offer SOMETHING to nearly everybody. The one and only cost is that...well, designing such a thing is nontrivial. It would require the designers to think very carefully about how the system is built and how to make that work well. But, personally,
that's what I think designers should ALWAYS be doing when they design a system, so
to me that is just "expecting the designers to earn the money they expect us to pay for their work".