Fail-forward combat

[Knaight]

In a very real way, making (a) new PC(s) to better respond to a lethal failure is "failing forward". Certainly more forward than forcing the players to keep their apparently ineffective PCs plus making those PCs worse by inflicting a permanent injury upon them.

Failing forward is more about continuity than anything - and while a new PC is going to likely be more able to mechanically contribute they don't have the history of an old PC. There's a discontinuity when they switch, and nowhere is this more apparent than in TPKs and near TPKs.

[kyoryu]

For me the obvious alternative to death is ... persistent injuries and setbacks.

When the PC hits zero, if they don't die, instead roll on a table, including things like fractured hand, eye injury, lose a foot, break a weapon, hireling dies, animal injured, lose 1d4 Str due to muscle strain, gain a madness, big scar, etc. I'd keep the chance of actually losing a limb very low, and restricted to forelimbs which can be replaced (hook hands, peg legs, etc, not whole leg, whole arm). This keeps combat nice and scary, with real consequences, without needing PCs to actually die. For a high magic game, change magic to make removing such injuries time consuming (eg 1d6 days with magic, or 2 months without).

I personally prefer plot-level setbacks.

But that goes against the desire for many people to have pre-planned, written out "stories" so meh.

[Nifft]

I actually like this better for that reason. Having a character develop flaws over time makes you more invested in that character, an optimal character is basically a race car and a somewhat damaged character who keeps on trucking is your high school car that you can't bring yourself to get rid of for a mini-van at 28.

Failing forward is more about continuity than anything - and while a new PC is going to likely be more able to mechanically contribute they don't have the history of an old PC. There's a discontinuity when they switch, and nowhere is this more apparent than in TPKs and near TPKs.

Here's where I'm seeing a useful definition: https://www.forbes.com/sites/mikemad...-fail-forward/

Failing forward seems to be about iteration and adaptation. It's about maximizing learning and minimizing the sorts of trauma and self-doubt which might reduce future success.

Increasing your failed investment isn't failing forward -- it's just embracing a sunk cost fallacy.

Continuity sounds like an anti-adaptation argument at best, anti-learning at worst. The players are the thing you want to optimize for, not the characters. Characters are useful only when they are useful. Making a player stick with a losing strategy is like staying in a bad marriage for the kids -- except in this case, only the kids are actual people. Don't punish real people for the sake of pretend ones.


Now there may be valid reasons to stick with one PC, to increase investment and continuity, but that reason is NOT "failing forward". Failing forward tells you to do the opposite -- try new things, recognize failures when they happen, learn from these failures, and then make changes based on what you learned. Don't enshrine a mistake just because you spent a long time making it -- that's just plain failing, not failing forward.

Again, there are probably solid reasons to do what you guys are advocating, but what you're advocating is NOT "failing forward".

[Knaight]

Here's where I'm seeing a useful definition: https://www.forbes.com/sites/mikemad...-fail-forward/

Failing forward seems to be about iteration and adaptation. It's about maximizing learning and minimizing the sorts of trauma and self-doubt which might reduce future success.

Increasing your failed investment isn't failing forward -- it's just embracing a sunk cost fallacy.

Continuity sounds like an anti-adaptation argument at best, anti-learning at worst. The players are the thing you want to optimize for, not the characters. Characters are useful only when they are useful. Making a player stick with a losing strategy is like staying in a bad marriage for the kids -- except in this case, only the kids are actual people. Don't punish real people for the sake of pretend ones.

Fail forward the RPG jargon is not the same thing as fail forward as listed as forbes - and the whole analogy where characters are sunk costs and/or misses the point of why the people who favor fail forward are playing RPGs spectacularly - which is generally more on the shared creative storytelling end, where the ups and downs of one group of characters that win some and lose some are interesting, and throwing a whole bunch of disposable characters at a challenge totally defeats the point of the style.

[Tvtyrant]

Here's where I'm seeing a useful definition: https://www.forbes.com/sites/mikemad...-fail-forward/

Failing forward seems to be about iteration and adaptation. It's about maximizing learning and minimizing the sorts of trauma and self-doubt which might reduce future success.

Increasing your failed investment isn't failing forward -- it's just embracing a sunk cost fallacy.

Continuity sounds like an anti-adaptation argument at best, anti-learning at worst. The players are the thing you want to optimize for, not the characters. Characters are useful only when they are useful. Making a player stick with a losing strategy is like staying in a bad marriage for the kids -- except in this case, only the kids are actual people. Don't punish real people for the sake of pretend ones.


Now there may be valid reasons to stick with one PC, to increase investment and continuity, but that reason is NOT "failing forward". Failing forward tells you to do the opposite -- try new things, recognize failures when they happen, learn from these failures, and then make changes based on what you learned. Don't enshrine a mistake just because you spent a long time making it -- that's just plain failing, not failing forward.

Again, there are probably solid reasons to do what you guys are advocating, but what you're advocating is NOT "failing forward".

Hmm, I can see your point ablut definitions. I suppose constructive failing might be a better term.

If I play a shooter where I come back instantly after death I am going to respond differently than if there is a distant savepoint, and way differently from a perma death like Fortnite. As death has more consequences I become more cautious.

Injuries and scars have as much effect or more on my willingness to be reckless or cautious as dying and spawning a new person, while also being better IMO for connecting the player with their character. Collab story telling is my biggest goal in playing RPGs.

I also wouldn't force anyone to keep a character they no longer want. If they feel like rerolling someone without a eye missing or a peg leg I would be on board, I just don't want changing characters to be a consequence of mechanical effects.

[Nifft]

Spoiler: An Earlier Post

Show

The conflict here is due to the requirement that deadly combat must exist in the same game as pre-ordained victorious heroes.

You can do things old school, and have normal mortal PCs who live & die as the dice-gods dictate. The victorious heroes are the PCs at the end of the campaign, who might not have existed at the beginning.

OR you can do things literary-style, where the main characters of the story survive all tribulations through to the end. The main characters faced off against the appearance of risk, but there was no actual risk because their ultimate success was written in advance.

If you try to have both at once, you encounter the OP's conflict.


Both styles are viable, though the former is more easily available in a game with deadly combat rules, and the latter is more easily adapted to a fixed story (like a book or a video game).

The questions you need to ask are:
- Do I want a game with PC death as a risk?
- If not, what am I willing to put at risk?
- If so, how specifically should the risk of death be signaled?

Fail forward the RPG jargon is not the same thing as fail forward as listed as forbes - and the whole analogy where characters are sunk costs and/or misses the point of why the people who favor fail forward are playing RPGs spectacularly - which is generally more on the shared creative storytelling end, where the ups and downs of one group of characters that win some and lose some are interesting, and throwing a whole bunch of disposable characters at a challenge totally defeats the point of the style.

So you're talking about games with pre-ordaned designated victorious heroes, and the stories of the struggles & failures & eventual success that gets them from session zero to their inevitable victory.

That's cool, but as mentioned in An Earlier Post that's incompatible with ye olde schoole lethal-combat games.

I don't think of that game style as fail forward, though. It seems very confusing to use that as a label for a game that's about:
- Non-Lethal Risks
- Designated Champions
- Guaranteed (Eventual) Success

By some metrics, you cannot fail in such a game. Clearly those metrics are irrelevant if the games are "spectacular", so it seems a lot more useful to define these games in positive terms which don't obfuscate their method.

Hmm, I can see your point ablut definitions. I suppose constructive failing might be a better term.

If I play a shooter where I come back instantly after death I am going to respond differently than if there is a distant savepoint, and way differently from a perma death like Fortnite. As death has more consequences I become more cautious.

Injuries and scars have as much effect or more on my willingness to be reckless or cautious as dying and spawning a new person, while also being better IMO for connecting the player with their character. Collab story telling is my biggest goal in playing RPGs.

I also wouldn't force anyone to keep a character they no longer want. If they feel like rerolling someone without a eye missing or a peg leg I would be on board, I just don't want changing characters to be a consequence of mechanical effects.

Let me try to think of a game that I think might be included in your description.

How about FATE (specifically Dresden Files RPG)? In that game, your character won't die mechanically. You'll have a chance to suffer "consequences" which have impact for a number of scenes or a span of game time, but which you can recover from completely given sufficient time and/or resources. If you "die" in an encounter, you're "taken out" (you don't contribute further to the encounter, and you don't get to help dictate the conflict's eventual resolution). You can give yourself a Consequence of varying duration instead of being "taken out", so you should only take a Consequence when you think you've got a shot at taking out your opposition.

If you get taken out, you lose -- and in the case of a TPKO, the GM dictates the conflict outcome unilaterally -- but you gain Fate Points for being taken out, so you gain an edge for your next conflict.

Hmm, I think FATE / DFRPG may suit both the general definition and the one that you guys are using.

[Knaight]

So you're talking about games with pre-ordaned designated victorious heroes, and the stories of the struggles & failures & eventual success that gets them from session zero to their inevitable victory.

Not at all - there's no guarantee of eventual success. Continuity is not the same thing as victory, and it could easily end up being about how they fail in their goals and then are slowly ground down running ineffective damage control until they're eventually dealt with for good by their victorious enemies. They fail, the failure persists, then the game moves forward, through different events than it would have moved through in the context of a success.

Meanwhile being able to substitute in new characters until they're successful is much closer to the designated victorious heroes style, as it's just a matter of which heroes end up in that slot. Similarly it's pretty easy to define metrics where you can't fail here, as you just keep getting more tries.

[Nifft]

Not at all - there's no guarantee of eventual success. Continuity is not the same thing as victory, and it could easily end up being about how they fail in their goals and then are slowly ground down running ineffective damage control until they're eventually dealt with for good by their victorious enemies. They fail, the failure persists, then the game moves forward, through different events than it would have moved through in the context of a success.

That sounds spectacularly depressing.

Meanwhile being able to substitute in new characters until they're successful is much closer to the designated victorious heroes style, as it's just a matter of which heroes end up in that slot. Similarly it's pretty easy to define metrics where you can't fail here, as you just keep getting more tries.

Hmm, it looks like you think "designated victorious heroes" is a term of derision, so you're trying to apply it to (what you perceive as) "my side". It's not, and it's not.

1) Ye olde school high-lethality games are not "my side", and they're not claimed to be better -- it's merely a style that is incompatible with "designated victorious heroes". That's why I contrast them.

2) Under lethality + new PCs, if the new PCs have different goals, then abandoning a failed goal isn't failing forward -- and it isn't continuity except insofar as players desire continuity. The point of failure demands that the player decide on goals for the new PC, and that might be a different ("discontinuous") direction.

You're not throwing more bodies at Gandalf's sub-plot after Gandalf leaves the party. You focus on what the other PCs are into.

[Knaight]

That sounds spectacularly depressing.

Sometimes bleak and hopeless is an aesthetic that can be enjoyable, if not necessarily fun qua fun (though beer and pretzels bleak and hopeless can be fun). After all character focused stories in other media are routinely tragic. That said, outside of the sort of games outright pitched as "you'll be attempting a futile, symbolic gesture doomed to failure" that tends to be more a hypothetical worst case scenario than what actually happens, which is more a mixed end than anything.

Hmm, it looks like you think "designated victorious heroes" is a term of derision, so you're trying to apply it to (what you perceive as) "my side". It's not, and it's not.

1) Ye olde school high-lethality games are not "my side", and they're not claimed to be better -- it's merely a style that is incompatible with "designated victorious heroes". That's why I contrast them.

I have no issue with the designated victorious heroes approach - I'm just saying that to a large extent the replacement character approach often fits them better, particularly if you substitute in comparable characters. It's analogous to the roguelike genre of video games, and the thing about them is that if you play enough you'll eventually win. It's a totally viable design, if not one I have any interest in with this particular medium.

You're not throwing more bodies at Gandalf's sub-plot after Gandalf leaves the party. You focus on what the other PCs are into.

Sure, but that's still going to involve Sauron - and at least in my experience the high lethality games are disproportionately likely to be main plot focused instead of character focused, precisely because of the character turnover.

[Psikerlord]

Heh, good point.

Usually the penalty for utter failure is ... you get to make a new PC, with perfect custom magical gear, who can better respond to whatever killed your previous PC.

Without PC death, the penalty for utter failure is... you must keep your PC who couldn't survive the previous frustration, except now your PC is worse.


In a very real way, making (a) new PC(s) to better respond to a lethal failure is "failing forward". Certainly more forward than forcing the players to keep their apparently ineffective PCs plus making those PCs worse by inflicting a permanent injury upon them.

I dont know if this is actually that common. When you die in our game, you start again at level 1, but advance twice as quickly till you catch up. No free magic items, etc.

[Psikerlord]

I personally prefer plot-level setbacks.

But that goes against the desire for many people to have pre-planned, written out "stories" so meh.

Well you can have both injuries and plot setbacks. Neednt be exclusive.

[Spore]

-Someone else dies. They failed to beat the dragon, and as a result, it burns down the village.

We played out the deciding battle in a war by skill checks, attack rolls and somesuch. It was pretty free form but each individual check meant something good or bad for a major NPC or large groups of soldiers. If we screwed up, we would have even lost (our country winning the war was not vital for the plot).

Here is what the DM did.

"My attack roll was a 17."
"The enemy commander got a 27. Sorry. Your battalion is dispersed and they take the opportunity to kill the old cleric in the midst of it, responsible for your soldier's resilience."

Many tears were shed because a good villain (and a good DM) usually threatens the things PCs like and love, and not the PCs themselves - unless they are heartless bastards and/or chaotic stupid.

[Nifft]

Sure, but that's still going to involve Sauron - and at least in my experience the high lethality games are disproportionately likely to be main plot focused instead of character focused, precisely because of the character turnover.

I started out playing in what I'd describe as a High-Lethality Sandbox, with the [Hex-Crawl] subtype. The world reacted to the characters, who were generally explorers and treasure-hunters and bandits (ostensibly cut from the cloth of Robin Hood or Zorro).

If you're on an Adventure Path, then I guess your path is going to persist no mater how many bodies you throw at it. But that's not the only style of lethal game.


Hmm, maybe Gandalf was a bad example.

How about Boromir? His conflict about wanting to use the Ring for humanity died with him, and the replacement PC (Gollum) brought a very different conflict -- though it was clearly the same player, and you can tell because of how both PCs tried to steal the Ring.

I dont know if this is actually that common. When you die in our game, you start again at level 1, but advance twice as quickly till you catch up. No free magic items, etc.

Hmm, that's surprisingly similar to Knaight's analogy of Roguelike games.

Using your rules, a player does get to rebuild her PC after learning more about the world, but putting the lessons into practice isn't quite as easy as getting to start as a level-appropriate character.

[Knaight]

I started out playing in what I'd describe as a High-Lethality Sandbox, with the [Hex-Crawl] subtype. The world reacted to the characters, who were generally explorers and treasure-hunters and bandits (ostensibly cut from the cloth of Robin Hood or Zorro).

If you're on an Adventure Path, then I guess your path is going to persist no mater how many bodies you throw at it. But that's not the only style of lethal game.

It's not the only style at all - but I'd argue that the specific context of the improvement of character building being desirable because a specific encounter went poorly makes a lot more sense in the context of that encounter effectively being repeated. A new character doing other things dodges that entirely.

Hmm, maybe Gandalf was a bad example.

How about Boromir? His conflict about wanting to use the Ring for humanity died with him, and the replacement PC (Gollum) brought a very different conflict -- though it was clearly the same player, and you can tell because of how both PCs tried to steal the Ring.

His conflict does, but it still fit within the parameter of the overall quest structure - even if you don't take the perspective that Boromir was on the same quest as everyone else until the ring basically killed him via manipulation he's still mostly working with the fellowship.

Hmm, that's surprisingly similar to Knaight's analogy of Roguelike games.

Using your rules, a player does get to rebuild her PC after learning more about the world, but putting the lessons into practice isn't quite as easy as getting to start as a level-appropriate character.

It occured to me after I wrote it that the analogy fit better for that sort of thing than where I used it, given that roguelikes generally involve losing your progress, for all that the game persists.

[Silva]

This is already present in Apocalypse World in a way: Every time you fill your health clock, you may, instead of dying, take a debility that permanently reduces a stat (crippled reduces Hard, disfigured reduces Hot, etc).

I think 2e changed that though. Don't remember how.

[Koo Rehtorb]

This is already present in Apocalypse World in a way: Every time you fill your health clock, you may, instead of dying, take a debility that permanently reduces a stat (crippled reduces Hard, disfigured reduces Hot, etc).

I think 2e changed that though. Don't remember how.

When life becomes untenable:
come back with -1hard
come back with +1weird(max+3)
change to a new playbook
die

[Conaldar]

Don't roll dice.

Dice control outcomes in D&D. It's how the game works. I've been running games since the early 80's and no matter how proficient a party was, the dice could betray them and end their adventure. Dead stop. Usually when you, as G/DM least expect it.

A more narrative approach works: just RP combat. That allows a clear path to the sort of outcome you're wanting. You can play D&D anyway you want based on group agreement (outside of Adventure League/Tournament play).

But rolling that d20, based on what you want, is counter-productive.

[Knaight]

Don't roll dice.

Dice control outcomes in D&D. It's how the game works. I've been running games since the early 80's and no matter how proficient a party was, the dice could betray them and end their adventure. Dead stop. Usually when you, as G/DM least expect it.

A more narrative approach works: just RP combat. That allows a clear path to the sort of outcome you're wanting. You can play D&D anyway you want based on group agreement (outside of Adventure League/Tournament play).

But rolling that d20, based on what you want, is counter-productive.

Dice work just fine for this - the way D&D happens to use them isn't ideal, but neither is just RPing combat if you want to have both the options of failing forward and success open, and to allocate that decision to a rules set with built in uncertainty. That doesn't have to be dice, but they're a great way to do it.

[kyoryu]

So you're talking about games with pre-ordaned designated victorious heroes, and the stories of the struggles & failures & eventual success that gets them from session zero to their inevitable victory.

No.

That's cool, but as mentioned in An Earlier Post that's incompatible with ye olde schoole lethal-combat games.

That's fairly true. These games typically don't have one-shot kills or other things. The games are more about finding out what happens in the world than not dying. That said, since 99% of the "fudging yay!" posts use "character death" as a reason to fudge, I doubt that the actual death rate in most games is that different.

I don't think of that game style as fail forward, though. It seems very confusing to use that as a label for a game that's about:
- Non-Lethal Risks
- Designated Champions
- Guaranteed (Eventual) Success

Well, you'd certainly have non-lethal risks in addition to lethal variety. Typically the games focus on in-world consequences to characters vs character death, though character death will be on the table in appropriate situations.

Guaranteed (Eventual) Success is not a thing AT ALL. Most of the games using Fail Forward are very much in favor of letting people deal with the consequences of the failure. You fail to stop the portal from opening? Cool. Now you're in a demon infested world, and the game has changed to surviving in that world. Have fun with that, chumps.

Designated Champions can be a thing, but isn't necessarily, any more than it is in D&D.

What Fail Forward really means is that, on a failure, something happens. What you don't see is:

"I pick the lock."
"You fail. What do you do?"
"I pick the lock again."
"You still fail. What do you do?"

and so on and so forth. In a fail-forward game, the presumption is that if you have infinite time and resource you will succeed - so what does failure mean? What happens that you don't continue? What resource or time limitation do you run up against? With fail forward, you'd do something more like this:

"I pick the lock"
"It's taking longer than you'd like. As you're working on it, a pair of guards comes around the corner".

Depending on feel, you could just have the guards heard but not seen, allowing them to react, etc.. Or just have them show up. In either case, due to failure, their situation has become worse.

How about FATE (specifically Dresden Files RPG)? In that game, your character won't die mechanically. You'll have a chance to suffer "consequences" which have impact for a number of scenes or a span of game time, but which you can recover from completely given sufficient time and/or resources.

Partially true.

First off, Consequences represent non-permanent impacts that you have - something that many systems don't bother with *at all* outside of death. So in most cases, you're likely to come out of a Conflict with some amount of lasting pain that will impact you far more than mere hp - and lasting pain that can't be mitigated until the appropriate time is passed. Apart from death, that's often more lasting pain than most other systems.

Secondly, there are Extreme Consequences, which never completely heal. They might be mitigated to some extent, but they negatively and permanently change your character.

If you "die" in an encounter, you're "taken out" (you don't contribute further to the encounter, and you don't get to help dictate the conflict's eventual resolution). You can give yourself a Consequence of varying duration instead of being "taken out", so you should only take a Consequence when you think you've got a shot at taking out your opposition.

Taken Out can absolutely mean dead. It doesn't have to, but it can. If you're Taken Out, the player that took you out, including the GM, has unilateral say on what happens. That can absolutely include death.

It doesn't have to, and because of that, it frees the GM to play very, very hard if they want. Since a PC can never be "accidentally" killed, the GM can pretty much go as hard as they want. Failure is super common in Fate games - to the point that I have to explain that explicitly to new players to Fate.

If you get taken out, you lose -- and in the case of a TPKO, the GM dictates the conflict outcome unilaterally -- but you gain Fate Points for being taken out, so you gain an edge for your next conflict.

No. If you are Taken Out, you're just Taken Out, and the GM decides what happens to you (possibly including death). If you bow out of the conflict *before* you're Taken Out, you can avoid parts of it, but the bad guys get what they want. That's the case where you can get a Fate Point or two to help in the future.

So, no, a game like Fate isn't really based on "guaranteed eventual success". It's not easy mode. You "lose" fights/conflicts/whatever usually 1/3-1/2 of the time. I screw with my players in Fate games really, really hard, and when I've asked them if they thought my games were easy mode (and these are long-term, traditionally minded players) they laughed in my face.

I can understand your assumptions - I actually shared them myself before someone on this very board corrected me, and got me to investigate the games more. And I found I liked what I saw - not because they were easy, or allowed "the story" (screw "the story") to go along the designated path, but because they gave players choice and let me cause so, so much pain to them.

[Nifft]

No.

Hey Knaight - this guy apparently wants to answer questions directed to you.

Is this guy your designated spokesperson?

[Knaight]

Hey Knaight - this guy apparently wants to answer questions directed to you.

Is this guy your designated spokesperson?

No, but we have sufficiently similar stylistic preferences and jargon that their answers all work fine for me. I'm also far from the first person to use the terminology in this thread, so those questions directed at me are all pretty open as far as I'm concerned.

[Silva]

When life becomes untenable:
come back with -1hard
come back with +1weird(max+3)
change to a new playbook
die

Thanks! Hmm that's interesting. I wouldn't expect that Weird increase, but it seems to fit the implicit setting well.

[Psikerlord]

Don't roll dice.

Dice control outcomes in D&D. It's how the game works. I've been running games since the early 80's and no matter how proficient a party was, the dice could betray them and end their adventure. Dead stop. Usually when you, as G/DM least expect it.

A more narrative approach works: just RP combat. That allows a clear path to the sort of outcome you're wanting. You can play D&D anyway you want based on group agreement (outside of Adventure League/Tournament play).

But rolling that d20, based on what you want, is counter-productive.

I am firmly in the opposite camp. I want the dice to decide what happens next. For me that is where the fun is, for me gameplay > plot. There will still be a great story, but it will be an improvised one, partly from the dice, partly what the PCs choose to do. If the party survives, that means something in and of itself - they earnt it.

[Mechalich]

I am firmly in the opposite camp. I want the dice to decide what happens next. For me that is where the fun is, for me gameplay > plot. There will still be a great story, but it will be an improvised one, partly from the dice, partly what the PCs choose to do. If the party survives, that means something in and of itself - they earnt it.

Using dice to produce consequences also has a conflict reduction element at most tables. The GM doesn't want to be arbitrarily murdering off PCs, that's going to sound mean no matter how deserved it is, to the point that it is usually worthwhile to actually get out the dice and roll even if the result is mathematically certain to produce death (ex. rolling Xd6 damage where x > than the character's hp).

It most situations, as the GM, you probably have a small number of scenarios for outcomes, ie. good, neutral, or bad, and the general range of both player choices and dice outcomes should dictate where things land. The players might plan really well but roll poorly and end up with a neutral outcome, or plan badly but roll well and still get the good outcome, but extreme results should generally only occur if the players do something extremely unexpected (this is usually bad, simply because the number of catastrophically stupid things one can do tends to outnumber the corresponding number of unbelievably awesome ideas out there) or if they dice produce a spectacularly unlikely result. in d20 this isn't even particularly unlikely since a two critical success or two critical failures in a row each have a 1 in 400 chance of happening, meaning it's likely to occur at least once a campaign.

[Silva]

I am firmly in the opposite camp. I want the dice to decide what happens next. For me that is where the fun is, for me gameplay > plot. There will still be a great story, but it will be an improvised one, partly from the dice, partly what the PCs choose to do.

Yep, and that's exactly what Fail Forward does.

[Max_Killjoy]

Someone might have said something along these lines already, but...

From what I've read in these threads, there are two kinds of "fail forward", and one is a lot less useful than the other.

One is "the game should have ways to continue forward even if the PCs fail at an individual task or at a whole scheme". The GM should have contingencies and/or improvisation on hand such that the game isn't stalled by the failure of a single roll ("Oh, goody, we're still here trying to pick this lock for the 37th time.") or a plan ("Well, we couldn't bluff the guards to get into the castle, what now?"). This one, I can get behind. Don't stall the fun, keep moving in a generally onward direction.

The other is when there's so much determination to move "down the road" that success and failure don't matter at all, the outcome is functionally the same regardless. Ends up looking a lot like railroading.

[Nifft]

What Fail Forward really means is that, on a failure, something happens. What you don't see is:

"I pick the lock."
"You fail. What do you do?"
"I pick the lock again."
"You still fail. What do you do?"

and so on and so forth. In a fail-forward game, the presumption is that if you have infinite time and resource you will succeed - so what does failure mean? What happens that you don't continue? What resource or time limitation do you run up against? With fail forward, you'd do something more like this:

"I pick the lock"
"It's taking longer than you'd like. As you're working on it, a pair of guards comes around the corner".

Depending on feel, you could just have the guards heard but not seen, allowing them to react, etc.. Or just have them show up. In either case, due to failure, their situation has become worse.

No. You're making a distinction where there is no difference, or you're talking about rail-roading on failure ("failroading"?). Neither of those is remotely fail-forward.

Here's a more detailed look at your first case:

Player: "I pick the lock."

DM: That took one time unit. Dice are rolled behind the DM's screen. There is risk! Danger! Uncertainty! You don't know if you can afford to take more time doing this! "You fail. What do you do?"

Player: Guys, we only have six time units in the castle before our excuse for being here runs dry, and we need at least 3 of them for the main operation and extraction. We can afford one loss here but we can't afford another. "I pick the lock again."

DM: Heh, now you realize why Count Muddlespont sneered at your threats. This castle's security is above your skill level. You start to doubt your ability to get away clean even if you did succeed. Also I'mma roll for wandering monsters and smirk about the result. "You still fail. What do you do?"

Player: Damn it, we're down to four time units! Oh my gods we're going to have to do something drastic!

DM: Now I'mma roll Hide vs. your passive Spot...

Instead of your "try-once-then-combat" Failroading™ thing, the more open approach allows the DM to just let the world react to whatever the PCs try to do. If they keep trying the same thing, the world may react the same way each time -- or it may react differently, depending on what the dice say.

Having random encounters, and having success be a thing that arises naturally from the actions of the PCs rather than being a way-station on your railroad of plot -- that's part of the game for me.

Taken Out can absolutely mean dead. It doesn't have to, but it can. If you're Taken Out, the player that took you out, including the GM, has unilateral say on what happens. That can absolutely include death.

It doesn't have to, and because of that, it frees the GM to play very, very hard if they want. Since a PC can never be "accidentally" killed, the GM can pretty much go as hard as they want. Failure is super common in Fate games - to the point that I have to explain that explicitly to new players to Fate.

Technically somewhat correct, but misleading: I'm discussing the difference between baseline D&D, where losing a fight is far more likely to result in death.

Repeating how FATE "taken out" can also result in death is downplaying the fact that the D&D equivalent is far more heavily slanted towards death.

No. If you are Taken Out, you're just Taken Out, and the GM decides what happens to you (possibly including death). If you bow out of the conflict *before* you're Taken Out, you can avoid parts of it, but the bad guys get what they want. That's the case where you can get a Fate Point or two to help in the future.

Wrong. You're not the only person at the table. You getting taken out means you won't participate in the resolution -- it doesn't mean no other player can participate, only you are excluded by having your PC taken out.

It's quite possible for your PC to get "taken out" and then for the rest of the group to retreat (shouldering your unconscious bulk or whatnot).

So, no, a game like Fate isn't really based on "guaranteed eventual success". It's not easy mode. You "lose" fights/conflicts/whatever usually 1/3-1/2 of the time. I screw with my players in Fate games really, really hard, and when I've asked them if they thought my games were easy mode (and these are long-term, traditionally minded players) they laughed in my face.

I can understand your assumptions - I actually shared them myself before someone on this very board corrected me, and got me to investigate the games more. And I found I liked what I saw - not because they were easy, or allowed "the story" (screw "the story") to go along the designated path, but because they gave players choice and let me cause so, so much pain to them.

I'm talking about eventual success without the risk of PC death, which is not the same thing as "easy mode".

This "easy mode" thing seems to be your straw man. If you were genuinely confused about the distinction between "easy mode" and what I'm talking about, let me know.

Anyway, hope that helped a bit.

[Silva]

@Nifft, I honestly don't know the concept of Fail Forward as you put it, and find Kyoryu is spot on with his definition. Failing Forward has nothing to do with "Designated Champions", it simply means keeping the game state moving/changing, instead of coming to a halt. The lockpick example is perfect, by the way.

What game have you seen describe the concept? Because the ones I've read or played (Powered by the Apocalypse games, Mutant Year Zero, Fate, Burning Wheel) all fall under Kyoryu definition.

[Nifft]

@Nifft, I honestly don't know the concept of Fail Forward as you put it, and find Kyoryu is spot on with his definition. Failing Forward has nothing to do with "Designated Champions", it simply means keeping the game state moving/changing, instead of coming to a halt. The lockpick example is perfect, by the way.

Which one is perfect -- Kyoryu's Zork example, or my more plausible table-top re-write?

What game have you seen describe the concept? Because the ones I've read or played (Powered by the Apocalypse games, Mutant Year Zero, Fate, Burning Wheel) all fall under Kyoryu definition.

I found the concept defined as a business & engineering practice pre-dating those games.

The games I've seen discussed so far seem to have tried to implement that practice, and gotten parts right, but also gotten parts wrong.

[NichG]

The business terminology arises from trying to understand how to design strategies for playing the game successfully in a universe which has to be taken as given. That isn't really relevant to the usage in gaming, where the problem is inverted: design the universe such that it gives rise to the game that we'd like to play.

The point of fail forward design in terms of game and scenario design has to do with challenging the perspective of formulating a game as a series of instances of obstacles with checks (either simple, like a literal die roll, or elaborate like a combat encounter) to see whether or not the characters can pass them. Such a design has two flaws - one, in that it is possible for checks to become irrelevant if they revert state on failure, because they may be tried an arbitrary number of times until success is statistically guaranteed; the other, in that it is possible for the game to enter a dead end state, in which there are no viable moves forward (at which point, the players+DM will inevitably fiat a continuation of the activity in the form of e.g. scrapping everything and rolling new characters, starting a new campaign, etc).

So instead of designing things from the perspective of pass-and-continue, fail-and-return-to-state type interactions, a fail-forward design philosophy focuses on forks - each point at which something is decided (via whatever mechanism) should not question 'does the game continue?', it should only and entirely question 'in which direction does the game continue?' taking continuation as a strict given. This generally means that you reduce the number of times that the system/DM/players engage with the mechanics to ask a spurious question, e.g. one where there's either only one inevitable answer or one acceptable answer.

The pre-existing business term isn't really meaningful here.