Fail Forward


Pathfinder Second Edition General Discussion

201 to 250 of 281 << first < prev | 1 | 2 | 3 | 4 | 5 | 6 | next > last >>

2 people marked this as a favorite.
Pathfinder Roleplaying Game Superscriber
PossibleCabbage wrote:

I personally let the players set a great number of the details themselves when they opt to, even if doing so gives them an advantage, provided they are being at all reasonable. Since one of the main jobs of a GM is to describe the world in a way that creates a mental picture that is shared by all. If we change "imagining the world" from a passive task (i.e. "listen to the GM and try to follow along") to an active one (i.e. "try to think of what sort of thing could exist here which would be helpful") I find it works much smoother.

Like RPGs are fundamentally "asymmetrical improv" with a bunch of systems grafted on, but there's no reason they can't be less asymmetrical. So if my players want to posit the existence of a convenient dark alley, secret tunnel under some ruins, a nearby hamlet, a chandelier rope, a scullery maid, etc. then I'm happy to confirm they are correct.

One of my favorite things about Exalted 3e is that they codified this in the form of players being able to make a "Fact Introduction" roll. Basically they make a knowledge check relevant to the fact they want to introduce, and if they succeed that thing - subject to GM interpretation - becomes canon.

So instead of "I succeeded my knowledge check; what fashion is popular in Nexus right now?" it becomes "I'm going to roll Fact Introduction; my character has heard that long sleeves and floral patterns are all the rage in Nexus lately."

@DM_aka_Dudemeister: 100% all of this. "Each roll should change the game state in some way" is exactly how I see F>F best working.


5 people marked this as a favorite.
Pathfinder Pathfinder Accessories Subscriber; Pathfinder Roleplaying Game Superscriber

Edit: Quick TLDR; my basic position is that John Lynch 106's definition is fine, but he's selectively ignoring it when it describes something that aligns with how he already does things because that's not how he got to the same result.

Quick semantic pedantry:
John Lynch 106 wrote:

The definition I could find is this

Quote:

Failing forward is the idea that you still get to unlock the door on a failed roll, but it comes at a cost.

That isn’t how skill checks are described in the rules for the playtest core rules.

Okay, first of all, that definition literally only covers doors. If we're going to play the definition/semantics game, we should call that out (because that's what rules lawyers do). Really, though, the important takeaway is that it isn't how skill checks are described by the rules, and that's completely true.

Anyway, "Fail forward" was redefined later in the thread by John as:
John Lynch 106 wrote:
Fail forward is when you allow the PCs to succeed with the same approach even though they failed to overcome the challenge with that approach. Often with a loss of resources or a minor obstacle that will take up table time but not meaningfully hinder their ability to reach the end goal.

This is the definition that we'll work with, because it's actually fine for my purposes. I disagree with John Lynch 106's interpretations of that definition, and I think that's where some of the disconnect is between some of the views here among the mechanically-oriented posters here. I'll try to avoid introducing any new examples to make my points.

For folks using other definitions of fail-forward, many depending on narrative meta-mechanics, I'm not really addressing these here, even later when I talk a little about designer intent. Suffice to say that I don't agree that fail-forward design requires these meta-mechanics or expansions of the definition beyond what John found/described.

John Lynch 106 wrote:
Ventnor wrote:
Also, not sure what the dwarf story added to the discussion. That solution can occur in any game, fail forward rules or no.
In a fail forward game this is what happens.
Quote:

Player: I go up to the guard and ask him to let us through.

GM: Roll a Persuasion check.
Player: I get a 2.
GM: He lets you through even though despite you failing the skill check. Scratch off 50 gp.
No amount of clever thinking is necessary. Because regardless of whether or not the players succeed in their approach to overcome an obstacle, fail forward means they'll always get to the next stage....
John Lynch 106 wrote:
Ventnor wrote:

I will disagree and say that failing forward is good, because it means something interesting always happens when you roll dice. If the PCs roll low on a survival check to track an enemy down, it doesn’t necessarily mean that they’ve lost them. But perhaps the tracker was so focused on finding the enemy’s trail that they missed the signs that a vicious owlbear calls this territory home.

And so on and so forth.

So the DC wasn't to "find and follow the tracks". The DC was to "look out for trouble while following this very clear and obvious path". It sounds like the GM just used the wrong table to determine the DC. It happens. GMs are only human after all. I've certainly stuffed up plenty of times and played with GMs who messed up. No-one holds it against the GM.

John, here's my problem with your interpretation of your own definition of fail forward. You keep busting out the No True Scotsman fallacy on the valid examples. Every time someone introduces a "failure at a cost" mechanic, if it's implemented in a way you like, you dismiss it as not being fail-forward, but instead just an incorrect labeling of the challenge/DC. If you don't like the implementation, you call it bad adventure design. It's all a matter of framing, and that has nothing to do with the mechanics except that people are sometimes framing things in ways you don't like.

For the tracking example, it is a literal example of a fail-forward mechanic. Maybe the DC could be better described as avoiding danger while following the trail, but that doesn't change what it is, which is a roll where no matter what happens, the PCs continue on the trail of their quarry while instead suffering tangential costs. I'm not super fond of this as a universal example either (I wouldn't personally use it as described for the "expert scout" that could and/or should be hard to track), but it's clearly fail-forward.

In your own example of the guards, I could take your previous argument and say that the DC was actually to convince the obviously corrupt guards to let you through without a bribe. The adventure could have printed: "One nearly fool-proof method to get through the wall is to bribe the guards, and a diplomacy check modifies the bribe cost, but the guards routinely take bribes (which can be learned from random townsfolk on a DC 15 gather information check, or the guards will suggest there is an 'expedite fee' to gain access) and therefore will always let the party through assuming they have sufficient gold." That's, by your definition, a fail-forward mechanic. It's literally your example. Is that inherently bad adventure design? What if there are other methods with different potential costs? What if a critical failure prevents the guards from taking the bribe because they think the PCs are working for their boss? What if a critical failure just means that those same guards report the PCs as suspicious which increases the challenge of a later encounter?

One more example:

John Lynch 106 wrote:
Quote:

Player: I try to unlock the door.

GM: Give me a disable device check.
Player: I get a 3.
GM: You fail to unlock the door which was enchanted with a lightning bolt spell to trigger upon detecting a failed attempt to unlock the door. You take 28 electricity damage and are deafened for 3 rounds.
Player: I check the door for traps. This time I take my time and take 20.
GM: You discover the lightning bolt trap and see that the enchantment was a one time use.
Player: I try to unlock the door.
GM: It takes a while but you manage to successfully unlock it.

That isn't fail forward. That's "Once the consequences for failing are removed, I simply keep trying until I succeed." Unless there's a consequence for failing (e.g. the party has buffs that will expire) there's no need to force the player to keep generating random numbers until they generate the right one. You can just narrate it taking a while and then they succeed.

That's not a fail forward mechanic. That's the "take 20" mechanic from the PF1e CRB.

How is that not a fail-forward mechanic? "Fail forward is when you allow the PCs to succeed with the same approach even though they failed to overcome the challenge with that approach. Often with a loss of resources or a minor obstacle that will take up table time but not meaningfully hinder their ability to reach the end goal."

It's literally: You will be able to open the door, it might just cost some HP (with extra steps). Frankly, almost all traps are fail-forward mechanics by your definition.

And really, here's the thing (for John and others that viscerally hate fail-forward design). When game designers talk about failing forward, particularly with non-binary success mechanics, they largely are giving pretty generic advice about avoiding a basic failure in adventure design for people who aren't you and haven't already internalized the good methods without ever using the term for it. Multiple avenues for success with different costs and benefits are also good design (just because the guards will always take bribes doesn't mean you can't tunnel under the wall). Having one super reliable way to continue the adventure isn't inherently a bad thing. If you feel that it ruins creativity, the fix is straightforward: change the incentives. If the players bribe everyone, maybe the costs keep going up as the party gets a reputation as easy marks. Maybe you make it clear they can't afford key items later. Maybe the cost is actually their favorite weapon. Just because there is a cost for "guaranteed success" doesn't mean it has to be a small one.

I'm a simulationist as a GM, which seems awfully close to the absolutist people are chatting about. I think John Lynch 106 and probably DeadManWalking have the same or very similar tendencies here. I don't generally tweak DCs and try to run every mechanic as printed, because I want players to be rewarded for understanding how the rules and the game universe work and get that sense of satisfaction when they think of a cool idea that makes sense and it just works. I also like when actions, even those resulting in failure, have predictable consequences. I don't agree with the folks declaring that every roll has to have interesting outcomes, because some failures aren't supposed to be interesting (and if success is obviously impossible I'll generally tell the player not to roll or let them roll and immediately tell them it fails and they don't see a way to make that approach succeed). There are exceptions here, too, where the point of the roll is to keep information secret (player asks for a check to tell if someone is lying; I'll go ahead and roll and say the person seems sincere). I think player agency is important, and if it makes sense within the world, and doesn't obviously violate game mechanics or introduce balance issues, I'll allow just about anything but impose appropriate consequences. I don't like the weird "a convenient maid will be nearby, but only for the purposes of following a failed check" answers for failing forward either, unless that's just the daily routine and not being added literally as a railroady crutch. But either I or my players will come up with ways that make sense and don't violate the fabric of the world or the adventure, so it hasn't been much of an issue. This is, I suspect, where John Lynch 106 can be both a talented, enjoyable GM and also one who enforces DCs with an iron fist. I tend to not design lots of obstacle-bypass methods in great detail in advance, because I rely on the pre-existing DCs for the world to do much of that heavy lifting for me. The difference for me, as I think it is for John Lynch 106 here, is that the existence of a convenient window, or maid, or whatever it is that I suddenly ruled is there is because it makes sense for it to be there, and if I was asked in a different context I would also say it was there, I'm not deciding the thing is there because it's required to allow the players to proceed.

Lengthy aside/example from my own game:
Many many years ago, running a 3.5 campaign with an evil party, the entire campaign arc was a consequence of the PCs suddenly attacking an NPC succubus (who was genuinely on their side), who immediately teleported away, then caused shenanigans with shapeshifting and general faster travel methods. There were so many fail-forward mechanics in that campaign, although I wouldn't have described them as such at the time, since that wasn't a term I'd heard.

One time, the party fell in a hole (literally, just a big hole, I think formed with a few castings of move earth, or maybe a lyre of building), and lacked climbing equipment/skills (plus were wearing heavy armor) because they were wildly overconfident and careless. Climb DCs were as printed, and most of the party (if not all of them) couldn't make the checks, primarily due to significant armor penalties. The wizard ended up using several casts of teleport or dimension door or something to get out, which required an extra day's rest to reprepare spells. All the trap did was delay them at the cost of some spells, which was really just time. That time cost meant that the next town they went to was an elaborate ambush of suicide-bombing gnolls equipped with necklaces of fireball (long story), because the succubus had time to set that up (which was the point of the giant hole).
Failing forward: The party fails a check to notice a trap, fails whatever save, but essentially no matter what, they'll still be able to make it to their destination (eventually), at the cost of time and/or some other resource (which is probably renewable, so really just time). More time means better-prepared foes later.
Simulationist with player agency: The trap design itself is a reaction to player choices. It targeted a party weakness, from an enemy known to them, using an ability within the expected portfolio of that enemy. The players did not prepare for that weakness. The players still tried to get out in creative ways (wizard tried to turn into a winged humanoid first, but didn't have good enough maneuverability to fly straight up and there wasn't room to do so diagonally, so it didn't work). Foe preparedness will vary based on time spent. They were somewhere around 10th level, so any number of methods could have gotten them out.


2 people marked this as a favorite.
MaxAstro wrote:
PossibleCabbage wrote:

I personally let the players set a great number of the details themselves when they opt to, even if doing so gives them an advantage, provided they are being at all reasonable. Since one of the main jobs of a GM is to describe the world in a way that creates a mental picture that is shared by all. If we change "imagining the world" from a passive task (i.e. "listen to the GM and try to follow along") to an active one (i.e. "try to think of what sort of thing could exist here which would be helpful") I find it works much smoother.

Like RPGs are fundamentally "asymmetrical improv" with a bunch of systems grafted on, but there's no reason they can't be less asymmetrical. So if my players want to posit the existence of a convenient dark alley, secret tunnel under some ruins, a nearby hamlet, a chandelier rope, a scullery maid, etc. then I'm happy to confirm they are correct.

One of my favorite things about Exalted 3e is that they codified this in the form of players being able to make a "Fact Introduction" roll. Basically they make a knowledge check relevant to the fact they want to introduce, and if they succeed that thing - subject to GM interpretation - becomes canon.

So instead of "I succeeded my knowledge check; what fashion is popular in Nexus right now?" it becomes "I'm going to roll Fact Introduction; my character has heard that long sleeves and floral patterns are all the rage in Nexus lately."

Oh hey, I really like this idea. It's a similar concept to some "I know a guy" mechanics I've played with in the past. Stuff like this can be a lot of fun, really puts the cooperative in cooperative storytelling.


The locked door scenario was used twice when "Fail Forward" was first introduced to me. The first scenario was trying to save a good dragon from an evil princess. If the Rogue beat the DC then they unlocked the door before the princess sacrificed the dragon. If they failed the DC, they unlocked the door but the dragon had been sacrificed. No matter what, they unlocked a locked door but the narrative cost was huge.

The second was trying to unlock a door while avoiding notice by a guard. If they succeeded on the DC, they unlocked the door silently. If they failed, they unlocked the door but the guards were shouting an alarm. Sure, the second scenario would have worked the same way if the lockpick roll failed and thus required the team to bash down the door.

These examples showed me that it was largely better to think of the reasons why the players were generating tests instead of the binary "this door has a DC of 15, do they unlock the door or fail to unlock the door?"

However, by the same merits, I'm not really sure the first example counts as failing forward. Their ultimate goal is rescuing the dragon, not unlocking a door. If they failed to save the dragon but still opened a door, I'm pretty sure the players will still feel like they failed and didn't go forward.


Pathfinder Lost Omens, Rulebook Subscriber
Vornesoul wrote:

The locked door scenario was used twice when "Fail Forward" was first introduced to me. The first scenario was trying to save a good dragon from an evil princess. If the Rogue beat the DC then they unlocked the door before the princess sacrificed the dragon. If they failed the DC, they unlocked the door but the dragon had been sacrificed. No matter what, they unlocked a locked door but the narrative cost was huge.

The second was trying to unlock a door while avoiding notice by a guard. If they succeeded on the DC, they unlocked the door silently. If they failed, they unlocked the door but the guards were shouting an alarm. Sure, the second scenario would have worked the same way if the lockpick roll failed and thus required the team to bash down the door.

These examples showed me that it was largely better to think of the reasons why the players were generating tests instead of the binary "this door has a DC of 15, do they unlock the door or fail to unlock the door?"

However, by the same merits, I'm not really sure the first example counts as failing forward. Their ultimate goal is rescuing the dragon, not unlocking a door. If they failed to save the dragon but still opened a door, I'm pretty sure the players will still feel like they failed and didn't go forward.

The narrative still advanced and thats the important thing. Of all the games I've played, I think the best have been the ones in which the players didn't actually succeed or barely succeeded in their goals. Maybe not right there and then, but they are always the ones my players talk about with the most fondness. (So long as it wasn't a TPK that is.)


5 people marked this as a favorite.
DM_aka_Dudemeister wrote:

I use fail forward mechanics all the time. I think a lot of people have some misconceptions about what Fail Forward means.

I mentioned this the other day on twitter, but the lack of academic study on tabletop RPG design is to the detriment of the community, as discussions get derailed and mired in semantics and definitions rather than the far more interesting discussion of affect, effect art and culture.

But here I go delving into a discussion that has devolved into semantics and definitions.

Fail Forward is not the following:

- Players succeed, or succeed at a cost no matter their roll. I saw this one bandied about a bit. That's not what failing forward is. Of course the dice and mechanics should be meaningful, otherwise what's the point of playing a game that has mechanics and dice.

- The GM moving the goal posts in the background. Again setting a secret DC and then adjusting it based on the player's rolls, it's functionally identical to the above.

What I feel Fail Forward is, and what it's for:

When players make a dice roll, no matter what they roll: SOMETHING INTERESTING MUST HAPPEN. If there is no interesting outcome on a successful, or a failed roll, then there was no reason to make a roll. (In PF1e this was what Take 10 and Take 20 were actually for).

Fail Forward means that on a failed roll: SOMETHING HAPPENS. The situation escalates, or there is a real cost that makes the difference between success and failure.

The issue is people assume "Forward" means to the end of the adventure, rather than the next bit of the narrative. Player characters are adventurers, by definition everything they do is an adventure.

They fail to scale the cliff:
"Cool, some of the rock crumbles away in your hand you start sliding down the cliff face, more rock is following you down! Oh god, that's a rock-slide, what do you DO?"
Maybe they try and push off the cliff and activate their ring of Feather Fall so the rockslide will fall safely under them (Acrobatics), find an...

This does have one issue though. As soon as the players fail only ONCE, that challenge is over. This is one of the things John Lynch said that stuck to me. Maybe the players tried to climb the cliff and failed. In your example, they now find the entrance, which means they do not get to try the "Fly over it" idea since it's pointless now. This means the "obvious" solution is always the right solution. They aren't really incentivized to experiment.

There are games where the players would be hit by the rockslide and then have to consider whether they now fly over or think of something else. Maybe now they can find the entrance IF THEY SEARCH. Seeing the consequences of attempting a check out of their league without much of a reward will educate players that they gotta actually think about the choices they make, because there are wrong ones.

It depends on the party, but after being a railroady/fail-forward GM for so long I've become disappointed in how I've raised my players to just sit around wait for me to tell them how the story advances without them taking initiative or trying creative things. I've since started doing my best to give them more independence, including the freedom to fail.


My question is- what is gained from letting the players roll over and over again to get the correct number, other than insulation from narrative consequences?


PossibleCabbage wrote:
My question is- what is gained from letting the players roll over and over again to get the correct number, other than insulation from narrative consequences?

That maybe they'll stop rolling and figure out something more intelligent.

EDIT to elaborate: Who is to say the key isn't out there on their side of the door? Maybe they can smash it (Usually with identical results to the fail-forward solution to this) or wait for the denizens of the building to try and use it. What if there's a secondary entrance from the other side? Maybe there's nothing essential behind it anyways! A a magic item for the Rogue the party clearly doesn't have.

I guess it depends on how available these alternatives based on the dungeon/adventure design. But the "bottleneck DC40 lock" is a pretty red flag anyways; it doesn't stop any adventure.

Keep in mind, I have advocated for Fail-Forward mechanics in this same thread, but the ones I consider good ones are pretty different from the examples others are coming up with.


1 person marked this as a favorite.
PossibleCabbage wrote:
My question is- what is gained from letting the players roll over and over again to get the correct number, other than insulation from narrative consequences?

I just can't help myself. I had to respond to this.

If there is no chance of success and no cost for failure then there is no rolling again and again. "You spend the next ten minutes failing to climb the cliff. It becomes clear that it isn't going to work and you must find another way to reach your goal. What do you do?"

Now as much as I'd like to respond to all the other posts it became clear I wasn't wanted so unless someone specifically asks me to respond to their specific post I won't. But for all the strawman arguments I got accused of making, I had to point out that PossibleCabbage was clearly arguing against something (that I think most here would agree) only a bad DM would actually do.


I think we can agree that if climbing the wall or unlocking the door is the only way to proceed, and if there are no repercussions for failing then there probably should not be a roll. Or at worst some kind of check for "how long did it take?"

But the issue I think is if the GM has planned for routes A, B, and C to proceed, along with recognizing that players are liable to come up with even more, when the players select the hardest, most obvious route and fail whether there should be consequences for not picking a better solution or whether nothing should happen and they are free to pick something else.

Like if the PCs are tasked to "get into a fancy party to which they are not invited" if the PCs decide "let's just bring a ladder, smash a window and get in that way" I figure the "fail forward" here is after a guard chases them away, security is more alert and posts somebody at the accessible windows. In doing that you make it clear "that is even less likely to work the next time, try something else." Plus you fail them forward because moving more security to the 2nd floor windows means there is less of a guard presence somewhere else.


2 people marked this as a favorite.
PossibleCabbage wrote:
My question is- what is gained from letting the players roll over and over again to get the correct number, other than insulation from narrative consequences?

Verissimilitude.

Also, avoidance of Sanity loss from trying to make up "interesting consequences" for every roll in the systems which basic gameplay loop is not really suitable for it, because rolls are too numerous and situations where a couple of failed rolls will swerve the "narrative" straight to a TPK are common.

Of course, a decent game should not force PCs to roll in situations for which Take 20 exists.


Verisimilitude is pretty subjective. One could argue that someone who will allow for unimpeded access to their locks and scalable walls for extended periods of time, simply relying on "they are hard to pick/climb" is someone who really hasn't thought through their security very well, and someone else should have robbed the place already.

Like an accessible lock that people don't check on with some frequency is fundamentally just security theater. I mean, some time antagonists just do ill-considered things, but if they aren't going to go to the level of "post patrols" then it's hardly a stretch to say they might not have managed to lock all the doors and might not have remembered to put their ladders away.


2 people marked this as a favorite.
PossibleCabbage wrote:

I think we can agree that if climbing the wall or unlocking the door is the only way to proceed, and if there are no repercussions for failing then there probably should not be a roll. Or at worst some kind of check for "how long did it take?"

But the issue I think is if the GM has planned for routes A, B, and C to proceed, along with recognizing that players are liable to come up with even more, when the players select the hardest, most obvious route and fail whether there should be consequences for not picking a better solution or whether nothing should happen and they are free to pick something else.

Like if the PCs are tasked to "get into a fancy party to which they are not invited" if the PCs decide "let's just bring a ladder, smash a window and get in that way" I figure the "fail forward" here is after a guard chases them away, security is more alert and posts somebody at the accessible windows. In doing that you make it clear "that is even less likely to work the next time, try something else." Plus you fail them forward because moving more security to the 2nd floor windows means there is less of a guard presence somewhere else.

Which roll did they fail here? The (common)fail-forward scenario is they just walk up the front door, fail the diplomacy/bluff to be allowed in but still make it in somehow, perhaps through a bribe or with someone keeping an eye on them. They wouldn't need to use the ladder gimmick, that would require the main attempt to be utter failure than they decide to try something else, or at least the players figuring it's not even worth trying the front door approach.

If I had to fail-forward this situation, I would make it so when they fail front door attempt they may notice an NPC open a nearby window... Or maybe overhear a conversation about a musical troupe being late to the event. Even those are a bit heavy handed if too obvious, though.


1 person marked this as a favorite.
Pathfinder Battles Case Subscriber; Pathfinder Pathfinder Accessories, Starfinder Roleplaying Game, Starfinder Society Subscriber; Pathfinder Roleplaying Game Superscriber

I find it curious that some folks seem to assume that Fail Forward is always going to be an option just because it is in the GM's toolbox. I will only be using it when the PCs, for whatever reason, are completely stopped by some obstacle that they simply cannot bypass, and thus the story would otherwise come to a halt. To use the cliff example: if the party is capable of flight then I would never use Fail Forward when they fail climbing checks since they can always fly around the obstacle. It may mean they have to camp out to prepare the necessary spells, but that's the price of being unprepared. For me, Fail Forward is simply one of many tools that I have as a GM to help ensure the game doesn't come to a screeching halt.


Pathfinder Pathfinder Accessories Subscriber; Pathfinder Roleplaying Game Superscriber

Emphasis added:

PossibleCabbage wrote:

Verisimilitude is pretty subjective. One could argue that someone who will allow for unimpeded access to their locks and scalable walls for extended periods of time, simply relying on "they are hard to pick/climb" is someone who really hasn't thought through their security very well, and someone else should have robbed the place already.

Like an accessible lock that people don't check on with some frequency is fundamentally just security theater. I mean, some time antagonists just do ill-considered things, but if they aren't going to go to the level of "post patrols" then it's hardly a stretch to say they might not have managed to lock all the doors and might not have remembered to put their ladders away.

But...like...that's how things work in the real world. Security is hard. It's asymmetric (attackers only need one exploitable route through, defenders have to cover many potential approaches), it's inherently inconvenient (if someone actually lives or works there, they're not going to to want to bypass 50 traps every time they come and go), and it's expensive (a PF1 permanent alarm has a 2500gp material component cost + any spellcasting required).

Security failures in game rarely break verisimilitude for me. It's usually the perfect or surprisingly comprehensive defenses that I have issues with, because they tend to require things that aren't level-appropriate or simply don't exist in the established rules.


5 people marked this as a favorite.

At Rico's request I'm replying to his post.

RicoTheBold wrote:
Edit: Quick TLDR; my basic position is that John Lynch 106's definition is fine, but he's selectively ignoring it when it describes something that aligns with how he already does things because that's not how he got to the same result.

Then let's make this abundantly clear: There have been examples in this thread of failing forward that (IMO) highlight why failing forward is a bad mechanic.

I'm going to pick one I've already posted, the Jason Buhlman example and an example from the DMG2 for D&D 4th edition that I've been told doesn't exist. I'll then post how I'll fix these examples.

Jason Buhlman Podcast wrote:

(not a direct transcript)

Jason: Finding the Hydra is going to be really hard. What do you do?
Player: Can I use survival?
Jason: THat is exactly how you would do it.
Player: Great. I get a 27.
Jason: That's enough. Although we have a new mechanic called the fail forward mechanic. Had you failed, I would have still let you find the hydra.
Nerdologists Blog wrote:
Failing forward is the idea that you still get to unlock the door on a failed roll, but it comes at a cost.
Dungeons & Dragons 4th edition Dungeon Master's Guide 2 wrote:
Eventually, a dead branch will stump you. You just can’t imagine a consequence of failure that takes the story anywhere fun or interesting. In this case, allow the players an automatic success. Either secretly set the DC of the check to 1, or drop the pretense and tell the players that they overcome the obstacle without resorting to a die roll. If this easy success feels like a lapse in story logic, describe a fortuitous event that makes the obstacle suddenly surmountable.

I think we can all agree that these use the "fail forward" mechanic in the same way, right?

They're all terrible examples of GMing/adventure design.

Jason Buhlman's Use of Fail Forward
Jason first framed the situation of finding the hydra as being really hard. He then revealed that he lied and that finding the hydra was actually going to be so easy that even a barbarian with -1 survival could have done it. He said "I would have just introduced some cost to make it seem like the die roll wasn't a complete waste of time."

Now I'm not saying Jason is a bad GM. Perhaps he was sick that day. Perhaps they were running low on time. Perhaps this is really how the fail forward mechanic is framed in Pathfinder 2e and he wanted to highlight this mechanic AS WRITTEN so that listeners would get an insight into how s*@~ some of the new rules are going to be about how this mechanic works in PF2e.

Now we all have our moments where we GM bad. It doesn't we're bad GMs. It just means we're human.

Here is how I would have framed Jason's hydra situation

Alternative Hydra Example wrote:
Pretend Jason: The hydra's path through this area is clear. However because of the terrain you are going to lose the path with quite a bit of regularity, making it difficult to determine which direction the hydra took. With enough time though, you could succeed. With a good survival check it would take less time then it otherwise would.

This gives all of the players the information they need to make a meaningful choice in hunting down the hydra.

It removes the b**%~*** of "finding hydras is so hard that even a -1 survival barbarian can do it" and it accurately frames the problem. The players can now try to employ magic items to make the tracking easier or perhaps even remove the need to use survival at all.

Correctly framing situations for players empowers them to be able to make the best decision they can. Now they might expend half their parties resources making it easier to find that hydra. That might then leave them underpowered when facing the hydra which could be a terrible decision. But it is a decision that the players chose to make.

Failing forward as used in this example removes the agency from the players. It presents a false choice to them and ultimately could have made it feel like the player's had no real impact on the outcome of this situation.

Nerdoglogists Example
We can assume that this stock standard scenario, using the fail forward mechanic, would have been presented like this.

Imaginary Lock Problem wrote:

GM: You find the door is locked. Unlocking it could be difficult.

Player: I whip out my master thieves work tools and give it a try. I roll a 27.
GM: Great. You unlock the door. Had you failed I would have still let you unlock it, there just would have been a maid on the other side. As it is, the coast is clear.

Now not all GMs are going to be that honest with their players. But again a false choice is presented and the players don't make the best they could because the GM is essentially lying to them.

A better solution would be

Alternative take on the Imaginary Lock Problem wrote:
GM: You find the door is locked. A cursory look reveals that the lock shouldn't be too difficult to unlock, which is what you've done for the past 10 levels when I present you with a locked door. However you know that if you aren't careful that unlocking this door could be quite noisy. What do you do?

Once more the players know what the stakes are (alerting anyone whose potentially on the other side) and can plan accordingly. Perhaps they do some more scouting. Perhaps they go to a door with a window so they can check no-one is on the other side. Perhaps the sorcerer expends their last high level spell slot to cast silence.

Regardless of what happens, the players can only blame their choices and the dice if they fail. There is no "Gotcha" GMing happening in the alternate version. There was in the GM proudly proclaiming that they were going to use a fail forward mechanic to make sure the PC's moved the story forward.

D&D 4th ed DMG2
I think we can all agree that lying to players (by secretly setting the DC to 0) is just bad GMing. Even before you get to a DC 0 check, it still says EVERY failure must ALWAYS have an interesting result. That actually highlights the exact proposal that Dudemeister said he would do. No matter what, something interesting happens on failing.

Not every skill check needs something interesting to happen on EVERY potential outcome. Doing that simply teaches the players that there is no consequence for failing and so they should try and use the brute force method every time.

------------
Basically in the above examples "fail forward" is being used to excuse (what I consider) bad GMing. In my replies in this thread, to try to avoid being antagonistic I've looked at ways to reframe people's examples to avoid (what I consider) bad GMing in order to avoid being more antagonistic then necessary. So if that's made you feel like I'm pulling out "No True Scotsman" I hope this post has made it quite clear I'm not trying to do that.

RicoTheBold wrote:
It's all a matter of framing

Yes. "Fail Forward" is being used to justify framing problems to players in the wrong way making them think they know what problem they're trying to overcome, when the GM is secretly making them overcome a different problem.

Listen to the podcast. Jason clearly says "tracking the hydra down is going to be hard" he then reveals "but thanks to this new thing we have called fail forward, despite how hard I said tracking down the hydra was going to be, you couldn't have failed".

Do you think incorrectly framing the problem to players is a good thing? If yes, then you have no problem with the Jason example. If no, then you agree that the Jason example is a case of bad GMing and he is using the label "fail forward" to justify that bad GMing.

RicoTheBold wrote:
For the tracking example, it is a literal example of a fail-forward mechanic. Maybe the DC could be better described as avoiding danger while following the trail, but that doesn't change what it is

Paizo adventures have had situations where no roll is necessary to get to the final outcome, but a roll might be necessary to avoid a problem along the way. Therefore if we take Jason at his word when he says "we have this new thing called fail forward" then we can agree that fail forward isn't simply "tracking someone down with a DC so low no-one can fail".

RicoTheBold wrote:
John Lynch wrote:
[Trapped door example]
How is that not a fail-forward mechanic? "Fail forward is when you allow the PCs to succeed with the same approach even though they failed to overcome the challenge with that approach. Often with a loss of resources or a minor obstacle that will take up table time but not meaningfully hinder their ability to reach the end goal."

Touche.

Your right. I do use fail forward mechanics as defined by me. Pathfinder's APs have also had scenarios like this for years.

Given we're told that PF2e's fail forward mechanic is something new, we'll have to see how they frame the mechanic in the Core Rules and whether or not they stick with the "bad GMing" examples I've provided, if they use it in the same way they've been doing for 10 years now or if they do something completely different to anything we've discussed thus far.


1 person marked this as a favorite.

I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.

Even so, I agree with John's analysis and the mechanic is still problematic for me as a designer and a GM even if the the PCs are in the dark about what happens. Why? Because it is basically demonstrating that PF2 doesn't have mechanically interesting ways to handle exploration mode without breaking its own rules and trivializing most of the skill challenges between encounters.

It is going to be pushing GMs into inventing "failing, but at a cost" scenarios for things like tracking monsters through the forests, climbing cliffs, and sneaking into castles.

Picking locks quietly should be a disable lock check and a stealth check. PF2 and probably PF1 and D&D turn the tension of multiple skill checks into something that feels tedious and boring to many players because extra checks make for an increased chance of partial failure and deciding what that can look like usually requires a lot of planning and skill as an adventure writer.

These kind of exploration mode challenges have not been well developed in most D20 games I have played. They look good in cinema, but game mechanics don't represent slow exhaustion and the tension of needing to not make little mistakes over a long period of time. This is why most games do best folding these into a paragraph of narrated expository text to set a mood and a tone for an adventure rather than a mechanical feature of it.

I do like John's suggestions for folding the these narrative set ups with a chance for the players to make one of a couple of choices for how they want to approach it, so that at least the choice is the players instead of the GM, but sometimes it can be good for failure to lead to a dead end instead of a new way forward or else the party never learns to step back and reconsider their position.

I think learning how to say "Ok, you've tried this path, it doesn't seem like your party is ready or trained or built to succeed this way, what other paths do you think you can take?" Rather than pushing them through a bad plan that compiles with bad rolls. And if there aren't other paths that the party can find, then the fail forward mechanic might be necessary, but it is from a position of failure as a GM, and it should be used as sparsely as possible.


2 people marked this as a favorite.
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.

Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.


1 person marked this as a favorite.
John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.

I don't disagree. I think all of this discussion about "fail forward" is much more about GMs and adventure designers failing than the party or the players.


2 people marked this as a favorite.
Pathfinder Lost Omens, Rulebook Subscriber
John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.

Dang, gotta get rid of all the illusions, invisible creatures and any npc using bluffing.


2 people marked this as a favorite.
Malk_Content wrote:
John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.
Dang, gotta get rid of all the illusions, invisible creatures and any npc using bluffing.

I think John, and definitely myself, are saying that any mechanic based on hiding or lying about how it works to the PC is a bad mechanic. Illusions, including invisibility work because players trust the GM is following the rules for them. Invisibility where the GM moves an invisible creature is or where the illusion is because they want to would be more analogous.


I thought of another example of a scenario where there a PF AP or module gives a situation where a die roll is called for but it is a must pass situation that basically calls for a lot of die rolls.

I will modify the example enough to make it generic:

The adventure begins with the party awakening inside a series of individual crates in a dark room, wondering how they got there and how they will get out of them. The check to get out of the boxes is difficult, but there is no real cost for failure. Adding an element of "making too much noise" would very easily turn an introductory situation into a TPK of monster/monsters swarming the party while still trapped in the boxes.

Why make the party roll? Or just have everyone roll once and for a STR check and let the player who rolls highest be the first out for the sake of bragging rights and then move the story along? Why bother trying to create a illusion that the situation is time sensitive or trying to make the "failure interesting" when the mechanics for STR checks are just plain boring.


Pathfinder Lost Omens, Rulebook Subscriber
Unicore wrote:
Malk_Content wrote:
John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.
Dang, gotta get rid of all the illusions, invisible creatures and any npc using bluffing.
I think John, and definitely myself, are saying that any mechanic based on hiding or lying about how it works to the PC is a bad mechanic. Illusions, including invisibility work because players trust the GM is following the rules for them. Invisibility where the GM moves an invisible creature is or where the illusion is because they want to would be more analogous.

Well I'd certainly a say that a GM who doesn't inform players of their style of gming during session 0 is a bad GM, but if players are aware and signed up for it then I disagree that it is inherently bad gming.

For the box, yeah either don't roll, make one roll that serves to highlight the strength of one character (the narrative benefits of this are great when the party doesn't yet know each other) or make one roll that modifies a further encounter e.g

Fail:you break out and have to fight one monster that came to investigate before you get your gear.

Success:you break out and get your gear before the investigator arrives.

Crit success:you break out and can change an allies fail to a success as you have time to help them before the investigator arrives.


1 person marked this as a favorite.
Malk_Content wrote:
John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.
Dang, gotta get rid of all the illusions, invisible creatures and any npc using bluffing.

Did you want a response or you just looking to score points?


1 person marked this as a favorite.

I honestly want to know, is there any point in continuing this discussion, or are you just intent on denying every single thing that everyone else puts forward? Save me a little time here so that I can remove my follow from this thread.

At this point, it really feels like the latter, and I don't feel like there's any value in me participating in this, especially given how you've misrepresented the situation from the Paizocon stream so badly that it's basically completely unrecognizable.

So tell me: how much am I wasting my time in this?

(Now see, obviously I failed my fail forward check. So even though I still realize I'm wasting my time on this, it's taken me way too long, losing me valuable time that I could have spent to do other things. At least I don't have to get to a wedding across 200 miles of forest in a day.)


1 person marked this as a favorite.

I responded to Rico's post because he asked me to. I also acknowledged he had a valid point and that I do use the fail forward mechanic.

If you want to explain how I have misrepresented Jason's hydra example, go for it. I admit I did not include a direct quote. So if you feel my quote misrepresented the exchange please let me know how.

If you dont want to engage me any further in this discussion you dont have to either. Tell me what you want from me and I will see if I can oblige.


2 people marked this as a favorite.
Pathfinder Pathfinder Accessories Subscriber; Pathfinder Roleplaying Game Superscriber

Emphasis added:

John Lynch 106 wrote:
Unicore wrote:
I personally assume that Jason is telling the audience this information more so than the players, and the intention would be that the GM would never tell the party that they couldn't fail.
Thats how D&D 4th ed DMG 2 uses it which IMO is bad. any mechanic that only functions well by hiding information from PCs is a flawed mechanic.

Let's not go overboard and say they're all bad (and invite a bunch of largely irrelevant counter-examples), but I agree that the GM should give the information that the characters would have to make appropriate decisions. If I were executing Jason's hydra-tracking example, I probably wouldn't mention the DC (hiding the information) and just adjudicate appropriately, but I'd probably also narrate the overall obviousness of the tracks and let them know what costs their choice gave/avoided (time loss) since their characters would reasonably learn that. I'd probably call out that the trail seems obvious at first glance and likely to remain so if the players were mulling over whether or not they had the capability to even attempt it, or if they were asking questions about/investigating the footprints or trail in the first place. Sometimes I tell players stuff like DCs after the fact, so they can better learn how to play the game and make choices from the narrative information their characters receive, and if the players read the rules they can expect the DCs to line up accordingly, but at the worst they'll still learn what type of success/failure they got, which narrows the range of potential DCs considerably based on their rolls.

All of that is hugely different from deliberately misleading the players as to the reasonably likely potential outcomes of their choices. I don't really agree that Jason was misleading here, though. Since the PF1 mechanic for failure in the Follow Tracks use of Survival involved losing the tracks and having to roll again to find them after an hour of searching (for outdoor tracks), easy tracks generally would just represent a loss in time. Since that's been the case for a long time, my interpretation of Jason's description basically boiled down to "I'm cutting down on the pointless rolls to get to the consequences faster." It didn't strike me as concealing or misrepresenting pertinent information from players, because lost time has been the expected cost of bad survival rolls on tracking for 20ish years of D20 gaming, and with experienced players they probably assumed as much anyway. As printed in the playtest rules, a failure is losing the trail and can be tried again after a one hour delay. Tracking an easy target is and has long been fail-forward by the inherent design of the rules (whereas a difficult target may well evade tracking).

I think it's better to say that fail-forward works best when the players have a good idea what their characters would know and how their choices interact with the rules (for the way we both seem to play without meta-narrative mechanics). The players get to take ownership of their decisions and the consequences, without that pesky deus ex machina feeling when their failures or successes, and therefore their choices, are irrelevant.

Doomsday Dawn Adventure 7 spoiler:
And in the Doomsday Dawn adventure, my players spent about five minutes thinking of a question to ask the Ashen Man, and got told (as per the book's instruction) that they were destined to ask that question and that the answer was irrelevant.
I didn't like running that, and if I were one of the players, I wouldn't like hearing that.

However, for situations where the characters wouldn't know the likely outcomes, I'm all for letting them make choices with limited information. Sometimes you're going through a dungeon and you just have to pick the left door or the right door.


1 person marked this as a favorite.

I wont discuss the validity of hiding DCs or the relevance it has to what I said at your request.

I will say:
1) You agree that you would have presented Jason's hydra situation differently. Personally when I hear "Something is going to be difficult" I dont leap to the conclusion that someone with a -1 skill modifier could succeed. Perhaps everyone here does. That's certainly different play experiences. And yet that was the exact scenario Jason presented to his players. I would be interested in knowing what the DC was and whether it was a DC 19 (this meaning it was impossible for anyone to not find the lair eventually).

2) Your saying retrying track attempts existed in PF1e and that anyone would always be able to succeed eventually. Given Jason said that his "fail forward" mechanic was new, it would suggest its not as simple as how PF1e worked. We will have to wait until the final rules to know for sure.


John Lynch 106 wrote:

Jason Buhlman's Use of Fail Forward

Jason first framed the situation of finding the hydra as being really hard. He then revealed that he lied and that finding the hydra was actually going to be so easy that even a barbarian with -1 survival could have done it. He said "I would have just introduced some cost to make it seem like the die roll wasn't a complete waste of time."

Let's start with the most obvious here. He said "a 28 is good enough so that you don't waste time. [...] If you fail this roll, you'll still find your hydra, it'll just take way too long."

Your description of Jason, summarized: "I made you roll even though it had 0 effect on anything, and would have had some cost to pretend it was relevant."

I don't think I need to state why these two statements are in no way equivalent, especially given they had less than a day to accomplish everything. Time is 100% a relevant factor in this, and wasting hours when you have less than 24 is not a thing you want to do.

Furthermore, Jason made these characters. He 100% knows that they can make it eventually. This is no way equivalent to "barbarian with -1 Survival", and pretending as though it is is incredibly disingenuous. That's literally a line you made up, and have been pulling out all over the place to justify yourself. Especially given that the party wouldn't even be attempting it in that scenario.

Lastly, this is literally no different from the standard Survival check. It just wastes valuable table time to go "roll a check. That's a failure. Anyone want to do anything in that hour that you're waiting? No? Okay, roll again." ad nauseam until someone either punches the GM in the face or succeeds. No one is interested in that.

Overall, there are just so many points that are different it's like comparing oranges to pineapple houses.


1 person marked this as a favorite.
Cyouni wrote:
Let's start with the most obvious here. He said "a 28 is good enough so that you don't waste time. [...] If you fail this roll, you'll still find your hydra, it'll just take way too long."

He said both those things AFTER the roll was made. Did he not?

Cyouni wrote:
Furthermore, Jason made these characters. He 100% knows that they can make it eventually. This is no way equivalent to "barbarian with -1 Survival", and pretending as though it is is incredibly disingenuous.

If you think Jason meant "we have this new mechanic called fail forward, where so long as your bonus is high enough to meet the DC, you can't possibly fail, success just comes at a cost if you don't roll high enough" then that's great. That's a mechanic I think that sounds like a good one.

BUT it's not new. That's exactly how PF1e worked for any mechanic that has "retry: Yes". Perhaps fail forward isn't new and Jason misspoke when he said it was new. We'll have to wait and find out.


John Lynch 106 wrote:
Jason Buhlman Podcast wrote:

(not a direct transcript)

Jason: Finding the Hydra is going to be really hard. What do you do?
Player: Can I use survival?
Jason: That is exactly how you would do it.
Player: Great. I get a 27.
Jason: That's enough. Although we have a new mechanic called the fail forward mechanic. Had you failed, I would have still let you find the hydra.
Nerdologists Blog wrote:
Failing forward is the idea that you still get to unlock the door on a failed roll, but it comes at a cost.
Dungeons & Dragons 4th edition Dungeon Master's Guide 2 wrote:
Eventually, a dead branch will stump you. You just can’t imagine a consequence of failure that takes the story anywhere fun or interesting. In this case, allow the players an automatic success. Either secretly set the DC of the check to 1, or drop the pretense and tell the players that they overcome the obstacle without resorting to a die roll. If this easy success feels like a lapse in story logic, describe a fortuitous event that makes the obstacle suddenly surmountable.

I think we can all agree that these use the "fail forward" mechanic in the same way, right?

I find them to be subtly different.

The Jason 'quote' is bad because he doesn't specify any consequences for the failure. A game where if you fail your survival rolls, the hydra eventually finds and ambushes you would be 'fail forwards' with meaningful consequences. Perhaps that's what he meant, but he didn't say it.

Then again, it's only a (not necessarily accurate) transcript of someone talking off the top of their head. Those are rarely good for precisely defining gaming terms.

"you still get to unlock the door on a failed roll, but it comes at a cost" is also a weak definition. For example, if someone with zero lock-picking skills attempts to unlock a well-made door, "you pick the lock the door but Something Bad Happens" is a fail forwards that breaks verisimilitude and fails to specify any plausible bad consequences.
A better fail forwards would be "the guards inside hear you and open the door to find out what's going on and now you've lost the element of surprise".

The "if you just can’t imagine a consequence of failure that takes the story anywhere fun or interesting, just let them succeed somehow" is saying to only do this as a last resort where the alternative is the campaign grinding to a halt. I don't know if anyone defines "fail forwards" in this way.


1 person marked this as a favorite.
Matthew Downie wrote:
"you still get to unlock the door on a failed roll, but it comes at a cost" is also a weak definition. For example, if someone with zero lock-picking skills attempts to unlock a well-made door, "you pick the lock the door but Something Bad Happens" is a fail forwards that breaks verisimilitude and fails to specify any plausible bad consequences.

That's what google produced. Well. The first search result was a post lambasting fail forward. I didn't think it would be particularly fair to present that option. I instead presented the first favorable search result. I've been assured that there is a much better definition. All you have to do is google it and it's the very first result. Alas my search result didn't find it and the person who was quite confident it existed has chosen not to post it themselves.

Matthew Downie wrote:
Then again, it's only a (not necessarily accurate) transcript of someone talking off the top of their head. Those are rarely good for precisely defining gaming terms.

I wasn't the person who hunted down Jason talking about fail forward and put forward his presentation of the mechanic as a "good example of fail forward." I simply used it because I've been told that his example was a good definition for it and he is working on Pathfinder 2e so it seems likely he would know how it will work in PF2e.

Unfortunately it seems we'll have to wait for the rules to really understand how Jason was using fail forward.

As for D&D 4e's DMG2. It is what it is. I don't play narrative focused games so I couldn't use any of their definitions of fail forward. I used the definition from the game I do have access to. After being told no game would ever use that style of fail forward.


John Lynch 106 wrote:
Fail forward removes all player agency. It says “no matter how much you mess up we WILL reach that predetermined outcome because nothing you do matters.”

I don't think the problem with "fail forwards" is anything to do with Player Agency.

The first definition of player agency I found on Google:

Quote:

1 The player has control over their own character's decisions.

2 Those decisions have consequences within the game world.
3 The player has enough information to anticipate what those consequences might be before making them.

By this definition, it sounds like a game where players had a 100% chance of success at everything would have more agency than one run the traditional D&D way.

If every time you try to pick a lock the door opens somehow, that's agency: you choose what to do, and consequences.

Loss of player agency is when you decide not to try to open the door, so the GM conspires open it for you.

What excessive or clumsy fail-forwards harms is (a) Dice Agency (the idea that the number on the dice matters, without which dice rolls have no tension) or (b) plausibility. Because in real life sometimes you try something and nothing happens.

Is there a term for when you try something, it fails, and so you have to think of something else? "Fail sideways," maybe?

For example, I try to talk the guard into letting me in. I fail a Diplomacy check. I try bribing him. I fail another Diplomacy check. He threatens to arrest me. I succeed at a Diplomacy check and he lets me go.

I have achieved nothing. The story has not moved forwards. But now I have to consider a plan B. Cast Charm Person. Wait until nightfall and sneak in. Create a distraction to lure the guard away. Look for a back door. Climb in through a window. Murder the guard. Kidnap his family. Rub your magic lamp and use up one of your three wishes.

A universal Fail Forwards mechanic could mean that you'd never have to stop and think of a plan B, because when you either succeed at plan A, or the situation changes (eg you are arrested and brought inside) so that the plan B is irrelevant.

I'll try and make an exhaustive list of types of failure results, just for completeness:
(1) Fail forwards with no meaningful consequences. (You fail to pick the lock, but the door just drops off its hinges.)
(2) Fail forwards with meaningful consequences. (You fail to pick the lock, but this triggers a trap and the entire door explodes, hurting you. If you survive you can proceed.)
(3) Fail sideways. (You fail to pick the lock, but you could now choose to bash the door down or look for another way to proceed.)
(4) Fail repeatedly. (You can keep trying to pick the lock until you get a natural 20.)
(5) Fail catastrophically. (You try to pick the lock; this triggers an alarm. A massive number of guards are approaching. You must abandon your mission and flee for your life as a fugitive.)

Anything I'm missing?


2 people marked this as a favorite.

1st of all, in the interest of continuing a good faith discussion, let me say that I am continuing to learn more about this topic and developing my own understanding of fail forward by participating in this conversation. Thank you to the people participating in it.

Matthew Downie, your idea to break the fail forward mechanic down into its different usages is a good idea and I like the general categories you have. I think generally, all of these do get called failing forward and that has lead to some confusion in this thread and with how people interpret the core principle that the game should be fun taking precedent over the game is always consistent to its rules.

How that gets in-acted though will always vary heavily from table to table and some times GMs think the game being fun, means that the party should always be moving forward through a predefined story, while others interpret it to mean that the story will always be moving forward whether it is towards where the GM intended it to go or not. A lot of the scenarios I see people present are getting debated around the assumption that the party needs to be moving relatively quickly through skill based challenges.

Sometimes, I think this is true. Gaming with younger players or players that have not built up a strong team dynamic can require a GM to find ways to move the story along faster when to prevent players from getting angry at each other and the GM.

But almost all of the "Fail forward" solutions I see being put forward in this thread are more about condensing skill challenges into single skill checks, or at least massively reducing the number of cross skill checks. I understand why that is a popular choice for GMs to make, designing combat scenarios is something well supported in this style of game, and the rules are fairly clear on how to do. Skill challenges get a lot of rules too, only we get told that we should ignore them 90 percent of the time and only try to understand them when they would be directly meaningful to the story.


John Lynch 106 wrote:
Fail forward removes all player agency. It says “no matter how much you mess up we WILL reach that predetermined outcome because nothing you do matters.” I have yet to see an example of failing forward with skill checks actually being good. If anyone thinks they have one I’d be willing to discuss it.
John Lynch 106 wrote:
If you enjoy games where the players will successfully reach the final battle no matter what, then that's great.

IMHO, fail forward is about trying preserving play momentum/pacing rather than directing play direction/outcome/trajectory. It's the idea that even failure will produce interesting, dynamic outcomes for the players that keep the game flowing.

Let's take climbing a wall. The player is attempting to climb a castle wall. Escaping? Infiltrating? Either way. They fail the check. Instead of simplying saying, "You don't make it up the wall" or "You fall," the GM may instead say, "Your struggles climbing the wall is causing a scene, and you draw the attention of the guards below who are now drawing their bows at you."

The player has still not succeeded at climbing the wall. They may be half-way up and struggling to get higher. But their failure introduces a new complication to the narrative that has to be resolved: hostile guards below. The GM complication does not give the player a successful outcome nor does it railroad them towards a set outcome (i.e., you climbed the wall!), but it does keep the story moving. Agency is still in the player's hands. If the player's goal lies on the other side of the wall, then it is up to them to find an alternative path, but their failure climbing it has complicated things for them.


1 person marked this as a favorite.
Aldarc wrote:
The player has still not succeeded at climbing the wall. They may be half-way up and struggling to get higher. But their failure introduces a new complication to the narrative that has to be resolved: hostile guards below. The GM complication does not give the player a successful outcome nor does it railroad them towards a set outcome (i.e., you climbed the wall!), but it does keep the story moving. Agency is still in the player's hands. If the player's goal lies on the other side of the wall, then it is up to them to find an alternative path, but their failure climbing it has complicated things for them.

I am not going to speak for John, but my continued issue with this model of fail forward is the assumption that the initial challenge was only interesting to the story if it was accomplished successfully. Weren’t the guards always out patrolling below? Doesn’t that mean the party needed to climb quickly and stealthily from the beginning? In that instance, couldn’t the climb check mean that the party has to make another stealth check to go unnoticed before they can make another climb check?

I don’t understand how tying the guards to the climb check is an example of failing forward and just tying multiple skill checks into one roll.

Over and over again it seems like “fail forward” is being described as the GM deciding what needs to happen next on a failed check, rather than asking the players what they want to do now that the first check failed.


1 person marked this as a favorite.
Pathfinder Roleplaying Game Superscriber

I think ultimately this comes down to a preference of narrativist vs simulationist GMing, and people who prefer one or another are never going to agree.

For example, the statement "a mechanic that relies on the GM hiding information from the players is a bad mechanic" is so alien to me and my players that my first reaction to reading it was "the heck?!" and I had to stop and think to figure out where that statement could even be coming from.

But if you are coming from somewhere where that statement makes sense, then I can understand why you hate fail forward. Because F>F operates under the premise that the game world changes in response to the player's roll, and that idea is alien to people who prefer simulationist games.

EDIT: Also, when Jason described F>F as a "new" mechanic, I think what he really means is "codified". Paizo has been including instances of F>F in their adventures for ages, they just haven't referred to it that way or talked about it as an official thing.

For example, in Hell's Rebels -

Spoiler:
it's impossible to "fail" the treaty negotiation with Cheliax. Even if you natural-1 15 out of 15 rolls, you still end up with a treaty with Cheliax, just one that isn't super fair to Kintargo.

For an even more direct example, Serpent's Skull

Spoiler:
uses the EXACT tracking mechanic Jason does on the road to Saventh-Yhi. Failed Survival checks don't result in the PCs failing to find Saventh-Yhi, they just result in it taking longer to get there.


3 people marked this as a favorite.

I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way. In the aforementioned "climbing the wall makes a racket, which rouses some guards" we shouldn't jump immediately to combat, but we should force the players to choose between "keep climbing", or "get down and hide" or "get down and get ready for a fight" or "get down and try to bluff the guards that you saw the malefactors and they went that way, etc."

I mean "the guards in this area will be alerted if the PCs make a racket in a nearby area" is a thing that is codified in a lot of APs. But I fundamentally see no issue with "baddies pop in and out of existence" in places where the PCs can't see this happening.


Pathfinder Roleplaying Game Superscriber
PossibleCabbage wrote:

I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way. In the aforementioned "climbing the wall makes a racket, which rouses some guards" we shouldn't jump immediately to combat, but we should force the players to choose between "keep climbing", or "get down and hide" or "get down and get ready for a fight" or "get down and try to bluff the guards that you saw the malefactors and they went that way, etc."

I mean "the guards in this area will be alerted if the PCs make a racket in a nearby area" is a thing that is codified in a lot of APs. But I fundamentally see no issue with "baddies pop in and out of existence" in places where the PCs can't see this happening.

That's the thing, though. While I totally agree with you, some people DO have a problem with "reality reshapes itself arbitrarily based on the PC's rolls", to try to phrase it negatively.


1 person marked this as a favorite.
MaxAstro wrote:

I think ultimately this comes down to a preference of narrativist vs simulationist GMing, and people who prefer one or another are never going to agree.

For example, the statement "a mechanic that relies on the GM hiding information from the players is a bad mechanic" is so alien to me and my players that my first reaction to reading it was "the heck?!" and I had to stop and think to figure out where that statement could even be coming from.

But if you are coming from somewhere where that statement makes sense, then I can understand why you hate fail forward. Because F>F operates under the premise that the game world changes in response to the player's roll, and that idea is alien to people who prefer simulationist games.

EDIT: Also, when Jason described F>F as a "new" mechanic, I think what he really means is "codified". Paizo has been including instances of F>F in their adventures for ages, they just haven't referred to it that way or talked about it as an official thing.

For example, in Hell's Rebels - ** spoiler omitted **

For an even more direct example, Serpent's Skull ** spoiler omitted **

Especially for the Survival check example, I don't think anyone, John or I included, would argue that it is wrong to decide to have the players roll 1 check to determine how long it takes them to track something, as long as that is made clear in the way the GM asks for the roll. I am perfectly ok with ignoring rules (and making new ones/new interpretations for them) as a GM, especially if the rule feels arbitrary or inconsequential to the current situation, I just want my players to know that is what is happening so that they do not character build, or strategize around mechanics that are not in play.

For example if my player says, "I want to use survival to track this creature through that forest back to its lair."

I might respond, "Ok, it has a bit of a head start on you, you are not going to be able to track it down instantly, but you can make a survival check to see how long it will take you to find it." In which case I have the player roll to find out how long it takes. But if I know that I want there to be a possibility that another creature ambushes the party while they are trying to track the first creature, I am going to be sure to ask the party, before they make the survival roll, how they are going to be proceeding through the forest, and that trying to focus on stealth, or being on the look out for an ambush, is going to add time to the trip and let that consequence be determined by a different roll than the survival check, and have them make, or secretly make for them, any additional skill checks that might be relevant at the same time they make their survival check (again, I am fine with GMs keeping secret information about a situation, I just want the players to know what kinds of knowledge is being kept secret and trust that I am not arbitrarily deciding what the outcomes of those secret checks are because it would lead to the situation that I would find most interesting).

What I wouldn't do is just have them make a check and then decide whether they stumble into a different encounter, or bump into a maid cleaning up the trail of the creature, but knows where its lair is located, based solely upon their survival check, without giving the players the agency to decide how they are traveling through the woods, and without making it clear that those decisions will affect the outcome of the exploration event. Players cant decide that they are traveling quietly, and searching every step ahead of them for traps, and having a look out keep eyes behind them and up through the trees, and tracking a foe as quickly as possible, and not make decisions about who is doing what and what is not getting done. In my opinion, that creates much more interesting dramatic tension then boiling all of it down to just one check.

And if I don't want to create all of those possible random encounters or alternative situations and just want the story to proceed, then I don't ask for checks in the first place, and either ask the players to help me describe the travel and environment in an interesting way (if it doesn't matter if I control these things and I don't have something already written up), or I give them a brief couple of sentences of exposition about the travel (because there are some details about the environment or the trip which could be important later). If I turn it over to my players and they create enough interesting details that I want something to happen with that environment or because of that trip, I bring it back later in the story so that it can feel fully integrated into the story and not push the players into thinking they can game the system by creating details that will benefit them in the moment.


Pathfinder Roleplaying Game Superscriber

Correct me if I am wrong but it feels like most people are saying more or less the same thing just wording it differently.

It seems to basically boil down to:

Players attempts to overcome challenge and fails, but magically succeeds anyway, potentially with some negative consequence, because failure at this point would stop the adventure is bad.

Players attempts to overcome challenge and fails, so player finds a different route through the challenge through their own choice of actions potentially with negative consequences is good.

I would tend to agree that the first option is a bad option, if the players failed the check, then the players should be the ones to figure out another approach. I would also agree that designing an adventure where there is a bottle neck where the adventure has to stop due to a failed challenge and there are no other paths through, that would be bad adventure design.


1 person marked this as a favorite.
PossibleCabbage wrote:
I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way.

So the debate essential comes down to whether it is better for the GM to A:

arbitrarily make up a consequence that allows the party to proceed after they (the GM) ask for a roll that they only realize was unnecessary to make a check after the player failed it and there would otherwise be no significant consequences,

or B:

for the GM to admit that the roll was probably unnecessary and to either allow the party to take 20 to advance past it, or inform the party that that pathway forward is beyond their skill and that they should look for another way forward?


4 people marked this as a favorite.
Pathfinder Roleplaying Game Superscriber

I would argue that it's a debate about whether binary pass/fail is better than nuanced pass/pass with a cost, but I'm trying to avoid actually arguing because I've realized the debate is entirely philosophical, basically.


1 person marked this as a favorite.
Unicore wrote:
I am not going to speak for John, but my continued issue with this model of fail forward is the assumption that the initial challenge was only interesting to the story if it was accomplished successfully.

IMHO, that's not an accurate way of framing it. As others have mentioned, a lot of fail forward approaches usually only recommend calling for a roll if there are interesting possible consequences for success AND failure. So it is quite the opposite assumption than "the initial challenge was only interesting to the story if it was accomplished successfully." It may even be more interesting if failure occurs. Either way, it will push the fiction forward.

Quote:

Weren’t the guards always out patrolling below? Doesn’t that mean the party needed to climb quickly and stealthily from the beginning? In that instance, couldn’t the climb check mean that the party has to make another stealth check to go unnoticed before they can make another climb check?

I don’t understand how tying the guards to the climb check is an example of failing forward and just tying multiple skill checks into one roll.

A minor quibble: I have said nothing about a party, as I spoke only of a singular thief.

A lot of advice around "fail forward" typically embraces the idea of following the fiction. Were the guards always there? I would say that they were a potential possibility that existed in the fiction so it is not necessarily that the GM is arbitrarily willing them into existence. We could even imagine that the thief succeeds and then the GM narrates that the thief sees guard patrols below who don't manage to see him. The GM may have felt that it would heighten the sense of success for the thief: they successfully dodged a bullet. Has the GM arbitrarily willed them existence? Possibly, but the GM has always had tremendous latitude when it comes to these things (e.g., wandering monster table rolls anyone?).

Then why didn't the thief stealth? I don't think that that following the rabbit hole of how the fiction reached this point is pertinent. As to whether or not an additional stealth check was needed once the guards appear, then I would say that largely depends on the fiction and what the GM regards as the best call. But then calling for an additional skill checks because the guards entering the fiction is a narrative complication triggered by the initial climbing failure. Use your best judgment call.

That said, I know a number of old school GMs who prefer "atomic rolls," where every action requires an associated roll. There are a number of games and GMs that prefer tying in a series of ideas and stakes into a singular skill roll. So for example in Blades in the Dark you are likely rolling for more than a binary climbing check (prowl?), but for whether or not you can ascend the wall with your loot. This ties stakes and intent into the roll, and a number of narrative complications can arise from this.

Fail forward a general GMing principle and not a hard-fast causal code of "if X happens, do Y."

Quote:
Over and over again it seems like “fail forward” is being described as the GM deciding what needs to happen next on a failed check, rather than asking the players what they want to do now that the first check failed.

It's not so much "the GM deciding what needs to happen next on a failed check" and more about how failure pushes the fiction forward in new ways.


1 person marked this as a favorite.
PossibleCabbage wrote:

I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way. In the aforementioned "climbing the wall makes a racket, which rouses some guards" we shouldn't jump immediately to combat, but we should force the players to choose between "keep climbing", or "get down and hide" or "get down and get ready for a fight" or "get down and try to bluff the guards that you saw the malefactors and they went that way, etc."

I mean "the guards in this area will be alerted if the PCs make a racket in a nearby area" is a thing that is codified in a lot of APs. But I fundamentally see no issue with "baddies pop in and out of existence" in places where the PCs can't see this happening.

And if they have a silence spell, what fail forward result will you have occur on a failed climb check?


Pathfinder Lost Omens, Rulebook Subscriber
John Lynch 106 wrote:
PossibleCabbage wrote:

I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way. In the aforementioned "climbing the wall makes a racket, which rouses some guards" we shouldn't jump immediately to combat, but we should force the players to choose between "keep climbing", or "get down and hide" or "get down and get ready for a fight" or "get down and try to bluff the guards that you saw the malefactors and they went that way, etc."

I mean "the guards in this area will be alerted if the PCs make a racket in a nearby area" is a thing that is codified in a lot of APs. But I fundamentally see no issue with "baddies pop in and out of existence" in places where the PCs can't see this happening.

And if they have a silence spell, what fail forward result will you have occur on a failed climb check?

Possibly none, the PC took extra precautions to mitigate failure. It is a general principle to apply when it makes sense and is good for the game after all.


1 person marked this as a favorite.
John Lynch 106 wrote:
And if they have a silence spell, what fail forward result will you have occur on a failed climb check?

Somebody dropped something, it might not be valuable but anyone who comes by might spot it and figure out that there are intruders about, so "should someone go back down and retrieve it" is a question the party now faces.

Silver Crusade

3 people marked this as a favorite.
Pathfinder Adventure Path Subscriber
Unicore wrote:
PossibleCabbage wrote:
I mean, the point of "fail forward" is fundamentally that each roll changes the world state in some way.

So the debate essential comes down to whether it is better for the GM to A:

arbitrarily make up a consequence that allows the party to proceed after they (the GM) ask for a roll that they only realize was unnecessary to make a check after the player failed it and there would otherwise be no significant consequences,

or B:

for the GM to admit that the roll was probably unnecessary and to either allow the party to take 20 to advance past it, or inform the party that that pathway forward is beyond their skill and that they should look for another way forward?

If a roll has no consequence for failure then do not ask for a roll. That's basic GMing.

(I know some GMs like to ask for rolls for no reason to keep players on their toes, but that just erodes trust in the GM in my experience).

Never ask for a roll if nothing interesting is going to happen afterwards. If the players fail and the next response is: "You fail, you can just try again though." Then you're just rolling until someone hits a number, it's not very interesting fiction.

Fail forward doesn't mean that there's arbitrary changes to the world state.

You can set up the stakes of fail forward rolls before asking for the roll:

GM: "You want to climb this cliff, but the jagged rocks are tearing at your rope you've been carrying since 1st level. The DC is quite high because the rope is slick with rain, and fraying even as you climb as fast as you can. DC is 20. You're a good climber though, its more about keeping the rope steady so it doesn't become so frayed as to be useless."
Player Rolls: "15."
GM: "Okay you get to the top, there's a dramatic moment as the rope frays and breaks unfortunately when you pull the rope up after you, you find that instead of 50 feet, only 25 feet of rope remains."

Respecting the fiction is an important element of what Fail Forward is.

A lot of people talk about the survival example. Tracking a monster to its lair.

As a GM start by setting the stakes:
"The DC to track this monster isn't too high, it's fairly big, and the ground is muddy. DC 15 seems fair, because it'll be crossing some streams and rivers. I'll say that if you fail the DC by 10 or less you lose a bit of time, and maybe stumble into a random encounter in the Forest (20% chance). You'll get another opportunity to try again. If you fail the DC by 11 or more you lose the trail entirely. You'll have to trek back to town unsuccessful, and the monster will have another chance to attack the outlying farmsteads."

The point is -
On success the players know what they win: Find the monster in its lair.
On failure: You lose time and maybe stumble into a random encounter. Can try again.
On Critical Failure: The monster will attack another victim. Hopefully you'll be able to set up some kind of stake-out or ambush, or at least you'll have another chance at hunting the beast.

The adventure carries forward no matter what either the players find the beast in its lair, or find the beast when it attacks a farmstead with a different clever plan.

A failed check doesn't just immediately turn into: "You lost the trail. Go home. The beast escapes. Go find another plot hook."

It's more than possible to use fail forward in such a way that there aren't secret off-screen guards who are suddenly alerted, or to bend reality around the players. It just means that GMs need to be far more considered before asking players to roll the dice.

Before asking for a roll:

- What can be won on success?
- What is lost on failure?
- Can failure be turned into success with a further cost (time, resources)?

If you can't answer all three questions, don't ask for a roll, just determine how long the thing takes.


1 person marked this as a favorite.
PossibleCabbage wrote:
John Lynch 106 wrote:
And if they have a silence spell, what fail forward result will you have occur on a failed climb check?
Somebody dropped something, it might not be valuable but anyone who comes by might spot it and figure out that there are intruders about, so "should someone go back down and retrieve it" is a question the party now faces.

And that's why I said fail forward removes player agency and why fail forward teaches players to go with the most obvious solution. Taking precautions is pointless, the GM will make the players face a different consequence. The GM ultimately controls the world and he can always pull out another "consequence" no matter how many precautions you take.

If I caught a DM doing the above (and granted it would take a while for me to catch on) I would stop enjoying the game just as I would if I caught them fudging dice rolls. If you enjoy fail forward as possible cabbage implements it then we simply enjoy playing Pathfinder for very different reasons.


John Lynch 106 wrote:
PossibleCabbage wrote:
John Lynch 106 wrote:
And if they have a silence spell, what fail forward result will you have occur on a failed climb check?
Somebody dropped something, it might not be valuable but anyone who comes by might spot it and figure out that there are intruders about, so "should someone go back down and retrieve it" is a question the party now faces.

And that's why I said fail forward removes player agency and why fail forward teaches players to go with the most obvious solution. Taking precautions is pointless, the GM will make the players face a different consequence. The GM ultimately controls the world and he can always pull out another "consequence" no matter how many precautions you take.

If I caught a DM doing the above (and granted it would take a while for me to catch on) I would stop enjoying the game just as I would if I caught them fudging dice rolls. If you enjoy fail forward as possible cabbage implements it then we simply enjoy playing Pathfinder for very different reasons.

I'm curious as to why that is a removal of agency -- is it because the player character isn't given some sort of opportunity to specifically correct the situation in the moment? Like some sort of reflex save or dexterity check to prevent the object from falling from their possession? (Sorry if this has been hammered on previously in the thread)

I think I get what you are saying for the most part and this thread has a lot of interesting discussion. It seems that "fail forward" is something that is said a lot by the designers, but mechanically does not seem to be supported by the system. How is PF2 is fail forward while having checks for 10 minute tasks like treating wounds or repairing objects? I think every prerelease game I've listened to so far has hand waved these checks once the GM realizes that everyone is sitting there rolling d20s for every increment of 10 minutes against no real threat or consequence. (Do you still damage someone when you critically fail treating wounds? That felt like the opposite of failing forward for me.)

As a player, I don't mind the hydra hunting example. Reaching a dead end can stink, and I think failed checks should mostly mean "you still make progress toward your goal, but it comes at a cost." The easiest cost to choose is time, but other than that it's tough to express a cost mechanically. It feels like the only thing that actually matters is hit points, so you end up succeeding but taking some damage, but then everything just ends up dealing damage, or comes up with a long winded way to ultimately deal damage, just so you can be healed in the next moment, so it starts to feel rather flat over time. Like you can always come up with a billion narrative driven costs, but if its not a thing on the character sheet that is being changed, then it's not really part of the system.

But anyway, sometimes it's about mitigating the real time spent in a session and fast forwarding to the next meaningful decision. If the PCs are pursuing the hydra, and you have to be done with the session in 45 minutes, then they're gonna find that dang hydra.

201 to 250 of 281 << first < prev | 1 | 2 | 3 | 4 | 5 | 6 | next > last >>
Community / Forums / Pathfinder / Pathfinder Second Edition / General Discussion / Fail Forward All Messageboards

Want to post a reply? Sign in.