D&D General [rant]The conservatism of D&D fans is exhausting.

[AlViking]

Because whoever owns the door isn't going out of their way to make life easy for the PCs.


Yes, that's true. That's why fail forward exists--to help minimize poor planning and design.

I'd rather instruct people on how to do and discuss better design instead of just throwing on a duct tape patch.

You missed the antecedent: the PC is chasing the assassin and wants to catch them.

No I didn't. They failed to catch the assassin but didn't somehow manage to catch them anyway. Instead they have completely different options to track him down. They failed to achieve their goal in this instance, now they have different alternatives to achieve it.

Have you mapped out the existence, location, and routine of every being in the location? Have you determined the squeakiness of every single hinge and floorboard? Probably not. So what does it actually matter if the hinge doesn't squeak on a success?

The point of the roll is "can the PCs get in?" Yes, they can--but there's a complication. Or no they can't, but they learn of another way.

Of course I haven't mapped everything. Meanwhile there will likely be other ways of getting into the house, including ones I hadn't thought of ahead of time. There's nothing wrong with "You try to pick the lock and nothing happens" because they can still achieve their goal, just not in the way they had planned. This happens pretty regularly in my games and is part of what the players enjoy.

So tell me, do you list all of these things ahead of time? If the players want to get through the door, they can: pick the lock, break the door down, find the hidden lever, knock

If so, and the players fail on all of these rolls, or simply don't think to do one, the game still comes to a screeching halt.

The game never comes to a screeching halt simply because they failed to achieve a specific goal. There are numerous opportunities and goals they can pursue instead. It's one of the main reasons I don't run linear campaigns.

 

[The Firebird]

What if...

You couldn't tell they were making it up?

"Is fudging ok if my players can't tell", and more at 11.

What if.... someone played in your game, and said afterwards that you were a poor GM because everything was preprogrammed or rolled randomly and therefore it felt like their choices didn't matter. Would you consider that to be a valid complaint?

I'd suggest they find a different table; our styles aren't compatible.

Also, why would you assume your choices didn't matter? If the GM is making it all up--which, by the way, every GM does when they write the adventure in the first place--then your choices are the most important thing, because that's what the GM is relying on.

I disagree, for reasons that I've described at length by this point.

 

[The Firebird]

Indeed. We can take that further, to make the point more clear.

The kitchen, which was explicitly placed in the scenario, requires a cook. And cooking fuel. And food. But the map probably forgot to include cordwood, or what exactly is in the pantry.

So, we have setting elements we know should exist, but aren't specified before play begins.

When the situation warrants, a GM is within rights to specify them as they deem appropriate, even without a chart, right?

Yes, certainly.

 

[Umbran]

Except succeed with complication is more accurate to what's actually occurred.

Oh, no! Someone used an older name for a technique! The HORROR!

 

[Maxperson]

Sure, but why? It's a dice roll that tells us the cook is there, which is something that you said you're fine with. Prefer, actually.

D&D General Post in thread '[rant]The conservatism of D&D fans is exhausting.'

Wednesday at 3:00 PM

One response then has been proposed by @Maxperson, which is that the table is used only because the fiction contains the right conditions. That seems to mirror @pemerton's careful and repeated explanation that the consequences of fail forward (such as a chef) can and ought to be prefigured in the fiction.

Yes. What I will do is something like the following.

The players decide to break into the wizard Alchaeus's tower. Alchaeus has two servants who live with him. His apprentice who also doubles as his personal servant, and his chef. While the players are discussing...

• Maxperson

• 

 

[Enrahim]

No, I believe the design task is to set a series of rules that can as completely as possible adjudicate any player proposal, using knowable, player facing rules. That most systems fail at this is not an indictment of the goal. Obviously, some abstraction is necessary, but you can get pretty far with a solid range of modifiers and some generic systems, like object hit points and hardness.

I'm not concerned with the precision of those rules as a simulation (though it is obviously helpful and easier to use them if they align with player's intuitions, so that can be a useful design guideline) so much as their predictability and universal application.

The GM as designer of last resort isn't a good thing, it's a patch to keep the game functioning. The less it occurs the better, and the more a designs celebrates it, the more I can assume it's not going to be a complete game, and the less player agency I can expect. My point was that if it does become necessary, the more complete the system, the less adhoc design work is necessary and the more likely a GM's ruling won't be disruptive to the player experience, because they'll have more existing rules structure to model it on.

Decign and principles are hard. If this principle as formulated here the game "Flip a coin whenever you are unsure" would be more complete than any of the mainstream RPGs. I actually think all post TSR D&D variants and their closest relatives have had minimising GM rulings as a virtue. However this need to be balanced against other concerns. For instance detailed rules for certain actions make these actions feel better in play, but it increases the need for GM rulings for the actions not covered compared to a bland generic system. So just by introducing action based rules you have made a sacrifice with regard to your completeness ideal.

Your proposal appear to be to adding on even more action rules in order to reduce the number of cases where GM rulings are needed/problematic. And while this indeed can work, the more you do it the more you sacrifice a different design principle: Simplicity and ease of play. And you seem to acknowledge that you won't be able to get to perfection anyway..

Modern trad game design is a balancing act between ease of use and granularity/richness. The way it try to counter GM rulings is not to increase granularity, but rather to make the grains fit as tightly together as possible, covering as much of the posibility space as possible given the desired level on the ease/richness scale.

so if you throw ease of use out of the window you easily can go to quite a bit higher granularity than today's mainstream, and with that increased granularity you might be in a position to provide better coverage than the less granular. But even this has an important limitation: if you become too granular the risk of significant overlap also increases. I guess you would agree that an "overcomplete" system where the GM has to chose between 3 different described actions that might fit the situation is as bad as the situation where the situation is not fitting any of the described actions?

I'm sorry, I'm also not totally sure what point you're referring to here, though that is an interesting question. I think the ability to support varying and updatable victory conditions is the definitional element of the form. The players agreeing to bind themselves to specific victory conditions ahead of time doesn't really impact that. Compare to a board game; outside of the variance of scenario play, the system of interaction can't be repurposed to broadly support a bunch of goals, and gameplay can almost never continue from the same board state after the evaluation of victory.

Not a point I'd considered before, but your example does make it clear the potential of the rules to be applicable to broad goals and continued play is more significant than the specific goals themselves, or whether play continues.

Have you heard about flux? Is that a TTRPG? Or if the range of victory conditions for that is not broad enough - have you heard of the card game MAO? (That one is an interesting case study, as that is also somewhat conforming to my attempt at defining a core aspect of TTRPG)

I think it is hard to get away from a player controlling a single fictional entity as a requirement for something to be considered a TTRPG. However with this in place I think our attempts at formulating "the rest" might be so that my requirement is slightly stricter than your. I think an unlimited set of possible actions will imply that the form can support updatable win conditions. I can however envision limited action games that do support updatable victory conditions. If such a hypotetical game seem to satisfy your thirst for TTRPGing, I have no objections to that. For me however I think I would feel like something essential is missing But I guess I could accept that the result would be distinct from what I would naturally talk about as a "board game". So you seem to be pointing to something essential.

 

[Maxperson]

People are responding as if the people who are taking exception to the fail forward example are trying to sell them on fail forward as technique, rather than show why we view the example as reductive analysis of other people's play. The cook example is representative of poor use of fail forward as a technique because the example does not include any of the telegraphing that goes along with it and the consequence does not follow from the established fiction in any meaningful way.

I'm not sure that's intentionally poor instead of just the facts of life when presenting a scenario online. Even @pemerton who posts the most detailed examples I know of on this site still isn't able to capture everything that led up to why he decided the way he did or why the players made the decisions they did. Real game play includes a huge number of details that a proposed scenario just can't begin to match.

It's something that affects every playstyle example that I can think of.

 

[zakael19]

Here's two examples from play of the sort of "Fail Forward" that I've previously defined (the scene changes on a roll), and that @clearstream cited from Daggerheart's rules text:

- (Stonetop, PBTA) The party is exploring an old barrow. There's sconces on the walls of rusted old iron, that flicker to life with silent eerie green flame when they approach. Marshall the Marshal declares he's studying one closely, trying to see if there's any way to like get ahold of it or perhaps take them back to Stonetop. I suggest he sounds like he's Seeking Insight here, he confirms and rolls. A lovely 4, so a Miss. I say "you can see that the flames are emanating from the blackened bones of a human, many of them - piled at the bottom... also as your companions slowly creep up the stairs here they start to hear the sound of metal scraping on stone, like the links of a massive chain dragging across the floor - and then you all start to hear an echoing rhythmic barking from up top - what do you do?"

- (Songs for the Dusk, FITD) The crew is racing cross-country in their buggy, a fast-mover located on radar coming up shadowing them. They know this has to be the Corundrum Paladin that's been making life hard for them (it was the Fallout from last mission after all). Their goal is to get to a sheltered spot Sol found on the map where they can turn and fight. I tell them that yeah, that's doable but this is risky as heck - they might totally flip the buggy here which will, well, complicate matters. I ask what they do, Ada describes how she's got all her processors focused on steering as Sol navigates, going full on rally mode, barking out instructions (he spends a Stress to Aid and give her +1d for her pool).

Ada rolls 4d6 and gets a 1,2,2,3 (ouch). That's a Miss. I describe the juddering of the wheel in her grip, the clods of mud flying everywhere, and then that sudden sinking sensation as the buggy just starts to tip too far and her gyroscopes can tell they're about to go over...and because this is a game with Resistance mechanics, I now ask if she wants to Resist the outcome (technically Push Herself, but shush). She does, I ask how, she describes fighting the wheel and buggy into a long side skid instead - and her overstressed servos result in a fair bit of Stress spent to do this. They're now sitting there stationary, facing back they way they came but not tipped over, only a moment to spare before that fast-mover is upon them so I ask..."what do you do?"

 

[FrogReaver]

One, picking the lock in a fail forward game is generating the result fully using a mechanic.

It actually doesn't. What it does is have a mechanic that generates a prompt for the GM to describe one of many possible fail forward states that the GM must then choose. The actual choice of fiction isn't dictated by that mechanic at all. But it's not just you. Most narrative proponents seem to elide this fact as well and instead substitute the notion that their narrativist 'mechanics are what generates the particular fiction'. But all those mechanics don't actually tell the GM what precisely that fiction should be. And again, that's perfectly fine and has many pros to doing it that way, but let's be clear about what precisely is going on.

Two, a random table is the most quantum-like state imaginable, because unless it's the most heavily-curated table imaginable, it will contain results that are make far less sense than anything the GM is likely to come up with on the fly.

Ideally the results all make sense on the random table and if one doesn't in a particular scenario then the GM should reroll on his table.

EDIT: I hit send too soon.


So basically a video game with a large but limited tree of options.

Not at all. I said "2) extrapolated directly from established and preauthored facts."

So the GM has enough established and preauthored facts to determine what happens next without a table. Great. We've just added to the established facts of the session, which then can be used in future decisions and then repeat this process hundreds of times. Video games don't do that. Video games have gotten better about giving players more choices, but it's always from a far more limited list than a ttrpg provides. Maybe this changes with gen AI in the future, but we aren't there yet.

So something made up to fit the circumstances is inherently less plausible than something rolled on a table?

The issue here isn't the plausibility. I agree fail forward can provide always plausible fiction. But that's rather beside the point and the issue.

By what makes sense.

Again, the issue is how, not whether it make sense when held up after the fact.

Or--and this is a wacky idea--saying, "Player, you managed to pick the lock but screwed up badly somehow in the process. What happens?

That's fine but in terms of the issues I'm discussing there's no difference in whether the player or the Gm does that.

Get the players to tell you what their player did. If you don't mind them narrating their successes, why can't they narrate their failures?

See above.

 

[AbdulAlhazred]

To the bolded: so in DW a character can never outright fail, or fail backwards, at an attempted task?

Yeah, that's a bit much.

No, because PbtA is not a system which requires Fail Forward! It is perfectly OK for a 6- to simply be "you failed, AND <bad thing X> just happened to you!" I am just saying that in DW/AW (probably most other PbtAs) IF you have a forward failure, then it has to be pretty explicitly "success + bad thing" because PbtA moves are triggered on actions, which are fairly specific things and not "when I set/act on a goal." @pemerton's case in BW is thus different, because BW adjudicates INTENT, so you fail your intent, but your action is carried out successfully. The difference, in the end, is not huge, and in a fictional sense 'non-success' can play out pretty similarly, but they do map a bit differently in terms of mechanics and process.