The Failings of Fail Forward

Like "yes, and..." (or yes-but, or yes-or-roll-the-dice, and so on), fail forward is one of those things that seems to get a lot of praise in the "indie" RPG/storygames camp, but despite being one of the largest (if not the largest) publisher of third-party Dungeon World content is something I disagree with.

Also like"yes, and", I found a variety of definitions for fail forward, but near as I can tell it's essentially a reinterpretation of failure: instead of a character trying to do something, failing, and having to think of some other way to keep going, you ultimately give the character want they want, but you also hit them with something bad along the way.

An "infamous" example of fail forward is when a character tries to pick a lock on a door and fails (though I've also seen examples involving climbing cliffs).

In a normal RPG, if the character fails to pick the lock, either because she rolls low or the Difficulty is beyond her skills, she just fails. Now, in a horribly written adventure in which the only way to proceed towards your goal is through a locked door, and the only way to proceed through the locked door is by picking it, this could be an issue...assuming it's also being run by a horrible DM that never deviates from the adventure-as-written, that is.

(Side Note: Frankly I think this would be an excellent-if-maybe-frustrating learning experience for both the GM and even his players: don't write an adventure with only one specific way to "win" or "move the game forward" or whatever, especially if it hinges on a one-time dice roll. Also, don't be afraid to modify an existing adventure.)

When relying on fail forward, you'd instead rule something along the lines that the character actually does pick the lock, but she triggers a trap, guards show up/are waiting for her on the other side (replace with wandering monsters as needed), breaks her lockpicks, takes a really long time, and so on. Fail-forward proponents claim that by going this route you prevent the game from "grinding to a halt due to one bad roll", but the problem with this is two-fold.

First, the consequences are often irrelevant to the character's capabilities and skills: a trap shouldn't suddenly be present due to a failed lock pick skill check, because a character's lock picking skill has nothing to do with the presence of traps. Same goes for a guard patrol and wandering monsters (and the weather getting worse, which I've seen pitched for characters that fail a Wisdom check while scouting or traveling).

(Another Side Note: One of the posts I read while researching the topic made the absurd claim that an objectively more "exciting" way of handling a lock-pick failure was to simply tell the player that their character totally knows a guy that can pick the lock for them: personally I'd be bored if the GM just spoon fed the solutions to me, but then I have no problem thinking of alternative strategies, failing, coming back later, dying, etc.)

The only two outcomes that seem reasonable at all are how long it takes and the broken lock picks, both of which can be handled in something like Dungeons & Dragons without the DM just arbitrarily declaring them to have happened in order to spare the players from having to think, which leads to issue two: fail forward robs the players of the chance to be creative and figure out a way around an obstacle.

As a GM I love seeing the kind of shit my players come up with. In the case of a locked door, maybe they'll try to bash it in, wait outside for someone to come through and ambush them, disguise themselves as guards, try stealing the keys from the guards, charm one of the guards, tunnel around the door, hire someone with the necessary skills/abilities, or *gasp* come back later when they're more skilled, or have an ability or item that'll get them through (like, say, a portable hole).

Fail forward comes across as an inconsistent "solution" for lazy DMs and/or entitled players that don't like to "lose", or even take the time to come up with a plan B: they wanna get through the door, so just have it open regardless of their skill check, and if they fail have guards suddenly appear for them to fight instead because that's easier (well, so long as the door is barring them from getting on with the "story").

Or, you could let your players try, let them fail, and if they're still alive let them try again. It may not be as quick or as easy, but when they do succeed it'll be more rewarding for both sides of the screen.

13th Age & Fail Forward
Someone pointed me to 13th Age's section on fail forward, so I figured I'd address it specifically since it has some pretty cringe worthy examples. It claims that outside of combat, "true failure tends to slow action down", and that, "a better way is to interpret failure as a near-success or event that happens to carry unwanted consequences or side effects".

The first example has a rogue make a Charisma check in order to befriend an officer on a ship. The rogue fails, and the GM interprets this to mean that rather than fail to befriend the officer, the character did in fact make a good impression, but the officer is now suddenly a cannibal.

The second has a character make an Intelligence check to determine the likely result of a competition. He fails, but instead of him failing to guess correctly, the GM determines that he did guess correctly, but a gang of gnomes controlling the bets lost a bunch of money and for some reason blame him.

So, what does the guy being a cannibal have to do with the character's Charisma check? The same thing a character's Intelligence has to do with a gang of gnomes losing out on a bunch of money: nothing at all.

much better way of handling the first situation is to have the character fail to make a good impression and just leave it at that. If the guy happens to be a cannibal, then he could also set his sights on the annoying character as his next meal. This adds tension or whatever to the game, and doesn't bizarrely punish the character for a completely unrelated reason.

For the second one, I dunno: the character guesses wrong, and then it's up to the GM to determine how the competition ends and thereby whether the character wins or loses money. Winning or losing money is sufficient all by itself, but I guess if the character was making a show of winning a bunch of money maybe someone would try to rob him.

Announcements
DriveThru is currently running a GM's Day sale, so for the next 8 or so days you can snag any PDF in our store for 30% off.

A Sundered World is out (and also available in dead-tree format)! If you for some reason don't want the entire setting, you can just snag the races and classes.

The Ghost has manifested!

By fan demand, we've mashed all of our 10+ Treasure volumes into one big magic item book, making it cheaper and more convenient to buy in print.

Lichfield is available for public consumption. If you want a concise adventure with a Silent Hill feel, be sure to check it out! Primordial Machine is also out, so if you want to catch a glimpse of A Sundered World<

16 comments:

  1. What are your thoughts on the compounding auccesses of the 4e skill encounters in the context of this discussion?

    ReplyDelete
    Replies
    1. @Adam: Oh man, it's been a while since I thought about skill challenges...I'd have to look through the later D&D books/adventures, because that's when WotC finally figured out how to make them work properly.

      Delete
    2. Adam,
      This is a great question. One I've contended with a fair bit in the 2 year 4E campaign I ran. We hacked skill challenges largely by using a modified version of "yes and". I call it "yes BUT". So they may roll low and move forward, but they don't "succeed". Instead, they move to another situation they must overcome. An example might be a Cleric attempting to channel holy energy into a relic to fuel a ritual. In 4E parlance, this might required 3 successes out of 5 tries. So, what do we do with the 3 potential failures? Well, the example I always go back to is that scene in Return of the Jedi where they're trying to "pick" the lock on the door on Endor. Han says, "I think I got it" as he sparks two wires. And then, the blast doors CLOSE! This was not "yes and" and guess what? It was brilliant! This is what I tried to emulate in my 4E skill challenges. So whereas the Han character might have said to the DM "I try to short circuit the lock". He rolls low and the DM says, "the blast doors close!". Great now they have to open another set of doors and we get the awesome scene where Han impersonates a Commander so they open the doors from inside. Similarly, the Cleric in the D&D example might fail to channel the energy properly and instead of fueling the ritual he speeds it up! Now, he has less time to complete the next challenge or something equally confounding. SO it's a failure, but the challenge still moves forward.

      Delete
  2. This comment has been removed by the author.

    ReplyDelete
  3. Let’s keep going with picking a lock. If you’re failing forward and the thief fails to pick the lock, having the lock not open is a perfectly viable option. What fail forward means is the PACE of the game never slows down because of a die roll. Thief fails to open the lock, and suddenly orcs show up to ambush the party! Fail forward doesn’t HAVE to mean “narrative from failure;” it can also mean “narrative, regardless of dice rolls.”

    Another thing to keep in mind here is that, for better or for worse, the community has largely coupled the concept of “failing forward” to a single role-playing system, the so-called “Powered by the Apocalypse” system. In those games, the GM has a list of different moves he can make, and a specific time when he can make them. If everyone is waiting on the thief, and the thief attempts to pick the lock, and fails, and everyone is looking at the GM? That is literally the textbook example of when a GM performs a move. In this example, let’s say the GM chooses “Show signs of an approaching threat.” Again; thief fails to open the lock, and suddenly orcs show up to ambush the party! Again, “fail forward” can simply mean “the game continues, no matter what the dice say.”

    ReplyDelete
    Replies
    1. "Let’s keep going with picking a lock. If you’re failing forward and the thief fails to pick the lock, having the lock not open is a perfectly viable option."

      I'd go so far as to say that the lock not opening is not only a viable option, but should be at least part of the result, if not the entire result, because that's the purpose of the skill check: to see if the character can open the lock or not.

      "What fail forward means is the PACE of the game never slows down because of a die roll."

      I'd argue that the pace of the game doesn't slow down because of a failed lockpick roll: characters tries to pick the lock, fails, and now they gotta think of another way to deal with it, and/or move on to something else.

      "Fail forward doesn’t HAVE to mean “narrative from failure;” it can also mean “narrative, regardless of dice rolls.”""

      How does a failed lockpick check resulting in an unopened chest or door NOT mean "narrative"? Why call the term "fail forward" if it can mean something happens no matter what you roll? In this context it sounds useless if not redundant: RPGs have always had things happen when characters do things.

      "Another thing to keep in mind here is that, for better or for worse, the community has largely coupled the concept of “failing forward” to a single role-playing system, the so-called “Powered by the Apocalypse” system."

      I know PbtA games are big on it, but I also assumed it was a part of FATE and heard that even 5E D&D had a bit on it, and the examples at the end of the article are lifted from 13th Age (which specifically credits Ron Edward, Luke Crane, and "other indie RPG designers" for its inception).

      Given the player base of 13th Age (and whatever Ron Edward and Luke Crane have done), I don't think a majority of the community specifically thinks of PbtA games when it thinks of fail forward, and I don't think that everyone defines it the same way.

      "In those games, the GM has a list of different moves he can make, and a specific time when he can make them. If everyone is waiting on the thief, and the thief attempts to pick the lock, and fails, and everyone is looking at the GM? That is literally the textbook example of when a GM performs a move."

      As one of the largest, if not THE largest, publisher of third-party Dungeon World content, I'm very familiar with how the game works.

      There's no confusion about what the rules say and what you're "supposed" to do in the game: my issue with fail forward is that at best it seems wholly unnecessary, at worst tries to convince GMs to hit players with unrelated consequences instead of having them think.

      "In this example, let’s say the GM chooses “Show signs of an approaching threat.” Again; thief fails to open the lock, and suddenly orcs show up to ambush the party! Again, “fail forward” can simply mean “the game continues, no matter what the dice say.”

      But even if orcs don't show up the game STILL continues. I'd say NOT having orcs show up is more meaningful, because as a player I wouldn't know that the only reason they're there is because of a completely unrelated skill check.

      Delete
  4. First, thinking from a PbtA mindset, I don't see "fail forward" but success-with-consequences and failure. In my mind, that nuance means a world of difference - failure is real and you may not open that door.

    Next, it seems to me that the majority of this dialogue is focused on convenient but, perhaps, inappropriate consequences of "failing forward". Throwing an Orc patrol at the party cuz the thief failed a lockpick roll is lazy GM'ing... unless it's narratively appropriate. Maybe it was already set up that they were being pursued and had limited time for thievery?

    In DW, a more capable GM would clearly see that there are other moves that would be more appropriate within the narrative, in addition to the failure.
    * use up their resources (broken tools)
    * reveal an unwelcome truth (trapped, also)
    * show a downside to their class (the mechanism needs the brute force of a fighter)

    Perhaps "fail forward" is actually a mechanism to save struggling DMs? Imagine that the thief falls on that lock - holy crap! Now my whole session/module is blown! Oh, whew, "fail forward" lets me fudge things a bit as I learn by trial-and-error to be a better GM.

    ReplyDelete
    Replies
    1. @allinonemove: A guard patrol/wandering monsters was but one of several possibilities.

      I included it because it's an example I have seen, and it's as illogical as "the weather gets worse" example I've also seen for characters that fail a scout/travel move in DW.

      Breaking lock picks or having the lock suddenly be trapped in order to have the door be open is just as lazy as throwing orcs at them.

      It's also inconsistent, unless you have every lock open regardless of the check (so long as you break something of theirs, or hit them with a trap, at any rate).

      "Perhaps "fail forward" is actually a mechanism to save struggling DMs? Imagine that the thief falls on that lock - holy crap! Now my whole session/module is blown! Oh, whew, "fail forward" lets me fudge things a bit as I learn by trial-and-error to be a better GM."

      This could be a VERY important learning experience for the GM: don't bar your "adventure" with a door that can ONLY be opened by picking the lock.

      I think it's better to just let the players fail and figure out their own way in (and maybe next time don't put in a door that can only be bypassed via lockpicking).

      Delete
  5. I usually do "fail forward" (that term is growing on my by the way) as "you can do this but you need to.." So with the lock picking example, I might say spend more time with it, bend a lock pick, have someone help move the door a bit, or (of course) try again.

    ReplyDelete
    Replies
    1. @Sectorbob: Something I miss from of all things 3E D&D, are the rules for taking 10/20.

      So, when you tried to pick a lock you could just take 10 and get a result of 10 + your Open Lock skill. Not good enough? Okay, you could take 20, which assumed you took 20 times however long an Open Lock check took to get 20 + your Open Lock mod.

      STILL not good enough? Whelp, that lock is just too damned good for you. No rolling over and over in the vain hopes that you'd get a nat 20 or whatever. No GM arbitrarily saying you can't do it, or you can, or whatever.

      Not that that necessarily stops you: back when I ran Age of Worms for the first time, the group didn't have a rogue, so the warblade had to spend several minutes or so bashing through all the locked doors they came across.

      Delete
  6. After rereading the examples in the 13th age book, I'm struck by one thing. They basically say to grant success even on a failure. I think that is where they're making a mistake. I think it would be less confusing if they simply said, "make sure the failure is interesting". That's really what I try to do, just to make sure that even failed rolls don't turn the game into a slog. But IMO failure need not be boring. I think some of the above comments are right. The thing that people are really trying to overcome with "fail forward" or whatever is poor adventure design or DM's who are not experienced enough to handle situations outside the binary. Love this conversation btw.

    ReplyDelete
    Replies
    1. @mbeacom: As I said with the whole "yes, and..." business, there are a variety of interpretations for fail forward, and most of what I've seen is in the "make a failure not really a failure as long as it's important so your adventure doesn't stop" ballpark.

      "I think it would be less confusing if they simply said, "make sure the failure is interesting"."

      Can you give me an example of a failure that just isn't interesting?

      "The thing that people are really trying to overcome with "fail forward" or whatever is poor adventure design or DM's who are not experienced enough to handle situations outside the binary."

      I'd argue that fail forward doesn't educate DMs on how to avoid those pitfalls.

      Actually, the closest I get to fail forward is maybe doing whatever DW says to do with the 7-9/miss results, so I guess that's something to ask you: has fail forward helped you avoid those sorts of scenarios?

      Delete
    2. "Can you give me an example of a failure that just isn't interesting?"

      1. You attempt to pick the lock and fail.
      2. You attempt to scale the rock wall and are unable to find proper handholds.
      3. You hurl the butt of your torch in an attempt to break the crystal fueling the ritual and miss.
      4. The trap you've set doesn't trigger.

      They are infinite. All of those situations CAN be interesting but when failure is described as nothing more than "your goal was not achieved", it can feel uninteresting. Where I think 13th age goes wrong, (and the indie revolutionaries pushing this orthodoxy) is that all that is really needed is better descriptions of failure. Rather than granting success with some impromptu challenge.

      "I'd argue that fail forward doesn't educate DMs on how to avoid those pitfalls."

      I would agree with that statement. I don't see it as an educational tool so much as a crutch for DMs who might struggle with handling failure in an interesting way.

      Delete
    3. @mbeacom: Personally I don't find the above examples uninteresting, so long as the were the only way to do a thing and now that we've failed we can NEVER open the lock or get around the cliff.

      I think THAT'S what makes things uninteresting: there's a cliff, and you MUST climb it, so we're gonna sit here while you all succeed at a string of Climb checks, and if you fail too bad adventure's over.

      "Rather than granting success with some impromptu challenge."

      It really sounded like "success-at-a-cost", but it was in the fail forward section. What bugs me about it is that it's not consistent, because 13th Age ALSO says that in combat it's "okay" to have failures with nothing else tacked on.

      "I would agree with that statement. I don't see it as an educational tool so much as a crutch for DMs who might struggle with handling failure in an interesting way."

      Yeah, it would be more useful if it would say something like "here are things not to do, but if you happen to do these things here are ways to sidestep them and keep the game going".

      As it stand it just feels like "do these whenever".

      Delete
  7. Many times in my games the thief tries to pick to lock, fails, and then says, can i try again? He keeps rolling until he gets a success. Maybe this failing forward can help make this situation more interesting?

    ReplyDelete
    Replies
    1. @Anon: Nah.

      In older editions of D&D, if you tried to pick a lock and failed, you could only try again when you leveled up (I think there were optional rules for trying additional times based on your stats or something else).

      I think the idea was that failing a lock pick attempt meant that it was currently beyond your skill, but I dunno since if you were a thief you didn't HAVE to put percentage points in a lockpicking skill.

      In 3rd Edition D&D this is all easily handled via take 10/20 (and the fact that skills state whether they can be tried again): very easy for a player try taking 10, and if that doesn't work you can just say "I'm taking 20 on my Open Lock check", and from there the player will KNOW if it's just beyond their skills.

      I think 4th Edition D&D, mebbe 5th Edition, only had the take 10 option, but I always houseruled taking 20 in as well since it just makes sense and avoids the GM having to "force" drama or have the character succeed anyway because "plot".

      Delete

Powered by Blogger.