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

[pemerton]

I'd argue that the largest determinator of success is therefore not the skills of the characters or even the fictional situation (as above, all three approaches are congruent with the fiction) but how many rolls the GM decides are needed before success. It risks a kind of "soft" railroading, whereby failure can be enforced by simply requiring more rolls - I've certainly experienced this and done it, albeit inadvertently.

Yes, this is one reason for adopting closed-scene resolution.

Or for adopting the AW approach where "nothing happens" is not an outcome of a roll.

 

[AlViking]

Unfortunately, in practice, what this means is that the PC's will nearly always fail. Because the DM simply piles on check after check until the PC's fail, and that failure will almost always be catastrophic.

You call it granular, I call it punishing for no reason. Having seen this in play so many times, I have zero interest in this kind of play anymore. The fact that most DM's simply cannot understand the math behind the checks means that it rarely, if ever, actually works.

As GM I want my players to succeed most of the time. They won't succeed 100% of the time at everything they try. As GM I want to challenge my player which sometimes means what they had planned doesn't work and they have to scramble to figure out having to do something else - which to me is half the fun of the game.

 

[TwoSix]

Yes, this is one reason for adopting closed-scene resolution.

Or for adopting the AW approach where "nothing happens" is not an outcome of a roll.

I have “liked” this post as a team-building exercise.

 

[AlViking]

All of this arguing comes down to 2 fairly different approaches. I prefer a more simulationist approach as explained in https://samsorensen.blot.im/new-simulationism*, others want a more narrativist approach or to add in fail forward.

My approach?

1. The character goes to the door, attempts to open the lock
2. As DM I think about the scenario and whether or not anyone will notice the lock being picked. If someone would hear I see no reason success or failure of the attempt would make a difference.
3. The player rolls poorly and fails.

That's it. Nothing else happens, the lock is not unlocked, there is no step 4. Other things may happen because a round has passed but as DM I am not going to introduce anything else to the scenario because of the failure. If I had determined in step 2 that someone or something inside would hear the attempt on success or failure I'll decide what happens based on them hearing the attempt.

Fail forward

1. The character goes to the door, attempts to open the lock
2. The player rolls poorly and fails.
3. The GM adds some complication that they feel is plausible for the current scenario. It may or may not be helpful, but something will happen. Examples have included a guard showing up, rolling on a wandering monster table and a grick appears, the door is opened anyway and there's a cook inside who cries for help.

That's pretty much it in an abbreviated form and of course I've left a lot out on the narrative side of things.

There are various reasons people do this or like different approaches. Narrative games are designed to work very differently from D&D and have a very different approach although you can of course take ideas from them. It all just comes down to preference and neither approach is better or worse.

I am not making any kind of judgement on how other games work, I just know that the narrative approach does not work for me based on games I've tried and streams I've watched. If it works for you, fantastic. Personally I'm thankful that there are so many options and different ways to enjoy our free time.

The horse is dead, it was beaten to death long ago. Long live the horse.

*I don't agree 100% but it's close. I also don't see anything particularly new about it, but that's a different issue.

 

[Maxperson]

What? Says who?

I don’t think there’s only one way to handle that in D&D. There’s certainly not only one way to handle it in RPGs.

Again, this is not true for all D&D. Certainly there are methods and means used to create dynamic environments where NPCs will move about or can be found in different locations.

Asking @pemerton what happened if the player succeeded on their roll to sing is a mistake. The player didn’t succeed… so it didn’t happen. There’s no way to say what would have happened except to guess.

It may be or it may bot be. It depends on how the GM preps or handles this type of situation. Even if it is fixed, the failed roll may just mean that she’s aware of the break in, and a successful roll means she is not.

Do you truly not understand that when someone is talking about the differences between two styles of play that he isn't saying there's one way to handle things or that it's true for all of D&D?

Oh I don’t think that’s in doubt at all. Though I think there’s likely plenty of variance from table to table, I would say that generally speaking a trad approach is going to involve more rolls than other approaches.

It’s not necessary to do it that way, but I expect most trad GMs lean that way.

Or that it's necessary to do it a specific way.

This is what I mean when I say that you read to respond and not to understand. Only the bolded paragraph was in any way a proper response to what I said. The rest of it was completely inapplicable because it didn't deal with what was actually said.

 

[Maxperson]

I didn't. @AlViking did. In my reply to AlViking, I pointed out that necessary conditions aren't causes, whereas the causes that @hawkeyfan and I had pointed to are INUS conditions.

Not all of them were INUS conditions, though. Take the example he gave of my being tackled for punching someone in the face. The "punch and witness" will not always result in my being tackled, so the "punch and witness" does not qualify as a sufficient condition.

 

[Maxperson]

I like it (pardon the pun).

What it saves are hundreds of posts sayng nothing but "I agree" or "Hear, hear!".

I agree.

 

[Pedantic]

I'd argue that the largest determinator of success is therefore not the skills of the characters or even the fictional situation (as above, all three approaches are congruent with the fiction) but how many rolls the GM decides are needed before success. It risks a kind of "soft" railroading, whereby failure can be enforced by simply requiring more rolls - I've certainly experienced this and done it, albeit inadvertently.

It's useful therefore to come up with a heuristic for how many tests are needed that's beyond the fictional situation (as demonstrated above, the same fiction can reasonably deliver multiple possible values)

Why is this necessarily a GM facing problem? You could space this at the system level with a more socially designed stealth system, tested in scenarios that match the intended level of competence. Off the top of my head, reworking Stealth as a fixed value that's spent down to perform actions seems directionally like a good start.

The impact of iterated probability is obviously very real, but that's as much or more an indictment of the system designer than the situation. That, and it's only a problem if the player is expected to resolve most situations by making action declarations that resolve as rolls with some chance of failure every time.

 

[Maxperson]

@AbdulAlhazred made this point some way upthread:
To me, it seems that the real issue is this:

@AlViking and @Maxperson are affirming some restricted version of the following principle: Counterfactual statements about the real world, and counterfactual statements about the fiction, should tightly correlate with one another.

That is why they insist that what would the GM have narrated, had the roll succeeds must correlate tightly with the way causation is working in the fiction. This then leads to an idea that the purpose of the dice roll and the associated decision-making about resolution is to directly model the causal process that is taking place in the fiction.

The RPG that I know that comes closest to an unrestricted version of the above principle is RuneQuest. There are two reasons that I say that @AlViking and @Maxperson are affirming a restricted version of the principle:

(1) There are big chunks of D&D's mechanics, including its surprise mechanics (as per my post not too far upthread), its stop-motion combat resolution (as per a post of mine further upthread), and other stuff too (eg at least some aspects of saving throws and hit points) that don't conform to the principle. It's that failure of conformity to the principle that explains why all the classic simulationist FRPGs that were designed in reaction to D&D (RQ, RM, etc) don't use these resolution mechanics, or at least try to minimise them to a great degree.

(2) They don't adhere to the principle in cases like the farrier you've been discussing with @Maxperson, which @AlViking also distinguished from the guard and cook cases:
The details wouldn't be added but for the player's question, and so the tight correlation principle is not adhered to.

As I also posted some way upthread, these restrictions on the tight correlation principle don't seem readily explicable except as a common way that D&D has done things. That's fine, but doesn't provide the foundation for some sort of principled attack on other approaches (eg for violating the tight correlation principle).

First, the combat rules can't really be used as any sort of indication for the use or lack of use of the principle. The combat rules use the stop motion method because if you tried to mimic reality, combat would be a God awful boring mess that takes 10-20 sessions to complete depending on the size of the combat. It's a necessary evil, not proof that some principle should not be or is not used.

Second, surprise in D&D has changed a lot over the years, so using 40somthing year old surprise rules to prove something in games being run now is sort of fail on its face. I haven't played in or run an AD&D game in 25 years, and I haven't played in or run Gygax's AD&D version in 36 years.

 

[Maxperson]

Unfortunately, in practice, what this means is that the PC's will nearly always fail. Because the DM simply piles on check after check until the PC's fail, and that failure will almost always be catastrophic.

You call it granular, I call it punishing for no reason. Having seen this in play so many times, I have zero interest in this kind of play anymore. The fact that most DM's simply cannot understand the math behind the checks means that it rarely, if ever, actually works.

It really depends on how the players play the game. If they are trying to be quiet when they enter, a single stealth check will cover the lockpicking, opening the door and entering the kitchen. Those are only all separate rolls if the players are reckless and just do it all without trying to be quiet. If it's "punishment," it's the players punishing themselves.

Assuming that they are trying to be quiet, the only other roll would be a perception check at disadvantage for trying to navigate a kitchen in the dark. Even with darkvision visual perception checks are at disadvantage. Stealth won't cover that situation, so it requires a different roll to get through without bumping into something and making noise.