Every 6 months I run a game for my old High School gaming group - so far one shots exploring different play styles and power levels of d20 Fantasy. This January I don't know if I'm going to stick with building a power level focused version of D&D like I have in the past or use it as a springboard for 5th Edition. If I were to do the latter I'd still want to build characters using my card stack system.
To explain, I have several stacks of cards that fall into certain types - levels in the 4 core classes, character race, character gifts, major equipment - and each player starts with one stack. They pick their favorite from the stack and pass the rest to the right. Eventually everyone gets a card from each stack which, taken in total, gives them their stats, class levels, race, special abilities and core equipment. It's a midpoint between pre-made characters and designing the characters at the table.
However it is very much keyed off of 3rd edition multi-classing. 5th edition doesn't mutli-class as cleanly, instead calling for a lot of specific decision points for a design in play character. 5E really wants you to start at 1st level and work your way through to give your character an organic feel, and the classes have internal option trees rather than promoting multi-classing and prestige classes. It keeps things uniform, but it also makes things messy for how I've done this in the past.
This speaks to some of my issues with 5E, not that I've played enough of the game to really form opinions on them. As I said before I think character creation is still too complex: you don't just select a race but a subrace, and rather than a +2/-2 on stats and some bonuses to skills/saves there are differing stat levels and special racial powers that aren't mimicked anywhere else. Classes are easy at first but have internal trees that force decisions at later levels that radically change the character. Feats are optional, but feel optional in the way that 2E non-weapon proficiencies were optional in that all future supplements will assume their use. At least backgrounds are consistent with their 'two skills and a social advantage' structure, which is nice, and you don't have points to distribute between skills the way you did in 3E, which could be a time saver.
It all makes me hesitate to use 5E for the January game. Maybe everyone will think it's brilliant and will start running games for their kids with it (it's probably good for that). But it doesn't feel innovative enough for me to one to showcase rules with it the way I have with, say, fanmail chips, noun+verb magic systems and decentralized power options in previous games - things that weren't in any of the games this group played with before that really show off new gaming tech.
There are a lot of good ideas in 5E, but aside from the Advantage/Disadvantage rules I don't see enough breaking new ground to justify some of the character crunchy bits complexity.