3e prestige classes are maybe 75% redundant in 5e.
[…]
Of those, 5e has no need for Prcs of #1-6. 7 is straight up a new class. So PRCs we could only be good for 8 or 9. Like Shadowdancer as rogues, rangers, and bards play Very differently.
While this may be subjective, I think a lot of PrC would ideally fall in the same bucket you place shadowdancers in. For example, Arcane Archers should at the very least be buildable off of a Fighter or a Ranger. Building them as a subclass means half the relevant audience is sidelined. And we’re not even getting into more interesting/exotic builds, for example, why wouldn’t an Arcane Trickster grow into an Arcane Archer? You could say they can and they simply need to dip into Fighter first, and maybe that’s fine, but maybe that does not fit the character theme as well as if there was a more direct path to get there. That is why I say subclasses are like straight jackets. Needlessly constraining.
Because the undead in D&D have been given the short end of a stick when it comes to alignment.
Baelnorn says hi.
I haven't read the whole thread but I would really not like the idea of subclasses being replaced by paths of....feat trees, I will be calling them that for simplicity.
For one, we can look to different systems as evidence that new players can get utterly overwhelmed by too many options from the get go. This happenned for Cypher, hence limiting of what your role gives you in 2e, it also happenned to Blades in the Dark, which had no class system - players were supposed to just pick up abilities from one big list, but many found it overwhelming.
Second, I think it would be much easier to abuse if all subclass abilities were converted into feat trees, especially with more elastic requirements it could lead to something like a Paladin taking early abilities of Oath of the Crown and late-game ones of Oath of Redemption, effectively picking best of both worlds.
These are all legitimate concerns, and difficult game design challenges. I do think they are surmountable, but should not be underestimated.
The gold standard of ultimate game design, Starcraft, famously had the motto of "easy to learn, hard to master".
Clearly, a sophisticated and interesting system which is "hard to learn, hard to master" is going to be disappointing for most players.
While this is admittedly hand-wavy, I have a vision of a skill tree or feat tree system where there is not only "one big list" to pick from, but there are also many curated "packages" or "recommended progressions" to make it easier to get started with. Just like the classes and backgrounds of the PHB have some preconceived (but optional) recommendations. "50 gp or the following specific pieces of gear", "4 x 1st level spells, the following specific ones are recommended", etc.
So you could have a "Battlemaster" package and an "Eldritch Knight" package which tell you exactly how to spend you "skill points" at each level. The end result is basically a class exactly like the one you would get playing 5.5e rules. But if you wanted to swap out one ability, there would be clear rules (you get this many skill points back, go shop around). A lot of this could be made easier with digital tools (e.g., "show me the skills I can actually afford with my free skill points and the prereqs I already possess", and then further filters like "show me only the options belonging to the martial paths, as I don’t care for magic in this build", etc). And if you wanted to do
tabula rasa you also could. All sophistication levels should be supported within the same system, at the same table. "Easy to learn, hard to master."