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

[FrogReaver]

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.

I’d suggest it’s a bigger problem. I mean at some point you get a feel for your DMs style and so in a sense learn this aspect for your given table. But making tactical/strategic decisions before you’ve ’found the pattern’ makes this nearly impossible. In practice this ‘unknown’ means players are much more conservative when trying things that are likely to rely on skill checks than they might otherwise be.

That said it seems like determining the appropriate level of zoom/granularity for a an action declaration in a narrativist game yields similar results.

 

[The Firebird]

You may or may not know the specifics. You know what’s desired… but the surrounding details aren’t specified until the roll occurs and then the GM establishes the fiction.

So in @pemerton ‘s example, if the dice had come up a success, the guard still may have made an appearance… it just would not have been in a negative way.

The insistence that “the guard would not show up on a success” is mistaken. It’s asking for and expecting a specific answer when there isn’t one.

I think the reason this isn't landing is because "no one knows if the guard would show up on success or failure" isn't any better, from my perspective, than "the guard will show up on failure and not on success". The guard's appearance is still being adjudicated as a result of an independent action roll.

No… YOU don’t know if the guard would have shown up in @pemerton’s game if the dice had indicated a success. No one knows that because it didn’t happen.

Does 'no one' include the GM?

I also don’t know how we can say for certain that the guard would show up either way in a trad approach… it really depends on the nature of what’s being attempted and how it goes.

Yes, we do. The guard is there--that is fixed.

 

[AlViking]

I pondered this a bit, and I think I have an answer for you. The critical notion is the commonality you express: "in each case the GM narrates something not because they considered the fiction and its logic impressed itself upon them but rather because something happened at the game table that prompted them to narrate the presence of a NPC/creature."

I think the crucial difference is the degree of freedom the GM has in their narration. In the farrier case they are completely free - they could just as well have narrated that there was no farrier without anyone calling foul. There is an outside compell, and that is when the decission need to be made, but a neutral referee with is the ideal for a living world would not feel any pressure toward the content of the answer.

In the fail/succeed roll case the GM is not free to narrate. They are bound by the result. On a fail they are not only forced to make an answer, they need to make sure that answer is interesting, and negative for the characters.

The wandering monster is interesting because that come in both flavors. You have games with hard player-facing rules for when wandering monster checks needs to be made, and modules often come with predefined tables. This situation is even more extreme than the fail succeed roll. Here the GM isn't only bound to when to make decission, they are fully bound to what content to introduce. This is what gives random monsters a really bad reputation for breaking immersion.

But wait a second! Wandering monsters and random tables is a hallmark of living world sandbox play! How can this possibly be if wandering monsters are even worse than forced fail forward? That is that when used in such play, the criterion of player side rules for their use is not present. Rather than being a game mechanics, they are a tool for the GM to help them stay neutral. These are usually made by the GM and tailored to the GM vision of how population dense the location is, and only populated by things they expect might make sense. They are only bound by timings they create, and if they find the fiction is critically different than what they expected when they made the tables, they are fully free to skip rols, alter frequency of rolls, or alter table results if that make more sense in the new situation. So they are actually not bound to time or content. The use of the structure and the random roll is to further eliminate the bias that might come in if they would have introduced things only on "feel", hence producing an even more authentic feeling world than they would have managed with full freedom.

Wandering monsters are a thing because of the old standby rule "When there is uncertainty, roll for it." If the characters are going through a tunnel where they can't avoid the giant spiders, there's no need for a roll. But going through a forest when the DM doesn't have a detailed enough map to know whether or not the specific area they're traversing has giant spiders? Roll for it! Even more important if you don't have largely stationary predator like the spiders, many predators can have ranges of dozens of square miles.

The creatures encountered based on a wandering monster table are not added to the fiction of the world as much as the roll indicates that you just happen to cross paths with the creature.

 

[Lanefan]

The conversation that is play. Like....that's literally how the game works. It is NOT simply "success on a search roll"--whatever that might be in a given system.

I gave @Enrahim a breakdown of how it works in Dungeon World.

Also? In nearly every possible game of this kind, this WOULDN'T be the only information. I'm only specifying that because that's what YOU told ME to do. I wouldn't have gone with that! Such a paucity of information is, frankly, not very conducive to good or interesting play.

"Such a paucity of information" just means the players, in character, have to do a little work and get their fictional hands dirty searching the house. And yes, resolving this might take some time, both in the fiction and at the table.

Pretty boring if you already know exactly where it is within the house and all you have to do is go in and get it.

 

[AlViking]

I never considered otherwise. I disagree with @Hussar's assertion that "failure will almost always be catastrophic." My objection was to multiple checks for a series of interconnected micro-actions serving a single task.

To use an actual example:
In campaign 2 of Critical Role, Marisha - as Bo - wanted to run up a vertical surface, leap across a gap, then follow up with an attack. Matt called for an athletics check (which Marisha negotiated to an acrobatics, because Dex-based character) to ascend, followed by another to jump, and then, of course, the attack. Personally, I feel the run and jump could have been folded into a single check. This was made all the more egregious by the fact that Bo was at a high enough level that Unarmoured Movement gave her the ability to run along vertical surfaces and liquids without falling, and no other threat had been established to warrant such a check.
This was not the only time Matt called for a series of checks that would inevitable ensure Marisha would fail.

There were times when Matt called for far more rolls than I ever would have, it's a balancing act and I don't think he always followed best practices. As you point out, chain together enough checks and you're bound to fail eventually.

However I will add a caveat to that. There are times for me that a successful check can determine whether you get a bonus. The character succeeds on a perception check before climbing a cliff, so they figure out the best path up. They get advantage on their climb check because they avoid a section of rock that looks unstable. Failure on the perception check doesn't mean they can't climb the cliff they just don't get advantage. Quite frequently I use checks to make subsequent checks easier and failure just means nothing changes. In cases failure may mean you need to figure out an alternative, sometimes it slows you down or cause a bit of damage but you can keep going. The number of checks to call for and the result of success and failure is always going to be both a balancing act on part of the GM and a preference.

Tying this back to Matt and Bo, if someone wants to do something truly difficult even for an experienced adventurer, then I do think there should be a pretty good chance of failure and that failure being based on a series of rolls can be part of an enjoyable game (and better than a single roll). I don't remember the the specific scene you're referencing but if someone wants to try something really difficult I'll let them know the odds of success may be low.

I guess my only real point, if there is one, is that I use checks in a variety of ways. Sometimes it's simple success or failure, sometimes it's make a check and you get a bonus on the subsequent action, sometimes it's a minor setback, sometimes it's not a question of whether or not something is possible it's just how well or how quickly you do something. But in all cases the result will be tied directly to what you are doing.

 

[Lanefan]

Like I said game rules don't prevent you using loaded dice for your rules. And how is anyone going to know to ban them afterwards? Are you going to boast of it.

I once had someone try using loaded dice in my game as a joke; it was blatantly obvious after a couple of rolls that those dice were sketchy, and the session then ground to a halt for a while because of course we all had to try them out.

But any reasonable person would consider it taking the piss and you would almost certainly be kicked from the table for using loaded dice.

Indeed.

Pulp Fiction has a continual plot device of a briefcase that when it's opened sheds golden light on whoever opened it. It drives the entire story and the characters all want it. But what it actually is is meaningless (and it literally is a briefcase with a lightbulb in it, but you never see inside).

Thanks for the explanation.

And this is a part of why most people don't play old school games. If we wanted to play a game of accountancy and inventory management most people find computer games do it better. Wanting to play burglars stripping the mansion down to the studs description by tedious description is a niche kink. But it's one encouraged by old school games (with XP for GP rules). If you look at the new darling Daggerheart it is so uninterested in money that it measures it in handfuls, bags, and a chest.

I've been meaning to have a look at Daggerheart. Seeing this, however, just made my interest plummet, and I hope it's not reflective of a broader-based inventory-doesn't-matter design philosophy.

 

[Faolyn]

Ah. I'd been assuming there was a longer time window between door attempt and deaths, in part due to my reading the original example as the house only just having caught fire.

In a trad game, these NPCs aren't going to have a lot of hit points[1], so between smoke (poison damage) and the flames (fire damage), they're going to die very quickly.

In a more narrative game, there (IME) usually aren't rounds that represent seconds or minutes; instead, actions take as long as they would logically take. so any attempts you make also don't take a set amount of time. In the real life, picking a lock can take anywhere from, say, 30 seconds to twenty minutes, or longer, depending on your skill[2], tools, and the type of lock. Thus your single lockpicking attempt (whether a roll is needed or not) would actually be five minutes long.

--

[1] If they did, it would suggest a higher level, and thus they would be less likely to need rescuing because they'd also have higher skills.

[2] In these games, "skill" doesn't necessarily mean a number of special ability.

 

[Lanefan]

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.

The bolded is where the mistake is made, if someone's doing it like that.

Ideally, in a granular resolution system, the degree of consequence to a failure would ideally be more or less in keeping with the degree of importance of the sub-task.

And sure, sometimes the most logical consequence of even a minor failure is something catastrophic, but that should be rare. Far more often, a failure here means you still have opportunity there to accomplish much the same end result. In the example of the burning house, failure to get in through the front door should still leave me options of trying the windows, the second story, the back door, etc. The only thing guaranteed not to work is trying the front door again using the same means I already failed at.

This is where "nothing happens" can be very useful - it's a minor consequence to a minor failure. All you lose is a bit of time. But that's not dramatic enough for some, hence the tendency toward more catastrophic outcomes.

That said, it's a matter of preference as to how often PCs should fail on average, and a fairly simple matter of adjusting the difficulty of checks to achieve the general failure frequency you want.

 

[Lanefan]

Hear, Hear!

Not for the first time, I find myself wishing the "like" options included a facepalm emoji...

 

[Faolyn]

I am so not going to ask my parents about how well they performed.

Then don't be ridiculous, because we're talking about the game, not your parents.