D&D 5E Consequences of Failure

[Ovinomancer]

The fourth doesn’t tell me what the character is actually doing. Picking a lock with thieves’ tools is a very specific action. I know exactly what that looks like. Attacking someone with a sword, again, pretty clear-cut. The fireball spell involves specific gestures and incantations. But “trying to read someone” is vague and unclear. I don’t know what’s being done by the character in the attempt to read the other person. Are they watching for micro-expressions? Are they listening for changes in speech pattern or cadence? Are they observing body language? There’s a lot of things “try to read” might mean. To me, it’s like saying “I try to unlock the door” instead of “I try to pick the lock with thieves tools,” or “I try to kill the orc” instead of “I Attack the orc with my longsword,” or “I use magic” instead of “I cast fireball.”

I come at it from the other direction. What happens if you fail?

I understand what can happen if you fail to pick a lock with lockpicks.
I understand what can happen if you fail to hit the orc with your longsword.
I understand what can happen if you fail to cast fireball at that area.
I don't understand what can happen if you fail to read him to see if

Thanks for letting me know I wasn't clear! I'll try to answer your question and explain my point differently.

In an open-ended game like D&D, I don't see how it would be possible for a DM to always be able to predict what the PCs will find important to be able to include it in an upfront scene description. Sometimes a PC will have an idea the DM didn't consider (and therefore didn't include all the relevant information for evaluating that plan), or it could be the PCs take the entire scene in a completely unexpected direction (rendering moot the DM's analysis of what information is important). I wouldn't classify either situation as a "failure to properly present the scene" on the part of the DM.

As a trivial example, a PC could deliberately change the topic of conversation in a non-confrontatinal social scene and watch for whether the new topic appears to make the NPC uncomfortable. Since the scene had been framed as non-confrontational, the DM hasn't yet had a reason (or opportunity) to telegraph that the NPC is/isn't concealing their emotional state, so the PC doesn't have anything to unique to this situation to structure their action declaration around. Accordingly, they'll probably go with something generic when describing their approach, like "... by watching their body language for signs of discomfort".

On the one hand, the PC declaring a goal (find out if NPC has knowledge of topic x) and an approach (by changing the topic of the conversation to x and watching their body language for signs of discomfort) seems to be exactly the sort of thing the PC should be able to do in Goal and Approach. On the other hand, the described approach isn't specific to some unique detail telegraphed by the DM (because the DM didn't know that such would be relevant or have an opportunity to include it before the player altered the scene) and would apply to just about any similar situation.

My question is, do you consider permitting an action declaration with a generic Approach to be in keeping with Goal and Approach as you see it? If yes, how is permitting a generic Approach functionally different than "pushing the Insight button"? If no, does that mean that under G&A certain types of actions (e.g. reading the emotional state of an NPC) are only available to the PCs when the DM accurately predicted that such actions would be important to them and included appropriate telegraphs permitting situation-specific Approach declarations?

Again, I don't see what you're positing. It makes no sense to me that a player can suddenly force a situation with an NPC where the DM is both determining that the NPC is lying/concealing knowledge AND can't do anything about it because it's a surprise. This makes me think you have a wrong conception of what my play looks like. If a player swings into a new line of interaction with an NPC, I will know if this NPC is lying, what their lying about, why their lying about it, and be able to react accordingly because I am the DM and already know all of these things. There's no surprise. I'm also struggling to imagine a situation where this NPC's lies are both important to the character interacting with them and where I've generated a scene where I've utterly failed to anticipate where said important lie couldn't come up and be left reaching for dice to figure out what to do.

So, No and No.

 

[iserith]

I like the idol example, and have a follow-up question: what if the player wants more information about Modra, rather than know more about the idol itself? Examining the idol more closely won't inform the PC whether (e.g.) Modra has any particular enemies in the Elvish pantheon. What action can the player declare for their PC at your table to find out such information, and how do you resolve it?

Chapter 7 gives a breakdown of each ability score and what sorts of actions might fall under them if the DM thinks there needs to be a ability check. For an action declaration to recall lore or make a deduction, that could be an Intelligence check to which one of various skill proficiencies may apply.

Therefore, a player simply says that he or she would like to try to have the character recall specific information and on what basis the character might have that information. That is the action declaration - drawing upon X to get Y information.

Do you just give the information to them or withhold it based on whether you think the character would know? Do you decide whether or not to give it to them based on whether they are proficient in Religion? Do you call for an INT check?

First thing I do is consider whether I failed to adequately describe the environment. If a player is asking for more info before acting, that's often why.

Assuming that passes muster, then I examine the action declaration as I explained it above. Is the task somewhere between trivially easy and impossible? Is there a meaningful consequence for failure? If those answers are both "yes," then it's time to roll. If the answer to either or both of those questions is "no," then I just narrate the result as either success or failure with no roll.

 

[pemerton]

Well, yes. If you, as DM, present an NPC where a large part of the challenge of the scene is determining that this NPC is lying, and you wish to gate that knowledge behind a check because you enjoy the play, then you'll find very little use for goal and approach because it actively fights against your goals of play.

This took me back to a post way upthread:

Generally then, if you want meaningful consequences of failure for a knowledge roll, you only make the roll at the point at which a character takes some action based on that knowledge.

In other words, if a frog-thing is hopping towards your character, you don't say (a) 'I'll make a knowledge check' you say (b) 'I'll hurl a flask of flaming oil at it, because I remember these frog-things burn real easy.'

Obviously, there are all kinds of interesting meaningful failures for (b). Frog-things might be resistant to fire. Or explosive. Many options.

The obvious sticking point is that many systems are predicated on the idea that success at (b) doesn't make susceptability to fire true - many people would say that's made true by the GMs notes or the Monster Manual. They will deny the player ever gets the authority over content.

So if you leave authority for content or situational creation with the GM, failed knowledge (or perception or arcana etc) checks deny learning what the GM has already decided, while successful ones allow for passive receipt of information.

But a successful knowledge roll can be used within an rpg system to share content authority at that moment. If that happens, you can end up at (b).

Similarly, although stealth has been discussed in the thread it's worth noting that a stealth roll could be accompanied by this statement 'I hide behind the old boxes stacked in the corner'. And the roll can be used determine what happens next. In this case the roll is determining whether the player is allowed situational authority to add some previously unnoticed or unimportant old boxes into play, or whether the GM gets to deny, complicate or aggravate this situation in some new and unexpected way.

Such play is perfectly normal, functional and fun, albeit quite removed from most D&D play (with, perhaps, the exception of 4e skill challenges, which were themselves an outlier).

So these two posts taken together made me think about some of the variety of roles that and Insight check might play in a RPG. (And other information-acquisition checks, like knowledge, perception etc are broadly parallel in this respect.)

(1) Make a check in order to acquire some information contained in the GM's notes. Normally triggered by a fairly generic form of words - eg I attend to the NPC's voice and manner to get a feel for whether or not she is lying. This sort of thing is as old as D&D itself, though its earliest form is more about finding secret doors than "reading" NPCs. This raises some standard questions about whether the GM makes a secret roll, the duty of the player to act on character rather than player knowledge, etc.

Gygax was clearly conscious of the potential infelicities of gating fictional content behind these dice rolls. Thus, on p 110 of his DMG, he says

it is your right to control the dice at any time and to roll dice for the players. You might wish to do this to keep them from knowing some specific fact. You also might wish to give them an edge in finding a particular clue, e.g. a secret door that leads to a complex of monsters and treasures that will be especially entertaining.​

If the information is a reward, then there can be a logic to gating it behind a check - RPGing traditionally has a degree of a gambling element to it. If the information is part of the "story", then the logic of gating becomes weaker, as the consequence of failure becomes not you miss out on a reward but rather you have a less-good story experience than you might have had.

(2) What can superficially seem similar to (1), but is I think importantly different, is using the check to affect the framing of a situation. In the D&D context, the classic version of this is a surprise check. But Insight can be used for a similar purpose. For instance, the 4e PHB says

if supposed allies spring an attack and you failed your Insight check to notice the attackers' traitorous intentions, you're surprised.​

As the surprise example shows, D&D has a long tradition of conditioning framing of situations on the outcomes of checks. Normally these are checks that the GM would call for (somewhat comparable to saving throws). The consequence of failure is that the player has a more adverse framing of the situation.

(3) There is the use of Insight that @chaochou describes - to establish some detail or piece of information about the NPC. In this case, the consequence of failure is that the NPC doesn't know or think or behave as the player (and his/her PC) hoped.

This way of resolving an Insight check requires the players to be engaged with the fiction, and to have some sense of what they want out of an interaction with a NPC before they declare the actions.

EDIT: Saw this:

I'm not following your argument here at all -- it appears you're arguing for a different approach based on the the idea that the DM may fail to present a scene adequately, and so asking for rolls is a way to protect against DM failure to properly present the scene?

So here's a (4): an Insight check can be used by a player to oblige the GM to establish additional information about a situation.

I wouldn't see thiat as about failing to present a scene adequately. There is always a limit on what the GM can narrate. S/he can't always anticipate what a player may find interesting or engaging about a situation. This use of Insight (and other knowledge-type) checks is one way (not the only, obviously) of responding to that possibility.

The consequence of failure here is that the situation is not expanded in the way the player hoped for.

I come at it from the other direction. What happens if you fail?

I understand what can happen if you fail to pick a lock with lockpicks.
I understand what can happen if you fail to hit the orc with your longsword.
I understand what can happen if you fail to cast fireball at that area.
I don't understand what can happen if you fail to read him to see if

Well, @Charlaquin's suggested example of watching the NPC;s face for micro-expressions doesn't seem any different to me in this respect.

But in either case I don't see much trouble in establishing a consequence for failure. The most obvious is an adverse reaction as the NPC says Why are you staring at me like that?!

 

[pemerton]

An attack is an in-fiction activity. Declaring an Extra Attack is just describing that same activity a second time. “I Attack the orc with my longsword,” and “I Attack him again,” or “I move over here and Attack this other orc with my longsword.”

Action surge does represent something being done by the character - specifically, they are “pushing themselves beyond their normal limits for a moment.” That is admittedly pretty vague, but since the rules clearly describe the exact mechanical effects of the action, it’s not really necessary for me to ask for clarification to adjudicate it properly.

So why can't the player of a 1st level fighter declare these actions?

 

[Bawylie]

So why can't the player of a 1st level fighter declare these actions?

I don’t know. Why can’t they?

 

[Charlaquin]

So why can't the player of a 1st level fighter declare these actions?

I mean, they could, but at that point the results - success, failure, or uncertainty resolved by a dice roll - would fall under DM discretion, since they don’t have an ability that prescribes the effects of such actions. Personally, if a player who didn’t have the extra attack ability described making a second attack on their turn, I would say something to the effect of, “sorry, that’s more than you can do in one turn.” If a player who didn’t have the Action Surge feature described “pushing themselves beyond their normal limits for a moment,” I would ask for clarification about their goal and approach, because that statement alone wouldn’t give me enough information to adjudicate the action without making assumptions.

 

[Xetheral]

Well, see, I wouldn’t expect my players to describe specific, pro-active attempts to catch the NPC in a lie. I find Insight to be one of those non-action “actions” that I would handle more like I do the example of the Dwarven idol below, or better yet, like a trap. Throughout the conversation, I am describing the NPC’s behavior, putting out telegraphs that might give the players an indication of the character’s emotional state, which they can interpret as they will. If the NPC lies, I’ll include a tell of some sort in my description. This gives the players something to follow up on to confirm their suspicions about the character’s intentions, by way of an action. For instance, if you notice that I keep describing the guy’s mustache twitching at the ends of some sentences, but not others, you might suspect that it’s indicative of those statements being false. Rather than take that assumption and roll with it, you might want to say that you’re watching for his mustache twitch to try to recognize a pattern as to when it does or doesn’t.

So how would you handle a player declaring an action to get more information about an NPC's emotional state (not necessarily just spotting a lie) when you either hadn't thought the NPC's emotional state was relevant (and so hadn't telegraphed anything) or the PC is interested in some aspect of the NPC's emotional state other than one you telegraphed?

Ahh. Now, at this point I think it’s important to note that this is how I personally would handle it. Other DMs who use goal and approach might have different answers. But I would be perfectly happy for the player to ask follow-up questions about Mordra, and to answer them according to the character’s background, Proficiencies, and how generally available I think that knowledge is. Some things I will just tell you, some might require research to learn, and that research might require one or more checks to resolve. But I won’t ever ask you to make an Intelligence (Skill) check in response to a question, because a question isn’t an action. Some goal and approach DMs might see this as “20 questions,” and I get that, but I’m personally not bothered by it.

Thanks for clarifying!

Again, I don't see what you're positing. It makes no sense to me that a player can suddenly force a situation with an NPC where the DM is both determining that the NPC is lying/concealing knowledge AND can't do anything about it because it's a surprise. This makes me think you have a wrong conception of what my play looks like. If a player swings into a new line of interaction with an NPC, I will know if this NPC is lying, what their lying about, why their lying about it, and be able to react accordingly because I am the DM and already know all of these things. There's no surprise. I'm also struggling to imagine a situation where this NPC's lies are both important to the character interacting with them and where I've generated a scene where I've utterly failed to anticipate where said important lie couldn't come up and be left reaching for dice to figure out what to do.

So, No and No.

Here's a basic example of the sort of situation I have in mind: the scene is a meeting with the PC's contact about their current quest progress. You've described the scene in that context and telegraphed anything you thought was relevant. The scene goes as you expect. At the end of the scene, before the contact leaves, a player makes the following action declaration: "I want to know if the contact knows anything about [issue related to one of the PCs' other quests], but I doubt they'd tell me if they did. So I'll bring up the topic in conversation and watch their body language to see if they startle or otherwise look uncomfortable when I mention it."

The DM can't have telegraphed yet anything related to this new issue, because it wasn't part of the scene until the character brought it up. So the approach "watch their body language to see if they startle or look uncomfortable" can't be tied to something unique about the situation provided by the DM's description.

You said such a generic action declaration at your table is impermissible, so if this sort of player-initiated action is possible at your table, how would it work? Or is there some other aspect of play at your table that makes the example inapplicable?

 

[Lanefan]

I come at it from the other direction. What happens if you fail?

I understand what can happen if you fail to pick a lock with lockpicks.
I understand what can happen if you fail to hit the orc with your longsword.
I understand what can happen if you fail to cast fireball at that area.
I don't understand what can happen if you fail to read him to see if

For the last one, it's simple: what happens is the conversation continues with none the wiser: you still don't know if the NPC is lying or not, and it then falls on you-as-PC to decide how big a grain of salt (if any at all) you'll be taking the NPC's words with; and-or whether to independently investigate or corroborate them later.

In a broader sense - and this might be significant, I don't know - this is one case where a roll (or action, or move) intended to resolve uncertainty instead results in continued or even increased uncertainty. I personally don't mind this at all, but others might.

 

[iserith]

I don’t know. Why can’t they?

They definitely can since there are no prohibitions with regard to describing what you want to do. The DM can, in this case, either say the attempt fails (no roll) or the DM can perform his or her role as mediator between the rules and the players as laid out in the DMG.

 

[Charlaquin]

So how would you handle a player declaring an action to get more information about an NPC's emotional state (not necessarily just spotting a lie) when you either hadn't thought the NPC's emotional state was relevant (and so hadn't telegraphed anything) or the PC is interested in some aspect of the NPC's emotional state other than one you telegraphed?

That’s not really a thing I’ve encountered in actual play, that I can remember. But, provided the action was declared in terms of a clear goal and approach, I would resolve it like any other action, by evaluating if it seems possible for the approach to achieve the goal, to fail to achieve the goal, and what the cost or consequences for failing might be, and adjudicate accordingly. If I could not glean a clear goal and approach, I would ask for clarification.

Thanks for clarifying!

No problem

Here's a basic example of the sort of situation I have in mind: the scene is a meeting with the PC's contact about their current quest progress. You've described the scene in that context and telegraphed anything you thought was relevant. The scene goes as you expect. At the end of the scene, before the contact leaves, a player makes the following action declaration: "I want to know if the contact knows anything about [issue related to one of the PCs' other quests], but I doubt they'd tell me if they did. So I'll bring up the topic in conversation and watch their body language to see if they startle or otherwise look uncomfortable when I mention it."

The DM can't have telegraphed yet anything related to this new issue, because it wasn't part of the scene until the character brought it up. So the approach "watch their body language to see if they startle or look uncomfortable" can't be tied to something unique about the situation provided by the DM's description.

You said such a generic action declaration at your table is impermissible, so if this sort of player-initiated action is possible at your table, how would it work? Or is there some other aspect of play at your table that makes the example inapplicable?

Well I’m not Ovinomancer, but that seems like a sufficiently clear action declaration to me. The approach of bringing up the subject the my want information about and paying attention to the contacts body language seems to me like it would have a reasonable chance of succeeding at achieving the goal of seeing if it startles him or otherwise causes him discomfort. I’m not sure it has a reasonable chance of failing. Like, either they react with surprise or discomfort or they don’t, either way you have your answer. It is possible that the PC could misread the NPC’s reaction, but I prefer to assume PC competence, so I wouldn’t consider that a reasonable chance of failure unless the NPC wanted to hide their emotional reaction to the subject being brought up. So generally I would have that action succeed without a check. If for some reason the NPC did have a reason to want to hide that reaction, then I would probably ask the player to make a Wisdom check against the NPC’s passive Charisma (Deception). And if the player wanted to add Proficiency for Insight, I would allow that. The consequence here being the opportunity cost.