So separating out "OC" style play, which in my experience gets pretty precious and can often lead to main character syndrome in a party that's not all there to play around with OCs, reading the (really freaking excellent)
Fabula Ultima game book + GM's guide I think gets to a lot of the a) fun in "neotrad" or whatever you want to call it [I think that the keeping of traditional style leveling + classes + skills + frequent and somewhat tactical battles is what separates neotrad from the more narrativist games -
@kenada provided a definition somewhere which boiled down to 'neotrad is a trad game using indie/story game techniques and systems'] and b) ways to run it.
On a) it makes it clear to all the players that everybody at the table must be as invested in moment to moment gameplay as the GM. Because players are steering the goals, they must agree to do so - accepting premise + world + risk. The world itself is co-generated during setup, and further defined in relation to the player's class and feature choices. You see
Daggerheart doing something very similar in its playtest. It puts limits on the GM, while inculcating techniques and game rules to maximize drama within scenes through mechanically tied bonds/identities/themes. It also makes it clear that the game is in service to a narrative arc that will be created via play, based on some starting conditions. Plot will happen, it's just not pre-designed beyond villains. Eg: "Metichor, Arch-Magus of Death is working on a ritual that will drain all life from the land. How will you stop him?" and the rest of the game is the players figuring that out - racing a clock (literal, it's using FITD clocks to indicate progress/threat).
I think most neotrad games will try and incorporate consequences to failure that avoid PC death.
Fabula Ultima explicitly says that PCs don't die at 0 hp unless they consent, they surrender instead. However, surrendering has permanent consequences - the bad guy's plans tick forward, the PC's bonds change to a darker place (and you invoke your bonds for mechanical advantage...), etc. So in many ways you've got a much richer outcome then a trad game, from a narrative/story perspective. People evolve from failure or success or setback, and then this ties back to mechanical uses. Finally, players accumulate metacurrency they can spend to invoke their bonds, change a scene (within bounds), and some other stuff.
On b) you get a set of principles and effective guidance via words and mechanics to implement them, a lot of it around seeking input from the players while being clear with them in terms of information + consequences. It's a bit long to type it all out, but I'll say that between the DM chapter in the play guide and the short GM's Guide, it contains probably some of the best GMing advice for this sort of game I've seen anywhere - presented in a manner that supplements the 4e DMG in my head as best game-running advice (towards narrative outcomes).
The Core GM principles espoused there are:
-Portray a Wonderous World (nothing new here, just make things unique but don't get bogged down in details - be short and then let players ask questions)
-Manage information. (usual stuff, you're going to be managing the info about teh world and people)
-Seek input. If class/identity relevant, ask for player input on teh fiction and game world.
-React to heroes' choice. (yeah, but does explicitly note
apply consequences of their freely chosen strategies and don't pull punches)
-Enforce transparency. Very narrativist, make sure people understand what's at stake before they make a Check/choose courses of actions.
-Keep the pressure high. (PCs want to do things, somebody will want to stop them, there's not enough time, make some hard choices, use public facing clocks)
-Think cinematically. Lots of video game framing reference here. I find myself doing this as shorthand in some of my 5e games too anyway - we're all younger gamers.
-Start and End Scenes. Hard scene framing with close-scene resolution. Once things are petering out, ask if it's done and if no players have anything they want to do - move on.
-Design interesting battles. No filler fights, no room by room dungeon crawls. Be intense and dramatic.