I've got a theory it can be narrowed down to two points of design aesthetic that 4e didn't account for:
- Mundane prowess must be modeled through general systems for it to feel mundane.
- D&D magic must be at least partially universalized to be recognizable as D&D magic.
That is, if you want your fighter to feel like a general reflection of the warrior archetype, it has to be primarily using or modifying actions that are available to all characters. I think this might be slightly less true of the rogue (given it originally was defined having access to a unique, additional skill system), but if you're going to write a skill system, the rogue should probably have class abilities that key off stuff written inside it.
The second point is that D&D magic has to be fairly universal;
fireball can't be something only your wizard does, it must also be something that NPC wizards cast and that monsters deploy.
This is a good one, yes.
There's a certain quasi-simulationist strain to traditional AD&D which strives to maintain the pretense that the things Fighters do are not magical.
It ignores the defying physics which is going on in killing (or just in fighting in melee AT ALL) a 10' troll, or 50' or 100' giant or dragon,with a 3' long sword, and squints past the narrative nature of hit points, to latch onto the use of "mundane" actions and mechanics for non-casters to do their thing. Special abilities which can only be used a limited number of times per day or per encounter strain the credulity of players accustomed to this view of the game/world. 1E Oriental Adventures included some, but Ki abilities can be rationalized as a strain of magic.
5E brought some back (Action Surge, Second Wind, Fighting Spirit, for example), but it does it in limited number. 4E put fighters and casters in the same mechanical framework, and for a lot of folks a really big part of the core of D&D is that casters DO NOT operate using the same framework as non-casters.
The same perspective (drawn from AD&D, IMO) also tends to assume that anything (magical or non-magical) a PC or NPC does could theoretically be learned by another PC or NPC. That PCs and NPCs are fundamentally the same and play by the same rules. This is one way 3E really fulfilled (IMO) one of AD&D's core premises. In a way no other edition ever did. Even AD&D. Changing away from that entire mindset/design philosophy again is almost certainly another reason a lot of 3E players found 4E objectionable.
I'll submit that out of every edition of D&D I ran since B/X, 4e was the first one that made encounter/monster design fun for me, rather than solely necessary overhead. I think because of the rock solid monster framework. Coupled with some decent at-launch encounter building guidelines, plus rapid iteration across published adventures (not HPE, though) and LFR -- to show how much better and stronger encounter design could get within 4e.
I will also give a shout-out to people pushing the 4e envelop like Fourthcore, the Worldbreaker monsters (was that Angry DM?), multi-stage boss monsters... it was a pretty glorious time for "monster + encounter tech" in my experience / opinion.
Oh yes, absolutely. While you can elevate 4E encounters by designing interesting terrain and situations (like one would in any edition), just out-of-the-box monster statblocks in 4E do a ton of the heavy lifting to make encounters interesting.
The monsters are designed to be flavorful and tactically dynamic in ways 5E monsters aren't and in ways 3E monsters take page-long statblocks and lots of homework to try to achieve.
The sample encounter/monster combos in the monster manuals were solid, but really just picking a small mix of monsters (two or three different Goblins, say, with different roles) and then using their abilities tactically at the table would almost always suffice for an interesting fight.