We Don’t Need Failure States After All

November 23, 2022

It’s Bulls$&% time again.

The other day, I watched a video published on YouTube by Adam Millard, Architect of Games. And he’s recently become my favorite video game design commentator on YouTube. That’s right, he even replaced Game Maker’s Toolkit. Sadly, GMTK’s been kind of hit or miss after his peak with Boss Keys. But I digress.

So, I watched this video. And I’ve now watched it three times through because my mind became a swirling miasma of scintillating thoughts and turgid ideas, to quote a certain psychopathic cartoon lagomorph. Once the literal brainstorm quelled — sometime after my second viewing — I found I had a totally different perspective on something lots of gamers have been fighting about for years.

Today, I’m going to turn the fight about failure states in RPGs entirely on its head. Because we’ve all had it wrong for lots of years. All of us. Even me. And especially you.

Designed to Fail:
Failure and Loss in Game Design

So, there’s this channel on YouTube called Adam Millard, Architect of Games. The titular Adam provides amazing game design analysis and critique. And the other day, he posted a video called Why Video Games Want You to Fail. Unsurprisingly, the video was about how game designers handle failure and loss. In video games. He talked about why loss and failure are important and why it’s important to handle them just right.

I’m gonna embed the video below because I think it’s worth watching. And because it’ll provide some context and background for the bulls$&% I’m about to spew. But I don’t think his video’s useful for TTRPG designers and game masters. I don’t think his argument applies to the TTRPG game space.

So, why did I watch it three f$&%ing times? Why am I sharing it? Why am I devoting thousands of words to it?

First, because the s$&% he says exactly echoes the s$&% lots of us in the TTRPG space have been saying for years. S%$& we keep saying. Yes we. I’ve said it all myself. Most of it anyway. That’s why I warned you against watching the video. If you watch it, you’re going to think it’s brilliant. “Wow,” you’ll say, “those sure are some great ideas and we should apply them to TTRPGs immediately!”

But the second reason I’m sharing the video — and writing a three thousand word essay about it — is because hearing Adam Millard, Architect of Games say all those things — which, by the way, are perfectly clever, logical, correct things to say about video games — hearing Adam Millard, Architect of Games spew the same pissing and moaning about death and loss and failure that GMs have been spewing for decades made me realize just how stupid it all sounded when applied to TTRPGs.

This one’s going to be a roller coaster. But ride it out with me, okay? I promise it’ll all make sense in the end. And it’ll give you permission to stop worrying about death and loss and failure in TTRPGs. Because it turns out, we don’t actually need that s$&%. We’ve got something better.

Anyway, here’s that video I told you about. Watch it if you have time. It’s good. But if you can’t watch it, no biggie. The article will still make sense.

Stories Losers Tell:
The Standard RPG Failure Narrative

Over the past several decades — yeah, decades — there’s been lots of talk in the online TTRPG space about failure, loss, and death in TTRPGs. Most of it’s been messy, complicated, confused, and loaded with nonsensical bulls$&% conclusions. Trust me, I know all about these conversations because I’ve been there. I’ve put forth my own arguments, started my own threads, and called out lots of nonsensical bulls$&%. Bulls$&% like “fail forward” and “make failure more fun than success.”

Gradually, there emerged a standard set of talking points — or maybe, I should call them pissing and moaning points — about failure, loss, and death in D&D and other TTRPGs. A kind of shared narrative about how TTRPGs handle that s$&%. Or fail to. And I’m going to distill it all down to its essence.

If you’ve ever talked about failure, loss, and death as game design ideas in TTRPGs, you’ve heard — or said — something like this:

Roleplaying games need failure and loss. Players can’t win every challenge no matter what. But losing isn’t fun and players don’t like it. So failure has got to exist, but it’s got to be fun. At least as fun as succeeding. Maybe more fun, even though that’s an utterly moronic idea. Moreover, failure can’t be a total bottleneck. A single failure can’t — or shouldn’t — ruin the whole game. Whatever the f$&% that means. And failure can’t be punishing either. The problem with TTRPGs — especially D&D but also especially lots of other TTRPGs — is that they really only have one failure state: death. And death is a bad failure state. The worst failure state. And maybe it shouldn’t even be possible because death makes players sad and besides, unexpected and random deaths ruin stories and also character generation takes a long time. So, obviously, D&D and its cousins need more better failure states than just death and those should be built right into the system even though neither I nor anyone else can explain just what the f$&% that would actually look like.

That’s it. Every conversation about failure, loss, and death in TTRPGs ever. More or less. With some editorializing, of course, because I’m me and editorializing is fun.

That video I mentioned above by Adam Millard? It analyzed failure, loss, and death in video games from a game design perspective. And it was very compelling. And I found myself nodding along. Because it basically echoed — or at least justified — every conversation ever about failure in TTRPGs. It argued that failure and loss are necessary — even vital — gameplay elements. But they’ve got to add to the experience, not detract from it.

Except what Adam Millard proved without even arguing it is that none of that s$&% really applies to TTRPGs at all. That whole narrative? It’s a bunch of useless bulls$&%. We don’t need it. We’re better off without it.

Making A$&es of Ourselves:
The Missing Assumptions in the Failure Narrative

So, there I was, listening to Millard and nodding along. “Yep,” I said, “we have exactly the same problem in the TTRPG space! I sure can’t wait until you put forth an elegant solution so I can port it into the TTRPG space and take all the credit for once again revolutionizing all my readers’ games.”

But the longer I listened, the more I realized something was off. Not in his arguments, but in the narrative I had in my head about failure in TTRPGs. And I’m not referring to the snarky editorial bits. I mean the narrative itself was based on a lot of assumptions. And it glossed over a lot of important distinctions.

Is there a difference between a failure and loss? How are we even defining failure? Are some failures okay and others not? Why? Is there a difference between failing an action and failing an encounter and failing an adventure and failing a campaign? What’s the difference? If there is — and there are — what kind of failure are we talking about when we demand D&D provide us better failure states? When we talk about death, do we mean character death in general or are we talking strictly about TPKs? Are character deaths really failure states? Why is bottlenecking bad? Why is it bad for an adventure to fail as the result of a failed action? Isn’t that always how it’s going to work? Doesn’t every adventure fail, ultimately, because of one failed action too many? There’s always a last failure. Isn’t it kind of like losing your last hit point? What would a failed adventure look like if that failure wasn’t the result of one failed action too many? And would it even be possible to quantify something like that given the emergent and holistic nature of narrative-driven games?

Round and round went my brain, picking at the narrative one detail at a time. I probed every crack, demanding answers of myself and then demanding justifications for the answers. How do I know that? Why do I think that? If that, then what about this? Is that true? Why? What’s the difference?

I’ve covered a lot of that ground before. I’ve written about failure in the past a lot. About how adventures and campaigns and encounters have to be “losable.” About how failure must always be an option.

But must it?

If You Can’t Lose, Why Try:
Why Failure Matters in TTRPGs (Maybe)

The whole Failure Narrative above is predicated on one simple assumption: that any TTRPG must include the possibility of failure. And that assumption comes literally from the TTRPG acronym itself.

A roleplaying game is a game of choice, right? If the players ain’t making choices — not just decisions, but actual f$&%ing choices — if the players ain’t making choices, they ain’t roleplaying. And that means they ain’t playing a TTRPG; they’re just playing a TTG.

See, roleplaying’s about playing a role, right? That’s why we call it role playing. Every character — and every real human — is ultimately defined by how they resolve internal conflicts. When the character’s got conflicting goals or values or ideas or desires, how do they pick the more important one? That’s what makes a character who they are. Whether they’re fictional puppets or real human boys.

Underlying RPGs is the idea that those choices matter. They’ve got stakes. They’ve got consequences. If a character can’t actually die, the character’s willingness to risk his life for his beliefs doesn’t mean jack s$&%. If you never have to give up the thing you don’t choose — if the choice doesn’t cost you anything — you’re just choosing “all of the above.” You can stay true to your values and live. Lucky you.

It’s costs and risks and consequences that give choices their meaning. If the characters can make whatever choices they want without ever stressing how s$&% might turn out or what they might have to live with as a result, the characters lose their humanity. And the game becomes a sure bet. So whether it’s the RP part or the G part of the RPG you care about most, you’re s$&% out of luck.

That’s how it be. No risks, no costs, no consequences, no RP and no G, so no RPG. And while you can have fun with that s$&% for a while, you can’t really care about it for long.

And it’s on that basis that GMs in the online RPG space — me included — argue that failure must be an option.

Don’t Go Hollow:
What Video Games Can’t Teach About Failure in TTRPGs

Right, let’s switch back to video games for a minute.

For a lot of years, I’ve argued that RPG GMs and designers can learn from video games. Hell, I personally think video games are the single best medium to study if you want to git gud at game mastery. Better than books and movies. Better than board games. Often better than the f$&%ing GMing guides ostensibly published to teach people how to run games.

Often, but not always. And in this case — the case of failure and loss and death in game design — I don’t think video games are the right thing to learn from.

Video games have a complex and tempestuous relationship with failure. One that’s usually oversimplified and misunderstood. For instance, lots of idiots claim video games only include failure and death because arcade cabinets used to charge you for every attempt you made to win. That’s based on truth, but it ain’t true. Want proof? First, games existed before video games and you could lose at those ones too. And second, video games have evolved away from life counters and game over screens — mostly — but they haven’t dumped failure or death.

But you know what’s interesting about death and failure in video games? It’s that, however they work in any given game, failure and death don’t lead to real, permanent failure. You never really lose a video game. Not anymore. And even when you did — when game over screens were a thing — the game didn’t explode or kill you for real or anything. You just hit the reset button or dropped another quarter and tried again.

You only really lost the game when you gave up. And until you did give up, there was a path to victory. Usually, it was, you know, playing again armed with the skill and foreknowledge that each previous attempt granted.

But that ain’t how TTRPGs work.

So, when we talk about failure in RPGs, what are we talking about? Really? What do we really want? Do we want points in an adventure beyond which the players just can’t succeed? Beyond which the princess can’t be rescued or the treasure can’t be found or the world can’t be saved? A real, honest-to-Gygax Game Over screen? Do we want to say, “well, you f$&%ed that campaign up royally. Grab some character sheets and let’s see if you can win the next one.”

Maybe. I’ve done that s$&% myself. “Sorry, dingdongs,” I’ve said, “I wasn’t kidding when I said the world was hanging in the balance. You blew it. The world is gone.” And, in the end, that’s just a different kind of TPK, isn’t it? “Sorry, you’re all dead; there’s no way to continue the game. Closest we can get is a new game in the same continuity with new characters to pick up the torch you losers dropped.”

If we’re going to figure out failure and loss in TTRPGs, then we’ve got to figure this s$&% out for ourselves. Because the video game model doesn’t work. The one where you fail and die but get to respawn or reload and try again. Or failing that, exit to the menu, select New Game and give it another go.

Though maybe there is a kind of video game model that works. And maybe we’ve got it already.

Picking the Right Model:
What Video Games Can Teach About Failure in RPGs

That Adam Millard video starts with this anecdote about failure. Or about not failing. He and his buds were playing this sci-fi sandbox survival thing called Space Engineers. That’s an original f$&%ing name, huh? Anyway, Millard and Company had crashed their awesome spaceship and stranded themselves. Rather than giving up and starting over, they kept playing. They turned their ship into a surface base, plundered the local environment for materials, and eventually saved themselves. And Millard points out this kind of emergent story s$&% is a hallmark of open-world survivally simulatory colony-buildy games like Dwarf Fortress and Rimworld and s$&% like that.

The thing about games like those is that, while you can quit whenever you want, you usually can’t reload a previous save or respawn and try an individual challenge again. If you f$&% up, you’ve got to live with that f$&% up. You’ve got to salvage the bits of the S.S. Dumba$&% and start hunting dilithium and boar spleens to build a new drive core.

Now, I should point out that most such games don’t usually have an explicit win state. There’s no end goal built into the game. And if there is, the game’s all wishy-wishy and laissez-faire about it. “You can try to reach the center of the galaxy if you want, but hey, if you just want to make the coolest ship you can and then point it toward the farthest star, that’s cool too. Whatever.”

And maybe that’s important. But I don’t think it is. What’s important is that those open-world-survival-and-simulator-and-colony-management games share two vitally important characteristics with TTRPGs: Persistent Continuity and Memory.

In a TTRPG, the game’s events form a consistent narrative, right? Every session — and every adventure — follows from the last. TTRPGs aren’t like board games where each session represents a distinct and isolated attempt to win. TTRPGs are about building a continuous narrative one session at a time. Moreover, when something happens in a TTRPG, the event becomes part of the game’s history. S$&% that happened last week? Last month? Last year? The game remembers that s$&%. And so do the imaginary characters that live in the game’s world.

Thus TTRPGs play out in worlds that react to every action. And evolve. If the princess died last week, the world doesn’t have that princess anymore. She’s gone. And if some cult killed her to free a demon prince, the world’s got a demon prince running around in it. And if the dumba$& players crash their spaceship, they can’t reload or try again. They’re stuck. And they’ll have to deal with that s$&%. Or quit. And if a character died in the crash? Well, the party’s got to get along without his skills and talents. But maybe they’ll meet some other crash survivor to fill out their roster.

And that combination of elements — Persistent Continuity and Memory — that s$&% changes the whole Failure Discussion.

The Broken Narrative:
RPGs Don’t Need Failure

Let’s go back to that Standard Failure Narrative thing. It starts with “actions must have consequences; therefore, failure must be an option.” The first line? That’s definitely true. It’s a TTRPG axiom. Actions must have consequences. But does it logically follow that failure must be an option? Is that the only way to ensure actions have consequences?

As established, TTRPGs have those Persistent Continuity and Memory things, right? At least, they should. It’s the GM’s job to keep the clock running forward, maintain unbroken continuity, and ensure the game world’s history continues to have happened as it did. Doesn’t that ensure that actions in TTRPGs do have consequences? They might not always be big or far-reaching, but they’re there. The consequence of a character face-tanking a trap is that they recover some treasure and buy a slightly better sword. That ain’t much. But sometimes, the consequences are way bigger.

When the PCs save a village from a dragon, the world remembers that s%&$. And it should act accordingly. The villagers should be grateful. The kobolds should hunt the PCs so they can harvest the heroes’ blood for use in a ritual to resurrect their fallen god. That’s how this s$&% works.

And beyond a few tracking tools that make it easier for the GM to manage some of the more common consequences — like enmity or reputation or some s$&% like that — this crap can’t be modeled at the system level. It’s not something you can do with rules. Why? Because there’s an infinity of things the players might do. And therefore, there’s an infinite possibility space for consequences. Figuring out after each session how the world’s different and what happens next? The system can’t do that s$&%. That’s why GMs exist.

But the real kicker here is that if you do all this s$&% right — if you keep the clock running and remember the world’s history and all that crap — then maybe it doesn’t matter if the players can’t lose an adventure. Or a campaign.

Because the world evolves, everything the players do matters. Whatever’s happening in your game today? It follows from everything that happened before. And if everything matters, then you don’t need to go out of your way to make certain things matter. They already do.

In my last adventure, the heroes went into a haunted forest and appeased a spirit by killing a savage that was despoiling its demesne. They also met a woman who’d gone blind because the forest was all haunted and despoiled and s$&%. And they also met a guy who was accused — possibly falsely — of arson.

The heroes won. They killed the savage, appeased the spirit, and un-despoiled the forest. Huzzah and forsooth; hail the conquering heroes. But the truth is they really couldn’t lose. The adventure was straightforward as hell. NPCs led them into the forest and pointed them to the spirit’s demesne. And when the heroes got there, it was pretty clear what had the forest spirits all hot and bothered. Short of dying in battle or just giving up and walking away, the players couldn’t lose.

But the heroes also had to put a blind woman in elder care — sort of — because she couldn’t live in the dangerous forest on her own anymore. Not blind as she was. And they didn’t dig into the arson case. Those aren’t failures, they’re just things that crossed the PCs’ paths. S$&% in the world. Might things have gone differently if the players had made different choices? Undoubtedly. That’s how RPGs work.

The point is, when the adventure was over, there was an arsonist in their hometown’s jail and a blind herbalist down the street. That s$&% wasn’t true last week. Some of it might have become true regardless of what the players did. The arsonist would probably have ended up in jail no matter what, but the herbalist probably would have died.

I didn’t build that s$&% into my adventure as failure states. They were just characters and stories for the players to see — and interact with — on the way to their pretty much predetermined success. I sure as hell didn’t say, “how can this adventure fail without me killing everyone?” I just said, “what s$&%’s in the world on the way? What other stories exist alongside this one?”

And when the adventure was done, I said, “okay, how are things different in the world now.”

I ain’t saying you shouldn’t include failure as a possibility. By all means, include it if you want to. I do. Sometimes. Sometimes I don’t. Hell, I usually don’t sweat it one way or another. And that’s what I’m saying. You don’t need to sweat it either. And all the pissing and moaning about failure states in D&D is likely a waste of breath. Because the idea of systematic, mechanical failure states in D&D? They’re probably impossible, but they’re also totally unnecessary.

Provided, of course, the GM has a good memory. Or takes good notes.


Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.

47 thoughts on “We Don’t Need Failure States After All

  1. Imo, the only real failure state in a ttrpg is for the gm to “lose the narrative”. In my experience, player failures mean very little, gm failures kill campaigns and play groups.

  2. Right, this all makes sense.

    It seems to me that failure states are less important if there are peripheral elements, such as the herbalist and alleged arsonist, to make the world come alive but become more important if there is nothing “distracting” the pcs from just walking up to the bad guy and beating him up.

    Take the arsonist and herbalist out of the equation and, like you said, the adventure is incredibly straightforward. So much so that if you have three of those in a row you’d get bored. Failure states can spice things up here. However if the gm -angry or otherwise- is skilled in sprinkling these little extra elements on top the whole thing can stay fresh and interesting as the game suddenly becomes non-linear.

    Mix the two up and I expect you could run a very long campaign that never gets stale.

    Does that make any sense?

    • What is ‘failure’ though? For example, say you’re negotiating with a Baron and you get a crap roll. Does that mean the Baron doesn’t give you what you want? Or does it mean the Baron gives you the bare minimum of what you wanted and you owe the Baron a favor? Now the character has something hanging over them and the GM has an existing hook for an activity the Player ought not take try to avoid when the GM has the Baron call in their chit. Sometimes ‘failure’ is just not having complete control over your future.

  3. I’ve started thinking less about whether the PCs can lose, and more about whether the Bad Guys can win. I’m sure we’ve all played and/or run that adventure where the annoying cult of getting things stuck in drawers or whatever is hanging out in their lair waiting to start the ritual until right before the PCs are seconds from kicking down their door. We could put some sort of clock on it, say that the PCs will fail if they don’t get there in time to stop the ritual, building in a failure state. But I’ve found it much more useful to think about what the cultists are actually doing. What does the ritual entail? Is there some sort of specific time it has to happen? Who needs to be involved in what capacity. This sounds like a lot of work typing it out, but the key is really that the Bad Guys are actually doing something. If the adventure is basically a still life of villainy, then you feel that urge to build in a specific, preordained failure state, but I think this a reaction to the lifelessness of the adventure, an ill-conceived attempt to build in the illusion of a dynamic situation.

    • Which is a great way to obsess over failure in a game with bad guys advancing a plot. But what about adventures where bad guys aren’t advancing a plot? That said, you’re capturing the spirit of what I’m saying: it isn’t the loss state that matters, but instead, just knowing what the world will look like if the bad guys win.

      • Basically, the “loss/fail state” can be summed up as: “A change in the world caused in a way that is not desirable by the players/party.” Like in your example having the princess sacrificed by the cult to bring forth the demon, when the party was tasked with saving her.
        Specifically, it is a result state regarding a specific event/task, resulting in some consequences, likely negative for the party.

    • Having moving villains, and complications is something that the Tension Dice articles really helped me with.

      Not in the context of the players failing…but in the context of, “the bad guys are doing something too”.

  4. Sort of tangential, but what about the “failure state” of a PC getting killed in a fight?
    The main thrust is that (for example) D&D 5e is built around combat as a central pillar of play, but when a player messes up they just take a bit more damage than they might have, or they go down. Then they get a bit of healing and are right back up again as if nothing happened, there’s almost no “failure” state at all, until a PC gets those 3 failed death saves and they’re out completely, until they make a new character that gets introduced or they get resurrected.
    I feel like this system is a bit punishing for its failure state of character death, so the system goes out of its way to try and make that a hard state to hit. Combine that with no lasting consequences of anything short of that final failure state once you get a long rest and I think that combat loses a lot of its potential impact for forcing play to adapt.
    I understand how you get the more interesting continuing play at the macro level, but at the micro level it feels lacking.
    This feels like a much more mechanical question, that may need more mechanical answers.

    • This is 100% just a question of mechanical balance (which can be summed up by DnD 5.0 healing is underpowered, except when it’s not) that has very little to do with the article in question. Extrapolating from the article, which is less an article and more just a random musing without a clear conclusion, pc death isn’t really a failure. Much like how saving the blind woman wasn’t a “success” but rather just “what happened”, a pc dying isn’t a “failure” it’s just another “thing that happened.”

      • It’s a long-running issue with D&D back to its old school days. Because you fight just as well at 1 HP as 100, the best strategy (for PCs and villains alike) is to pile on each enemy one by one until they’re completely dead. If being at 1 HP left you crawling around helpless and magical healing didn’t exist, characters wouldn’t waste their attacks on nearly incapacitated characters; then a party that narrowly won would just be forced to take a few days to recover, not go off and get some Raise Deads. That said, it is just a mechanical issue and there are plenty of benefits to the HP system too.

  5. I’m glad you said this stuff. I also wished for another few thousand words on why it’s ok to kill PC’s and how it’s not the GM’s job to solve problems for the characters. I’ve been frustrated for years with people talking about failing forward and how killing PC’s is bad for the story.

    • I think you can summarize the case against character death as:
      Many games, in their published adventures and their direct advice to the DM, encourage a structure and continuity where the player’s only emotional investment in the game world is in how these events effect their one character. Which means that if their character dies, you have 50/50 odds the player just stops caring one way or the other if the adventure succeeds or fails. And if it doesn’t it’s because you’ve tacked a whole epicycle onto your initial structure to get investment in this new single character and to tie that character into the structure. And then this same game, that has handed you a structure that can’t survive a TPK without revision, tells you that the players should be in mortal danger every single day of their lives.

      • It’s an issue with modern adventure design, and players, in my opinion. Too much of it is very “epic adventure” in nature. It’s Lord of the Rings, not Game of Thrones. If Frodo and Sam dies on their way to Mt. Doom, then the story ends. If someone dies in Game of Thrones the story goes on. Heck even beloved characters dies in Game of Thrones!

    • I hate this modern focus on “the story”.

      I keep thinking about my favorite Angryism; “The story is just something we tell after the game”.

    • 100% agree, I’ve DM’ed for some well intentioned but decidedly poor players who put their own damned selves in trouble, and they paid for it as anyone who is that careless should

  6. I am a little confused. That may be, because I’ve worked a long day and I’m on the train for an hour and I have to work on Saturday now and I didn’t have to yet yesterday, which means my weekend is f^%$ed.

    Anyway, I remember a sage GM saying, “start writing your adventure from the butt end. How can it turn out”? Which includes “failure” – which jives with the article. The party “failed”, so the world is now ruled by the evil princess instead of the noble dragon. The world remembers that s^$ and moves on, oblivious of the losers.

    What about actions, though? The same sage GM said, “if the action cannot fail, don’t ask for a roll”. If an action can fail, so can a series of actions. And a string of series.

    Are you talking about two different kinds of failure? More than two, maybe? Or do I have end-of-workday dementia (absolutely possible)?

    • I believe what Angry is talking about is, instead of worrying about “fail states” and “win states” GMs should simply be concerned about the result of the players and NPC’s actions; “failure” and “success” will emerge naturally as the state of the world and the players’ wishes collide

      Using the example from the excellent Adam Millard, there is nothing in Space Engineers that requires or wants you to keep a ship in orbit in particular, but crashing the ship is still a “failure” nonetheless because the players wanted to have a ship in orbit but their choices and their consequences (as dictated by the game) led to the ship being stuck on the ground

      I also believe that this is more relevant at a higher narrative level rather than an action-by-action level

    • Uh huh? You see how this mess happens? I mentioned precisely that in the article: that discussions of failure never ever start with discussions about what failure means and the different kinds of failure. I then didn’t discuss that but expected you all to infer from context which particular type of failure I meant.

  7. This article is making me rethink a lot! I think it helps me see why I struggle to actually enjoy TTRPGs even after years of playing.

    Take character levels and loot in D&D; they support remembering repeat profits gained in a hostile environment, and that environment doesn’t really need to change much or have much ‘memory’. Character deaths in such a game make sense, because the primary memory system is tracking each individual character’s progress. In fact, it’s almost better for the hostile environment to NOT change, because when it changes it undermines the player’s knowledge for avoiding character death in the same way later.

    This is very ‘video-gamey’, though. It doesn’t make much use of the GM’s flexibility and creativity. It’s boring to run. Video games do this style of game so much better.

    I totally agree with your concepts here, especially that continuity is its own reward in a TTRPG, which is precisely the thing that I’m realizing a dislike due to personal taste.

    • I’m sure no one is likely to read this since the article is old, but just in case… My RPG world turned upside down after this article. I run things totally differently, now. Here is my new procedure:

      Players decide when to gamble the dice. The GM just sets the scene and adjudicates results.

      When the players describe what they do at pivotal moments, the GM tells them the good things and bad things that will happen from the action. The players can choose to settle for the outcomes or gamble, but if they gamble then they must risk rolling dice to get only good outcomes or only bad outcomes (success/failure).

      This allows the GM to create devastating failures that the players opt into risking.

      Super simplified examples:
      GM: The door is locked.
      Player: I pick the lock.
      GM: Sure, but your lockpicks will break.
      Player: No, I want to keep my lockpicks in-tact.
      GM: Okay, roll to pick the lock and keep your picks, but if you fail then you cannot pick the lock at all and your picks break.

      GM: The NPC tells you she’ll let you stay here if you agree to remain and protect her from the next bandit raid. If you agree, then we’ll play out the bandit raid.
      Player: I want to convince her that I’ll help her, but I’m not planning to do it.
      GM: So, you’ll need to roll Deception, but if you fail then she won’t let you stay at all and she’ll turn hostile toward you.

      GM: There are a couple of guards watching the courtyard.
      Player: I want to sneak past them.
      GM: It’s well-lit and there’s not a ton of cover, so they’ll spot you eventually, but you’ll be pretty close to the other side when they do.
      Player: Can I roll Stealth to avoid getting noticed?
      GM: Sure, but if you fail then the guards will surprise you and get to ambush you immediately.
      Player: No way, I’d rather just get noticed when I’m near the other side.

          • I’ve wondered what it was about FATE that you didn’t like, Angry. That makes sense. I think you also mentioned how player goals and character goals don’t align (again in relation to compels.)

            And I agree. It is an *interesting* approach but it wouldn’t be my cuppa hemlock.

          • While its fair to compare it to Fate, it doesn’t feel the same to me. I’ve played Fate, but right now I’m alternately running and playing in a dungeon crawler campaign with pretty basic D&D rules and no meta-currency shenanigans.

            One of the old mantras I’ve heard a lot about running old dungeon crawls that have a lot of deaths is to give the players lots of chances to avoid rolling the dice (and another is that combat is a soft fail-state). The theory seems neat, but I’ve literally never before managed to pull it off until now. All of the normal procedures suggest asking the players to roll dice for ‘risky actions’, which is just about everything in a dungeon crawl.

            With this shift in thinking, we’ve had some great interactions. The players come up with ideas for things to try, the GM gives them a clear cost for each of those ideas, and then the players choose which one they’re willing to go through with or gamble for a better result.

    • It wasn’t a dogwhistle, it was an explicit reference to a very positive and often unappreciated aspect of the Dark Souls games and the Dark Souls community. I’m not afraid to publicly like and support what I like and support. No dogwhistling for me.

  8. I feel that “failure” (however you define it, but I’d say player perceived failure) is only good to make success taste sweeter. That is all.

    Truly difficult/challenging scenarios (if you want them) are not binary failure/success end states, but more the “pick your poison” type in my own limited experience/opinion.

  9. Feeding into this problem is the whole structure and form of published adventures.

    (Sidenote: the Alexandrian points out – I can’t find where – that published *narrative* adventures often teach DMs the wrong ways to prep for sessions, and are bad models for DMs writing their own adventures … because published adventures look more like, and have a structure more like, fantasy novels rejected from actual publication for profit. For a start, they tend to force DMs to thumb through pages and pages of turgid prose and backstory that doesn’t often help to run the actual game. Try subjecting yourself to one of the eye-bleeders that Paizo and/or WOTC called an “Adventure Path” sometime, you’ll quickly get the point.)

    Implicit in most published adventures is that you follow a given pathway, and if the party wipes halfway through, welp, Critical Mission Failure and that’s about it. Their structure – as if you are characters participating in a narrative that exists outside the players’ actions – in part pushes DMs away from allowing players to fail … because nobody really, at a visceral level, wants to stop reading a good story halfway through. Not once they’ve identified with the characters and have some investment in the story. The very structure of most published narrative adventures shoves continuity to the background and focuses on *where the party is going next*. And therefore has to, implicitly, try to force DMs to not permit failure, because the adventure’s failure state is: you don’t get to the end of the book, you didn’t reach the final encounter.

    Compare that to the sandbox or the hexcrawl. Much, much less force delivered by an outside narrative. Yes, there may be overarching factions advancing their schemes, but the party is not put, by force of the product, in the forefront of advancing or foiling those schemes. Scenario – not narrative. Don’t prep plots, prep contingencies. These get closer to the mechanic at the heart of a RPG: set the scene, ask for reactions, adjudicate the results, repeat.

    • This really resonates with me, Marcus Aurelius. I’m running Wild Beyond the Witchlight currently and there’s very little in the module that points to, supports, or allows consequences based on the PC actions. If the PCs are captured, they can escape with no changes. There’s no time pressure. There’s no plan that the villains are pursuing in the absence of the PCs. It all eventually culminates in a Mass Effect 3 style Red, Blue, Green ending (pages 204 – 207) where several of the “endings” don’t actual end the adventure and the DM is left wondering what to do next. I’m trying to run the adventure with a “don’t prep plots” mindset and thinking about how the world evolves in reaction to the PC’s actions but doing so takes a lot of extra effort.

      For what it is worth, I also remember the post about published adventures being bad examples and I also can’t find it. I’m not sure if it was The Alexandrian or another site I follow. But it was great to see The Alexandrian brought up here– “You got your chocolate in my peanut butter!”

  10. A video game fail state is a do over. It’s revert to check point whether to a save, some point in the stage, the beginning of the stage, or the beginning of the game. This is not applicable to a pen an paper RPG because there is no start over at the room’s entrance, or at the entrance of the dungeon, or back in town to do over. Screw up and have to retreat to try again and the world remembers that it’s not your first try.

    Failure in an RPG is the failure to achieve an objective and enduring the consequences of that outcome. It is not a defined state, but a continuum of less desirable outcomes. It’s can the heroes save the day, and if so at what cost versus how will the heroes save the day, where that they will is a forgone conclusion.

    Where a do see a case for a “fail state” is in a Morrowind versus Skyrim kind of way. The former will let quest givers die and will prompt you that you are living in a doomed world, where is the latter will not allow a quest line to become incompletable. Morrowind’s gameplay has failure and reset state and its story has a fail state, while Skyrim only has a gameplay failure state and no story failed state.

    The need for a failure state in a pen an paper RPG is the need for a state other than success. If the chosen ones to save the world sit around the inn all day instead of going on adventure, the world will in fact end, and they could die while on adventure leading to the world in fact ending.

  11. I think the core misconception about “failure states” in TTRPGs is that people assume a story is okay to end at “And then Darth Vader cuts off Luke Skywalker’s arm, before telling him he’s his father. In despair Luke throws himself into the chasm below. THE END!”

    There are many “failure states” in written media, but they are followed by Act 3. Vader Kills Obi-Wan, that’s not a failure state of the movie, it’s a sad moment, but then the story moves on to Luke and Han working together to destroy the Death Star.

    I would argue that any adventure which is designed in such a way that there can’t be “good” a conclusion to a story, is a poorly written adventure. Heck Rogue One ended up with a TPK, and yet the story continues, they still managed their mission. The players rolled new characters and played as the heroes of the Original Trilogy. (Then 30 years later the DMs little sister wanted to run the same campaign, but pretended it was set in the future of that old campaign, and all her players are silly edge lords)

    Even in video games, as mentioned in the article, there aren’t really any other failure state than “stop playing.” (Heck Fromsoft games even incorporates that idea into the core of it’s plot!)

  12. When I first started DMing I thought continuity as you describe was all that was needed. So I made a world filled with concurrent problems and complexity. I allowed the players to go what ever direction they wished, and they hopped from one issue to the next without solving any of them, and grew frustrated at the “lack of coherent theme”. They found genies over here, and fiends over there, and it was all connected in ways they didn’t understand because they dipped so shallow into each. After the fact I realized that the adventure had to be packaged more succinctly, without “diverging”. This basically meant that they didn’t want to decide which adventure to work on. There could only be one issue at a time in the world, and that had to consume the whole campaign. Other players love the complexity and agency of multiple issues. Just wanted to say you are acknowledging a fundamental of TTRPG, continuity, but there is more to figure out after that!

    • There’s a middle ground where you start introducing reasons for them want to go places and being more free with information as to how the world interacts with itself.

      Why are there envoys from Kingdom A here? because there are envoys from Kingdom A everywhere! Why are they getting in your way? because you caused trouble for their uncle back in the previous region and news of vendettas travel fast.

      Of course, there are still going to be groups that just look at you expectantly whenever there’s any ambiguity about what the next plot point may be. I’ve noticed it’s especially prevalent in groups that are used to playing published Adventure Paths

  13. Makes sense to me. Failure is a misnomer. It’s just what happened as a set of choices. Real example from a campaign I’m DMing: choosing brute force over diplomatic means to quell a rebellion seemed like a good idea at the time but in an IRL year later when the demon king invaded who the party over that time found out the rebels had warned the king about and the king thought they were trying to scare people to over throw him… well they wished the rebel faction’s resources were available to assist because the king didn’t support the party when they brought it up. Not a “failure” state. Just how it went down.

  14. I’m having trouble understanding how this isn’t just semantic, in the end. You can think of results as “just what happened,” or you can think of them as successes when they benefit the PCs and failures when they hinder the PCs… but I’m not seeing how this affects the actions nor approach of the GM in either case.

    Maybe I’m just not getting it.

    • It’s really a matter of how you, the adventure-designing GM, think about this stuff in the design stage. How much you tie yourself in knots trying to “include” this stuff rather than just trusting the game to provide consequences and that even if there are no consequences that anyone would called “failure,” you’re getting all the benefits you need without purposely trying to create a “failure state” or break an RPG system to include some kind of “mechanical failure alternative to death”

  15. I think it’s useful to distinguish consequences, losing, and failure.

    Consequences are game states that result from things that happen in the game. In soccer, the consequence for an egregious foul is the offending player is sent off and their team plays the rest of the game down a player. Consequences may or may not be dictated by the rules of the game.

    Losing is a final game state defined by the rules of the game. It requires rules that define when the game is over, and rules that pass judgement on the players. In chess, the game ends when one of the kings is captured. The player who’s king is captured loses. Once you lose you can’t keep playing. The game is over.

    Failure is a feeling, a subjective experience of the people playing the game. It requires players to have a desired outcome that isn’t guaranteed, and the ability to act to bring about the desired outcome. If you don’t care about the outcome, you can’t fail. If you have no control over the outcome, you can’t fail.

    The primary aesthetics you are aiming for in your game can help determine which of these elements you should spend time worrying about.

    Fantasy needs plausible consequences to maintain the fantasy. A game of pure fantasy doesn’t need an ending, failure, or winners and losers, just consequences that make sense within the fantasy world.

    Narrative needs endings. Stories end, and good stories end at the right time. Losing can interfere with a narrative if it is possible to lose before the story is resolved. A purely narrative game doesn’t need winners and losers, or failure.

    Challenge needs failure. A game feels challenging when it’s hard to get what you want, but you can get it if you play well enough. Designing for challenge is hard because you don’t have direct control of how players feel. Having objective, clear win/loss conditions is a useful tool for designing challenging adventures because you can virtually guarantee challenge-seekers will play to win. It lets you dictate an outcome for those players to be invested in. Then you just have to design an adventure that convinces the players that it’s their own damn fault when they lose, or that they definitely would have lost if they weren’t such amazingly skillful D&D players. You can have a challenging game without an ending or winners and losers but it is harder to design. Video games usually pull it off with execution challenges, but those don’t really exist in TTRPGs.

  16. I think this old video by TotalBiscuit does a great job of explaining what a failure state really means to a game: https://www.youtube.com/watch?v=6bvX4hzqcqc . Basically, he argues that all games absolutely need failure states, but failure states can be a lot more nuanced than “You die”. Your herbalist example is precisely a failure state; the players want to save her, and if they don’t, they have failed. That’s more than good enough to make a game.

  17. I like the comparison to open ended games like DF or Rimworld. There is no specific win state, it’s more like choose what’s important to you and try to accomplish that. If you don’t succeed, both you, and the world, will be different, and you can try again, or find something else to do. Toss a few “this will definitely change the world for the worse” states in there, and a lot of “this will help you get strong enough to prevent the other bad things” states in there, and let the players figure it out. I’m just here to adjudicate.

    That being said, I’m not above using “as the last as your strength leaves your body, your consciousness fades to black.” followed by some new experience the PC finds themselves in when they awaken. If players want to keep their PC’s alive I can do so without ruining fail states for them.

    In one campaign, PC death was tied to the BBEG plan and not only would the PC survive after death, but also be restored with a significant buff, BUT the BBEG’s plan would also advance. A mixed blessing for the PC, but a fate worse than death for everyone else (except the bad guy of course). You never really have to kill a PC because they ran out of hit points, but given the option, a lot of players choose to die and start over when the alternative is even worse.

    Lots of ways to play with fail states, even in TTRPG’s, and I think a lot of it is how much RP and how much G each table prefers. Good games let you restart quickly because sitting at a table while all your friends continue playing is a little sucky. Good RP lets consequences take the forefront, and an invested player will remain interested anyways because the RP is fun on it’s own.

Leave a F$&%ing Comment (Limit: 2,500 Characters)

This site uses Akismet to reduce spam. Learn how your comment data is processed.