Failing Forward

[The Crimson Binome]

It's actually easy to avoid contradiction without a single authority. You just undertake (as a rule) not to contradict anything that has been stated as true in the game. You also, by implication, agree not to assume that anything that has not been said is necessarily true.

I can see how that would be a matter of custom and preference, then. The strong GM model would say that one person is in charge of all details, because that way you just need to check any new detail against what that person knows. Your distributed model (also mentioned in a previous post, which I failed to respond to) has you only need to check against that which has been officially declared.

I still think my preferred model would work better for me, since the GM is free to spend any amount of time working out the background details between sessions, but I could see how the alternative would work well if you don't have a GM or if the GM doesn't want to spend out-of-game time in figuring that stuff out. I honestly think that I would have difficulty, as a GM, remembering whether or not something has been declared - when I act as the sole oversight authority for the entire game setting, I know that whatever I have decided is true regardless of whether the players know about it, so I don't need to track whether any of it has been declared yet as long as it's consistent with my underlying model.

 

[Lanefan]

"Fail forward" is about preserving momentum in play. It is part of a broader approach to RPGing in which the GM is expected to frame the PCs into challenging and/or confronting scenes (challenging and confronting both to the PCs and the players), and when the GM stopping to ask "What do you do now?" is a rare pause in the dynamics of play, rather than the norm.

This has nothing to do with railroading or "tunnel vision".

We-e-ell, in some ways, yes it does.

The DM asking "What do you do now?" is in fact the DM handing the momentum to the players and saying "Here, you drive the bus...or at least decide where it's going." If the DM is expected to frame every scene as opposed to the characters putting themselves into situations of their own choosing now and then I'd say there's a very real temptation on everyone's part to get on the railroad (which of course for some groups may be just fine) and thus it can't be discounted for the purposes of this discussion.

Lan-"what do you do now?"-efan

 

[pemerton]

The DM asking "What do you do now?" is in fact the DM handing the momentum to the players and saying "Here, you drive the bus...or at least decide where it's going." If the DM is expected to frame every scene as opposed to the characters putting themselves into situations of their own choosing now and then I'd say there's a very real temptation on everyone's part to get on the railroad (which of course for some groups may be just fine) and thus it can't be discounted for the purposes of this discussion.

If there is one thing that I would lament if it became a rare thing then " what do you do now?" would probably be it. That most basic of questions drives a game of meaningful player choice. Failing forward facilitates players who never have to really think of anything significant.

And here I'm with you. It's also part of the basic GM advice in Apocalypse World and many of the legion of derivative games - to ask after almost every GM action "What do you do now?"

I think this is a case where some familiarity with the technique in actual play can help.

There are two ways of asking "What do you do now?" They are illustrated nicely in this excellent post by [MENTION=99817]chaochou[/MENTION]:

As someone who does a lot of scene-framing, I'll start by saying this: it concerns things like intent and game procedure and these are things which are a lot easier to do, or see in action, than to write about. . . .

In my view, the fundamental cornerstone of scene-framing is character. . . . things that talk directly about personality, goals, flaws, relationships, dependencies, desires, problems. . . .

Written down it looks easy. At the table it is not. If you play a FATE game, where 5 players each have 10 Aspects, you may have 50 goals and flaws and problems competing for time and if you've got 2 ideas for each of those then you've got 100 possible scenes before play even begins. Then people start interacting with each other and NPCs and before you know it you've got thousands of potential directions to take play.

At which point feeling out the players becomes necessary. So if someone threatens an NPC like this - 'Drop the gun or I'll burn your damn house down' you might say 'That would be a cool scene...'. If you get a good vibe back from the table, well if it's still appropriate by the end of this scene then as the spotlight comes back to that player you could say 'Okay, so you're outside Jed's ranch with torches and oil somewhere just after midnight. You're starting forward when suddenly you hear the tail of the rattler as it rears up right in front of you'. Previously stated goal - burn house down. New complication - rattlesnake. . . .

When I cut into Jed's ranch and a rattlesnake, I'm not asking 'Are you good enough to deal with a rattlesnake?' I'm asking 'How much are you willing to risk in order to make good on your threat?' The scene is not there to process the outcome of success or failure, it is there to reveal more character to be used for future scenes.

This is why games written to this style (Burning Wheel, Apocalypse World and it's spin offs like Dungeon World and Monsterhearts, FATE to some extent, Dogs in the Vineyard for sure) don't tend to penalise failure particularly hard. Failure simply imforms the next scene. . . .

Finally, one tell-tale sign for this style is to watch your own GM-ing. Starting scenes is easy. But how are you ending them? And how are you moving on to something else? If you say 'What do you do now?' you are not playing using scene-framing. If you cut to a place where nothing is happening right now you are not playing using scene-framing. Scene-framing in the most aggressive sense means going 'bang!' - straight into the conflict, straight into the action. And again. And again. When it hums like this, as each scene unfolds everyone at the table is alive with ideas as to what the next scene will be be.​

"Fail forward" is a technique best-suited to scene-framing play, that is, play in which narrative dynamism is front-and-centre, and momentum is only rarely lost.

Of course, in scene-framing play, the GM needs to listen to player action declarations: when the GM tells the player that the rattlesnake rears up in front of the PC, the player is expected to declare some action in response - ie they declare what it is that the PC does now.

But in scene-framing play, the GM shouldn't be asking the players "What do you do now" as part of the process for transition between scenes. The GM should be framing the PCs (and therefore players) into the next confrontation.

And flipping it around: if you're playing exploration/discovery/GM-world-building style, rather than scene-framing, then you don't need "fail forward", because it is completely fine for the game to come to a halt, for momentum to be lost, and for the GM to look to the players to kickstart things again. This is the sort of play that [MENTION=29398]Lanefan[/MENTION] and [MENTION=66434]ExploderWizard[/MENTION] are familiar with and prefer. It has very little in common with the Apocalypse World sort of play that [MENTION=87792]Neonchameleon[/MENTION] is mentioning in. And as I've tried to explain, the question "What do you do now?" is playing very different roles in the two approaches.

Finally, anyone who thinks that scene-framing play is about railroading is suffering a fundamental misunderstanding. "Railroading' is about pre-determined events where choices don't matter. Scene-framing is about leaving everything open - events, the presence or absence of rattlensakes (or curses or tripwires or . . .) - until a new fictional situation is narrated by the GM in response to the events, context etc generated by the previous scene.

Here are two links, to reports of my first and my most recent Burning Wheel session. Both involved use of "fail forward", as I've mentioned in more detail upthread. Read them and you'll see that neither was anything like a railroad. How can it be a railroad, for instance, to have the mace carried down the mountain stream from one PC to another when, until the various action delcarations were made, I didn't even know that one PC would be in the cave looking for the mace, or that another PC would be near the stream at the foot of the keep following servants doing laundry - this being the result of earlier action by the PCs which resulted in an NPC priest's vestments being dirtied and hence needing laundering? Where is the predetermining, or the stifling of player initiative/creativity?

 

[pemerton]

Puzzles and traps based games are utterly destroyed by the players changing the puzzles.

<snip>

If the game is about explorers or archaeologists or as fish out of water exploring strange new worlds and boldly going where no one has gone before then making the PCs more in tune with their environment is precisely what you don't want to do.

However, it is possible to have successful play in which the PCs are exploring the gameworld and solving puzzles, without the GM having predetermined everything.

This won't be the sort of exploration/discovery play that [MENTION=6775031]Saelorn[/MENTION] and other posters in this thread prefer, because at the metagame level there is no exploration of a pre-established setting.

But in the fiction exploration and discovery are taking place, and at the table the players are still acquiring information about the backstory that is being generated by the GM, and are piecing it together to develop their overall picture of the relevant backstory.

Here's a link to an actual play example from my 4e game.

 

[pemerton]

So question... in a situation like the above... when has the trap actually been sprung (when do I suffer it's effects??). It seems like just in looking for the trap and failing you've now put my character into a situation where he's (partially??) sprung the trap just by looking for it. Not sure this would be cool with me as a rogue since it would mean when I fail at searching for a trap (not necessarily doing anything to spring it) I then end up guaranteed to suffer it's ill effects... if I fail to disarm it... well, and since I'm the rogue I'll be searching (and failing) at a higher rate than other characters.

This seems to be something of an artefact of D&D's action resolution system: wizards and clerics use spells, which rarely fail; fighters use combat, which has its own resolution method (and to the extent that it embodies "fail forward", that is via the hit point and related mechanics); and rogues use the skill system, which is where "fail forward" in the paradigmatic sense is most likely to see application.

That said - if you think my characterisation of fighters is too narrow (either in general, or relative to your own play experience with D&D), and if you think that I'm being too sanguine about spell-casting and that it often does fail, then the concern should go away. The rogue's failures to find traps see him/her framed into a difficult new situation; but equally the casters failed attempt to charm the chamberlain sees him/her framed into a difficult new situation.

I think there is another issue here that is distinct from fairness between classes, which is about the extent to which the GM is at liberty to frame the PC into a situation which presupposes activity (like walking around a room) that the player hasn't expressly declared.
[MENTION=97077]iserith[/MENTION] has already discussed this in some detail.

I think that, for various mostly historical reasons, D&D players are especially cautious about this. One time, in my 4e game, discussion with a coven of hags was being resolved as a skill challenge. At a certain point the Pact Hag said something to one of the PCs, and I deemed her to be using one of her command-type abilities and told the player that his PC had stepped from A to B, where B happened to be a trapdoor that the hag then operated by pulling on a rope. The player didn't contest the issue, and from my point of view that is legitimate fictional framing in the context of a negotation with a Pact Hag - in the end, after all, the players succeeded at the challenge (in part by dealing with the spiders in the pits below the trapdoor) and got what they wanted from the hags.

But I think a lot of D&D players, including many who post on this forum, would think that that was unreasonable GMing.

 

[Lanefan]

This seems to be something of an artefact of D&D's action resolution system: wizards and clerics use spells, which rarely fail; fighters use combat, which has its own resolution method (and to the extent that it embodies "fail forward", that is via the hit point and related mechanics); and rogues use the skill system, which is where "fail forward" in the paradigmatic sense is most likely to see application.

That said - if you think my characterisation of fighters is too narrow (either in general, or relative to your own play experience with D&D), and if you think that I'm being too sanguine about spell-casting and that it often does fail, then the concern should go away.

Spells fail (or don't do quite as intended) in my game quite often, as I have harsh interruption rules and frequently require casters to aim.

I think there is another issue here that is distinct from fairness between classes, which is about the extent to which the GM is at liberty to frame the PC into a situation which presupposes activity (like walking around a room) that the player hasn't expressly declared.

Yep. In fact your DMing style and mine might be almost opposites: I want to be reacting to what the players (as characters) proactively do, while you want the players (as characters) to be reacting to what you proactively frame them into.

Put another way, if I'm reading you right you want to jump from the end of one encounter pretty much straight into the start of another; where I want to know what happens in between and how (and why) they get from one encounter to the next.

I think that, for various mostly historical reasons, D&D players are especially cautious about this. One time, in my 4e game, discussion with a coven of hags was being resolved as a skill challenge. At a certain point the Pact Hag said something to one of the PCs, and I deemed her to be using one of her command-type abilities and told the player that his PC had stepped from A to B, where B happened to be a trapdoor that the hag then operated by pulling on a rope. The player didn't contest the issue, and from my point of view that is legitimate fictional framing in the context of a negotation with a Pact Hag - in the end, after all, the players succeeded at the challenge (in part by dealing with the spiders in the pits below the trapdoor) and got what they wanted from the hags.

But I think a lot of D&D players, including many who post on this forum, would think that that was unreasonable GMing.

I'd certainly ask what the hell happened to my saving throw against her command ability were I the player, as it doesn't sound like this character got one...

Lan-"but if I'd got a save and failed it, what you did is perfectly fair game"-efan

 

[Manbearcat]

Play example (sblocked for space). I believe I've used it before so it may be familiar to some folks.

[sblock]

Otthor's player:

He turns to Saerie so she can see his lips. "We have little time before that comes in. Find the dog and bring him back to us. Even if your theory is correct and he is deaf, perhaps its just age. Maybe you can still speak to and understand him as you do with other wild beasts." He points toward the common building. "I'm going there. With that storm, surely we're going to lose the tracks out of this place and to wherever the refugees went. Maybe I can find information on their path there. Besides, we're running low on supplies <he points to her waning quiver> and perhaps I can salvage something there." He pulls the remaining bundle of arrows from his quiver and places them Saerie's own quiver. "We can investigate that cellar in the morning."

He grabs his gear from a nearby table and turns to her again before he sets out. "Don't be alarmed when you see a fire. After I'm done there, I'm going to put the bodies of the poor settlers to rest...and my own mind."

Saerie's player:

I thank Otthor for the help in getting my injured bear-friend settled in and for the quiver-replenishing bundle of arrows. Before he leaves with his ominous last words regarding the fire, I relay my agreement with the plan. I will do everything I can to find the dog and bring her back with me.

I move to the last known spot where I saw the dog; the crest of the hill which overlooks the drop into the settlement. Somewhere beyond the now-open gates is where the the old canine currently lies. Depending on how terrified she was, which may be very, she would be either quite near or quite far. The snow is thick enough here that finding the spot where I last lost sight of her is a triviality. But to track the old girl down may be another thing entirely. The wind is already steadily picking up in the exposed tundra up here. I wrap my scarf over my face and pull my hood up over my head.

Hunt and Track (Wis)

5, 3 + 2 = 10

I follow the creature’s trail until there’s a significant change in its direction or mode of travel. I also determine what caused the trail to end.

GM (Me):

The wind dries your eyes to the point of pain. Your training ensures that you easily pick up the odd lope of the dog, its gait clearly weary, hobbled by age and its recent torment. It must know the territory well as it appears to have made a go for a line of snowdrifts that it could dig into and hide behind as the gusts pick up speed.

When you spot her, she has dug into the side of the drift facing away from the wind. Laboring over the effort, her tongue is out as she is in full pant. She doesn't notice your presence and as you watch her, you can see the fear in her eyes and the hunger betrayed by her gaunt form.

Saerie's player:

I slowly get into her line of sight so as not to startle her. When she sees me, I'll carefully pull out some dried meat from my pack. I'll get down on all fours, assume a non-threatening posture, and entreat her to a free meal, given in good faith.

Parley (Cha)

My leverage is food for the starving dog.

5, 1 - 1 = 5. Fails.

Mark 1 xp

GM (Me):

Her ears perk up. The dog looks interested in your offering. However, if she is deaf she doesn't need ears to perceive the thundering herd of reindeer bearing down on you. She, like you, can feel it in the ground. Your mind is ushered back to the Winter Wolf's words regarding a maddened realm near the two great bodies of water in the highlands where "...herds of reindeer would stampede each other and tear each other, and themselves, to pieces." Whether they're simply running from the fury of the storm that is hot on their tails or deranged creatures intent on your harm is impossible to say at this distance (far, 10 creatures).

The scared dog abruptly bounds out of her carved hole and rushes to your position where she might see the obscured threat. When she sees what is on its way, she tucks her tail between her legs and runs in a circle behind you, looking to you with uncertain eyes.

The reindeer are closing fast.

Saerie's player:

I spend the shortest of moments evaluating the situation. I want to know if this herd is behaving normally or if they look like they're intent on harming us. Also, is the snowdrift a reasonable location to obscure is from the herd if we have to hide.

Discern Realities (Wis)
6, 2 + 2 = 10

3 questions and + 1 forward when acting on the answers.

* What is about to happen?
* What should I be on the lookout for?
* What here is useful or valuable to me?

GM (Me):

1) The herd are bearing down precisely on your position. Its so tringulated that they either want the cover of the drifts, which would be odd, or they're coming after you specifically.

2) A predator behind them that they're running from or absolutely no fear when they near you. If they're of their right minds, these creatures are typically unnerved by humanoids.

3) The drift may just provide you enough cover to hide. The hole she carved out is really your only shot as you don't have enough time to fully carve out your own.

Saerie's player:

Knowing that making a stand out here against that herd would likely be foolhardy, I pick her up and rush to the drift. I widen the hole she made and we both get into it. When we're in, I collapse the roof of the hole, hoping the reindeer didn't spot our escape.

Defy Danger (Dex)
2, 4 + 2 (+1 forward from DR) = 9

Success but hard bargain, worse outcome, or ugly choice.

GM (Me):

In your efforts to hold the struggling dog to get her in the hole, she kicks your quiver which spills the majority of your arrows (leaving you with 1 Ammo remaining). You can gather them and make a stand before the beasts set upon you, or you can sacrifice them and get in your "hidey-hole." Your choice.

Saerie's player:

No question. Get in the hidey-hole with the dog and let the threat pass us by. I cringe when I hear the tell-tale snapping of arrow shafts as the herd tramples them underfoot. We get out when the herd passes fully.

GM (Me):
You've won the trust of the dog. She follows you back to the settlement. Your interactions with her reveal, quite clearly, that she is indeed deaf.

[/sblock]

This is, of course, Dungeon World.

There are lots and lots of dangers that the player is aware of when setting out on this harrowing excursion for the dog:

1) Waning daylight.
2) Frozen wasteland.
3) Looming blizzard on the periphery that could change direction at a moment's notice.
4) A quiver (though recently modestly refreshed) that isn't rich enough in arrows for a protracted skirmish.
5) Saerie is without her injured bear companion.
6) Saerie is down some HPs.
7) A land filled with dangerous creatures that are inexplicably going mad with murderous bloodlust.
8) The dog is terrified and starving...an unpredictable and desperate creatures makes for a dangerous creature.
9) We're already aware of creatures morphing and the PCs have just had an encounter where the ruined town they were seeking is bereft of all life but one dog. The common house was a house of horrors akin to the final scene in Aliens; a gestating abomination hatchery where many/most (?) of the former inhabitants were in pods and changing into something unfathomable...connected to each other and guarded by a sentient mass of tentacles and teeth. They've seen signs of this before in the open tundra (a burst cocoon suspended in a glacial moraine, gore leading off into the wilderness). Abominations obviously lurk in this place.

But the PC in question (Saerie) feels bound (literally - by 2 of her 4 bonds) to track down this terrified, starving, old dog that sprinted out of the common (horror) house when Otthor and Rawr (the other PC and the bear companion) approached and the tentacle mass smashed the front doors into splinters and attacked them. The formerly trapped, now free, dog ran straight out of the settlement.

The PCs knew (a) that deranged, psyche-assailing sounds and images flooded this place, (b) every living thing here is gone or changed...save for this lone dog. So, bound by her duties and yearning for clues, the PC decides to defy all of these potential dangers and set off headlong into the arctic tundra in search of the dog.

Getting into the nuts and bolts of the above sblocked instance of play, there is a specific moment where the conflict is escalated. From there stems a snowballing situation that turns into impactful decision-points. The PC's Parley move with the dog is an outright failure (which, of course, earns her xp). There are many dangers I could have made manifest from this. I could have had the starving dog attack her (Turn the Move Back on Them). I could have had the blizzard suddenly and violently change course, cutting her off from the settlement so that she must find shelter or likely perish from exposure in this frozen wasteland (Reveal an Unwelcome Truth). I could have had a monster (perhaps a hidden tundra yeti) ambush her or a false snow floor swallow her up into a crevasse (Use a Monster, Danger, or Location Move). I could have done any number of things that made sense given all the dangers that lurk in this place.

I chose to introduce the ominous thundering of the maddened reindeer herd (Show Signs of An Approaching Threat). Why? Because it escalates things dramatically and creates an interesting decision-point for the character. Things could snowball very, very badly for the player from this point depending on what they do and how they roll. It also realizes the foreshadowing a did in a prior encounter with a Winter Wolf and his Dire Wolf pack when I made this threat a latent one.

It fails the situation forward. Her intent in this scene is to befriend the dog, get it back to safety, confirm her suspicions (the dog is deaf therefore invulnerable to the psyche-assailing affect happening at World's End Bluff), and attempt to communicate with the dog to find out what happened in the now-ruined settlement (she can speak with animals). If I just kill the dog with a monster or if I make the dog unreachable (either because it fights her to the death because it wants to eat her), then that is a hard failure that makes her intent unrealizable. So I complicate the realization of her intentions with major problems that can quickly turn into mortal ones.

How badly does she want to rescue this dog? She could have made an action declaration that she frightens the dog off in the direction of the herd (creating interference with a death sentence for the dog) and melted into her surroundings (she has a Camouflage move that would have made this trivial for her), evading the oncoming threat of the maddened herd. Nope, she looks for sufficient cover for she and the terrified animal, grabs it, and takes cover in a nearby snow drift. In the desperate scramble, the protesting dog upends her quiver, spilling 2 of her 3 Ammo onto the snow.

Another choice. How much does she want that 2 Ammo? An archer Ranger transiting a deadly frozen wilderness with only 1 Ammo (with no confirmed means nearby to Resupply) is a recipe for disaster in Dungeon World. She can save the arrows but face the herd (and whatever shakes out of that...which would almost surely be a chase scene...with an old, starving dog as a liability...). Or...she can secure the dog and her safety and just deal with the unfortunate ammo deficiency.

- There is no railroad here.

- The player has agency going in to the conflict (awareness of the dangers and the stakes) and agency during the conflict to affect the trajectory of the scene. Increased specificity in action declarations (or my own increased demands for specificity on those action declarations) and intensified, discrete resolution of micro-component-parts of each action declaration (requiring several more rolls rather than effectively abstracting things by saying "yes" because we aren't focusing on non-thematic, conflict-neutral, minor actions whose resolution might lead to tedium and pace-atrophy), and process-sim rendering of fallout by the GM (the dog hates/attacks you is the only possible outcome of a failed Parley with Leverage as food for the starving canine) wouldn't have increased player agency.

- Narrative momentum never stalls and a dynamic scene which could have borne itself in any number of ways depending on differing Player Moves, resolution of those Moves, and corresponding GM Moves. This scene could have ended with:

a) PC death
b) dog death or at least the resource/asset being lost to the PC
c) player getting lost in the frozen wilderness with or without the dog
d) player being stuck out in the blizzard with or without the dog
e) player's resources (HPs, Ammo, Adventuring Gear, Rations, general gear/weapons including her cold weather gear which protects her from having to Defy Danger from the elements) becoming utterly diminished (instead of just partially) for the future adventure/journey
f) discovering something interesting or terrible about the highlands setting/mysteries.

The dog could have met an unceremonious end out on the frozen tundra. Or it could have become Rations for the PCs in a barren wilderness where foraging is an impossibility. Or it could have been just another dog (not a point of interest for the PCs). Instead, this dog became a dear, hobbit-like companion for the PCs and a linchpin for the game's future.

That is how proficient use of the Fail Forward technique and (coherent, high-utility but low complexity) system work together to achieve their intended result - snowballing, "play to find out what happens" emergent story and player agency dynamically affecting the trajectory.

 

[Balesir]

I still think my preferred model would work better for me, since the GM is free to spend any amount of time working out the background details between sessions, but I could see how the alternative would work well if you don't have a GM or if the GM doesn't want to spend out-of-game time in figuring that stuff out.

Sure - you should use what works for you. I would add to the times the alternative would work well "when you want to deliberately leave the background flexible to accommodate stuff the players indicate they want in the game". This can be a major aim when working with things like 13th Age's backgrounds and "one unique thing".

I honestly think that I would have difficulty, as a GM, remembering whether or not something has been declared - when I act as the sole oversight authority for the entire game setting, I know that whatever I have decided is true regardless of whether the players know about it, so I don't need to track whether any of it has been declared yet as long as it's consistent with my underlying model.

Could be, but, on the other hand, if no-one remembers a detail you contradict, does it matter? That is part of the philosophy behind the Universalis mechanism for "consistency" - if no-one is bothered about the erstwhile "fact" turning out not to have been a fact at all, there isn't really any need to make it so. Maybe the characters misremembered? Maybe it was misunderstood? Maybe it was all a dream... (!)

 

[ExploderWizard]

This is why games written to this style (Burning Wheel, Apocalypse World and it's spin offs like Dungeon World and Monsterhearts, FATE to some extent, Dogs in the Vineyard for sure) don't tend to penalise failure particularly hard. Failure simply imforms the next scene. . . .

Finally, one tell-tale sign for this style is to watch your own GM-ing. Starting scenes is easy. But how are you ending them? And how are you moving on to something else? If you say 'What do you do now?' you are not playing using scene-framing. If you cut to a place where nothing is happening right now you are not playing using scene-framing. Scene-framing in the most aggressive sense means going 'bang!' - straight into the conflict, straight into the action. And again. And again. When it hums like this, as each scene unfolds everyone at the table is alive with ideas as to what the next scene will be be.[/indent]

"Fail forward" is a technique best-suited to scene-framing play, that is, play in which narrative dynamism is front-and-centre, and momentum is only rarely lost.

Ah. That clears things up quite a bit, thank you.

Machete don't frame scenes.

 

[Bedrockgames]

"Fail forward" is a technique best-suited to scene-framing play, that is, play in which narrative dynamism is front-and-centre, and momentum is only rarely lost.

Of course, in scene-framing play, the GM needs to listen to player action declarations: when the GM tells the player that the rattlesnake rears up in front of the PC, the player is expected to declare some action in response - ie they declare what it is that the PC does now.

I think this explains much of my confusion here. I have been struggling to understand Fail Forward (every time I think I have it, it slips away with new information). I definitely seem to take a much different approach to GMing from scene framing, so that is probably why I am struggling to see how Fail Forward would fit into one of my sessions.