I adored the playtest for the 5e sorcerer before the 2014 PHB. I haven't read it in years but my memory is this.

• Hit die were definited by the subclass, with the two example subclasses, wild magic and draconic having a d6 and d8 respectively (look how the DNA of this carried into the PHB with the draconic soul sorcerer getting a +1 per level to hitpoints, but then isn't explored in any future subclasses.

• Sorcery points and spell slots are a singular meta-currency like spell points.

• Class features generally use spell points too.

• Most excitingly, subclasses have features that come online when you are low on sorcery points.

That final element has this amazing interplay where you feel that you're burning your humanity (or species neutral equivalent term) as you use magic, and your innate or monsterous side comes through, it was a really cool design and I'd love to see it taken even further with a subclass that also incorporated hitpoints into the flow somehow, meaning that you are a tank before you cast your spells, but literally burn out your life force as you do so, revealing the monster underneath. It could be really cool for a vampire or something else that has an interesting interplay with harm, healing and magic.

The concept didn't survive playtesting for three reasons, it's execution wasn't fantastic, it was far more unique than the other classes and it was far more difficult to learn than other classes. It's a shame because I'd have loved to see the class iterated on and explored further in playtesting.

I've often wondered why I've never seen an idea that explores combining Sorcery Points with spell slots but not spell points by using 1st level spell slots as sorcery points and offering far more low level spell slots than currently offered, but still less than the 20 that they have by 20th level. Then lock their subclass features behind having only x amount of spell slots remaining. Sorcerers already fill a niche of being the most reliable low level spell slot slingers with their flexibile casting feature, and with this, they'd have an incentive to burn them up quickly to access the meat of the character fantasy; the subclass.

This would have a totally different resource management to any other class and be narrativly weighted, as to let your true self out (like the avatar state, your vampiric or draconic self, your raw chaos magic etc).

To be fair, the bulk of that is the ability trees, and since you only choose one (of seven, IIRC?), the vast majority of the 28pgs is irrelevant to each individual build.

Not necessarily.

A. They might have really liked the flavor and/or mechanics on the UA version and the homebrew version was different enough that it wasn't as interesting to them

B. OP might have taken long enough researching Mythic and it's alternatives that the player had time to reconsider their choice. The player might have been split between the two anyway

C. The player might have realized that running a homebrew/UA class was going to be a lot of extra work for themselves and/or the GM, so they decided to do something easier

Honestly it's fine as long as everyone is in a similar power band. All you do is scale up encounter difficulty slightly by adding monster abilities or extra opponents. All power levels are relative to the campaign's conditions, after all.

What sucks is when one person is vastly stronger than their peers and sees no problem with this. Those people never learned to share and should be shunned.

I allowed KibblesTasty's Psion, Laserllama's Psion, or alteratively, the 2020's UA: Psionics Revisited

Honestly, the Light/Darkness one more or less became the Gloomstalker, I believe? Oh, and there's no OG mystic build that isn't broken AF, so I assume OP meant another thematically adjacent class entirely.

MCDM just released their psychic class The Talent. Though I haven't played it yet I have played with their other classes. The Beastheart and Illrigger are fun as fuck and I can't wait to play a Talent.