From http://www.enworld.org/forum/showthread.php?333362-Fixing-the-Fighter/page45&p=6073774#post6073774
This is an approach to GMing which I got from Pemerton's discussion of how he runs his games. I've used it in my last two 4e campaigns (one ongoing), albeit initially without full awareness of what I was doing. I've been finding that for 4e it seems to work better than the traditional (more gamist-challenge in process-simulated environment) style I use in my other FRPGs such as Pathfinder Beginner Box or Labyrinth Lord.
My idea of GM-led Pemertonian scene-framing:
1)GM sets up scene that derives from prior events but is framed to be interesting - as opposed to process-simulation where scenes are not 'framed' but derive from procedural generation, eg random encounters, d% event tables.
2)Resolution of the scene is left entirely open and up to the players - as opposed to hard railroading where there is a required scene resolution. And
3) Future scenes are largely determined by players' choice/action in past scenes, as opposed to linear AP style play where scenes are pre-written along the set continuum of the adventure. But in looping round to #1 the GM is guided more by what would be a cool/interesting/fun result than by Simulation concerns - though for a D&D world the two may not be hugely different.
The way I've been doing Pemertonian scene-framing it mostly resembles Sandbox play quite closely, with occasional elements of AP style linear play where I'm using a linear adventure (Heathen, Orcs of Stonefang Pass) more or less as written. But I try to open up those adventures for more of a Pemertonian approach, eg I tweaked the dramatic climax of Heathen to create more of a Narrativist style dramatic moment that raised questions of actual moral choice for the PCs, and I inserted a dragon into Stonefang Pass that led to a great dramatic moment when a player 'stepped on up' and talked it down.
One difference between Pemertonian subjective scene-framing I use in 4e and the kind of objective content generation I used in eg my Pathfinder Beginner Box game is that in the scene-framing approach the encounters are subjectively tailored - when my 4e Forgotten Realms Loudwater group met an Ettin, it was because I thought an Ettin would be a good encounter for them (there was foreshadowing of its presence and they could have avoided it, mind you) in all the circumstances. It was a tailored encounter.
Whereas when my PBB group met a Gray Ooze, it was because that was what the system & environment generated - it was a status quo encounter.
IME, the two are often not very different in-play, but over time tailored encounters lead to a much lower lethality level and less of a revolving door of PCs. I still kill PCs in Loudwater - first session TPK, another 2 temporary & 3 perma-deaths in the 32 sessions since - and scenes can be framed as "today is a good day to die", as at the climax of my Southlands campaign where the PCs had repeatedly screwed up and were left making a hopeless last stand at the bridge against a thousand Horde Ghouls and their Necromancer overlord (note that they chose to make the last stand, to give their allies time to evacuate the doomed town). But the general level of random PC death with scene-framing is much lower than with status-quo Simulation, tends to come at moments of dramatic climax*, and so often feels dramatically appropriate when it does occur.
*There was that one time both GM (me) and player screwed up what should have been a low-lethality encounter- I levelled up an encounter with wolves by stupidly making them all Elites; the Wizard player won Init, stepped forward in front of her allies and cast Burning Hands on the whole pack in round 1... *ouch*.
This is an approach to GMing which I got from Pemerton's discussion of how he runs his games. I've used it in my last two 4e campaigns (one ongoing), albeit initially without full awareness of what I was doing. I've been finding that for 4e it seems to work better than the traditional (more gamist-challenge in process-simulated environment) style I use in my other FRPGs such as Pathfinder Beginner Box or Labyrinth Lord.
My idea of GM-led Pemertonian scene-framing:
1)GM sets up scene that derives from prior events but is framed to be interesting - as opposed to process-simulation where scenes are not 'framed' but derive from procedural generation, eg random encounters, d% event tables.
2)Resolution of the scene is left entirely open and up to the players - as opposed to hard railroading where there is a required scene resolution. And
3) Future scenes are largely determined by players' choice/action in past scenes, as opposed to linear AP style play where scenes are pre-written along the set continuum of the adventure. But in looping round to #1 the GM is guided more by what would be a cool/interesting/fun result than by Simulation concerns - though for a D&D world the two may not be hugely different.
The way I've been doing Pemertonian scene-framing it mostly resembles Sandbox play quite closely, with occasional elements of AP style linear play where I'm using a linear adventure (Heathen, Orcs of Stonefang Pass) more or less as written. But I try to open up those adventures for more of a Pemertonian approach, eg I tweaked the dramatic climax of Heathen to create more of a Narrativist style dramatic moment that raised questions of actual moral choice for the PCs, and I inserted a dragon into Stonefang Pass that led to a great dramatic moment when a player 'stepped on up' and talked it down.
One difference between Pemertonian subjective scene-framing I use in 4e and the kind of objective content generation I used in eg my Pathfinder Beginner Box game is that in the scene-framing approach the encounters are subjectively tailored - when my 4e Forgotten Realms Loudwater group met an Ettin, it was because I thought an Ettin would be a good encounter for them (there was foreshadowing of its presence and they could have avoided it, mind you) in all the circumstances. It was a tailored encounter.
Whereas when my PBB group met a Gray Ooze, it was because that was what the system & environment generated - it was a status quo encounter.
IME, the two are often not very different in-play, but over time tailored encounters lead to a much lower lethality level and less of a revolving door of PCs. I still kill PCs in Loudwater - first session TPK, another 2 temporary & 3 perma-deaths in the 32 sessions since - and scenes can be framed as "today is a good day to die", as at the climax of my Southlands campaign where the PCs had repeatedly screwed up and were left making a hopeless last stand at the bridge against a thousand Horde Ghouls and their Necromancer overlord (note that they chose to make the last stand, to give their allies time to evacuate the doomed town). But the general level of random PC death with scene-framing is much lower than with status-quo Simulation, tends to come at moments of dramatic climax*, and so often feels dramatically appropriate when it does occur.
*There was that one time both GM (me) and player screwed up what should have been a low-lethality encounter- I levelled up an encounter with wolves by stupidly making them all Elites; the Wizard player won Init, stepped forward in front of her allies and cast Burning Hands on the whole pack in round 1... *ouch*.
