D&D 5E Modeling Uncertainty

[Lanefan]

If you mean using a secret DC, the issue I have with that is that the player has only the strength of his roll, not the difficulty of the challenge, on which to base his estimate of his probability of success. So, for example, on a low roll the player will probably guess he has failed, but what if the trap was an easy one and he has in fact succeeded? That's a false negative, which is something I'd like to see, but not I think in a good way; narratively I think it takes something away from the game.

Secret DC and secret roll - in other words, player says she's trying something (check the door for traps), then the DM rolls some dice and narrates success, failure, or something in between...any one of which might be put in more or less vague terms. "You check it and are sure there's nothing there" is a lot different than "You check it and you think there's nothing there, but can't be sure".

Main reason for doing it this way is to hide the difference between complete failure to find an existing trap and complete success in finding out there's no trap there at all, as in the character's eyes the end result is exactly the same - she finds nothing but has no way of knowing if there's really nothing there or if she simply missed it. Which means the run of play thenceforth should also be the same, which it probably would not be if the player knew she had rolled a '2' on her find-traps instead of a '19'.

Perhaps the DM could create some categories ("easy", "extremely hard", etc.) to share with the player, but at that point we've both added a new element to the roll and kept something secret, which I'm claiming are the two requirements to making this work.

Alternatively we could use the opposed roll. As long as it's kept secret then the player will doubt his chances of success. Again, though, he's not going to know what his chances were in the first place; all he has to go on is his own roll, with no knowledge of what he was facing.

Which is just how it should be: the character (usually) doesn't know the difficulty of what she's up against until she tries to deal with it. If it's something she can look at ahead of time e.g. how hard will it be to climb this wall then sure, she can get an idea of what she's facing. But most of the time she'll have no way of knowing in advance how tough it'll be to disarm a trap or find a secret door (if there's any there to find at all), and nor should she.

Same with social interactions - sure, the guard may have just broken out in a sweat; which could mean he's lying...or that the slow-acting poison in his lunch is taking hold (most of the other guards look a bit uncomfortable too, if anyone bothers to look)...or just that he really needs to pee.

My players know they can trust me to try my best to engage, engineer and support a fair, fun game.

But they rarely, if ever, trust any NPC, door or innocent wildlife I introduce - at least at first. Because they're players and this is D&D

A perfect end result, I'd say.

Lan-"the trick is to ask oneself 'what would the character see?' and go from there"-efan

 

[Guest 6801328]

Ok, now I'm getting depressed. Either I'm doing a really bad job explaining what I see as the limitation....or I'm just crazy.

Secret DC and secret roll - in other words, player says she's trying something (check the door for traps), then the DM rolls some dice and narrates success, failure, or something in between...any one of which might be put in more or less vague terms. "You check it and are sure there's nothing there" is a lot different than "You check it and you think there's nothing there, but can't be sure".

Main reason for doing it this way is to hide the difference between complete failure to find an existing trap and complete success in finding out there's no trap there at all, as in the character's eyes the end result is exactly the same - she finds nothing but has no way of knowing if there's really nothing there or if she simply missed it. Which means the run of play thenceforth should also be the same, which it probably would not be if the player knew she had rolled a '2' on her find-traps instead of a '19'.

But how does this work mechanically? Let's say the roll succeeds, and the DM says, "You think you succeeded." First, how does the DM choose between "you think you succeeded" and "you failed". There are basically three alternatives:

1) The DM says "you failed" when the roll fails. In this case saying "You think you succeeded" really means "you succeeded", so you may as well just say "You succeeded."

2) The DM always says "you think you succeeded" regardless of whether it is a success or a failure. In this case the player really has learned nothing useful by the attempt. Sure, maybe it has reduced the chance that he's going to die to a trap, but it hasn't actually changed his decision making. He's decided to open the chest, and finding/disarming the trap is simply something he does before opening chests, but it has zero impact on the suspense of the scene. Because he has no idea if his odds have changed.

3) The DM sometimes says "you think you succeeded" on a failure. This seems the most interesting scenario, but then the question is: how does the DM choose? Maybe it's something specific, like any roll within 5 of either side of the DC gets the "you think" answer. In that case we are in fact using a new rule, although one that gives the player less information than what I'm proposing. (It's a 50:50 chance any time the DM says "You think.") Or maybe the DM just uses his own judgment and improvises, but in that case the player has nothing to base estimates on, so will act as if it's the same as 50:50, anyway. This is what I was referring to up-thread with "100%/50%/0%."

Which is just how it should be: the character (usually) doesn't know the difficulty of what she's up against until she tries to deal with it. If it's something she can look at ahead of time e.g. how hard will it be to climb this wall then sure, she can get an idea of what she's facing. But most of the time she'll have no way of knowing in advance how tough it'll be to disarm a trap or find a secret door (if there's any there to find at all), and nor should she.

Same with social interactions - sure, the guard may have just broken out in a sweat; which could mean he's lying...or that the slow-acting poison in his lunch is taking hold (most of the other guards look a bit uncomfortable too, if anyone bothers to look)...or just that he really needs to pee.

A perfect end result, I'd say.

I think maybe what we need (for anybody sufficiently interested) is to step through a specific example with specific numbers.

I'll start: I've got a character with +5 in Insight, and I'm talking to a guard, and I say to you, "I want to see if I can tell if he's lying."

What do you, the DM, do?

 

[Tony Vargas]

But how does this work mechanically? Let's say the roll succeeds, and the DM says, "You think you succeeded." First, how does the DM choose between "you think you succeeded" and "you failed".

The wonder of taking resolution behind the screen is that it really doesn't matter anymore how it works mechanically. You create the impression there's legitimizing mechanic going on behind the screen. That's all you need. You can come up with as convoluted or simple a mechanic as you want to use, even vary it a little.

Since you want to keep at least part of the roll in the open, though, a variant like you propose seems like a fine idea.

 

[Guest 6801328]

The wonder of taking resolution behind the screen is that it really doesn't matter anymore how it works mechanically. You create the impression there's legitimizing mechanic going on behind the screen. That's all you need. You can come up with as convoluted or simple a mechanic as you want to use, even vary it a little.

Ok, but what does this all mean from the perspective of the player? What information are they using to estimate their probability of success? I posit that it's a total crap shoot; the player's best strategy is to assume a 50/50 probability.

Since you want to keep at least part of the roll in the open, though, a variant like you propose seems like a fine idea.

I do like keeping the player rolls in the open, but that's really a side issue. The main issue is that without any objective/neutral information for the player to process (that is, the sort that comes from dice rolls), the player only has the words, intonation, body language, past history of the DM to base a decision on. And part of what I'm suggesting is that DMs cannot reliably convey any probabilities more nuanced than 100%, 50%, 0%.

 

[Lanefan]

Ok, but what does this all mean from the perspective of the player? What information are they using to estimate their probability of success?

That's just my point - in a case like this the player flat-out shouldn't know their probability of success!

The +5 Insight tells the player that compared to most people their character's odds are better than those of an average Joe; no problem there, and if that's the best Insight value in the party then that's probably the right character to try this. But the player (in character) shouldn't know any more than that - they've never met this guard before; they don't know whether he's a brilliant liar or not (maybe his capacity for hiding the truth is why he got the job in the first place), or whether he's even lying at all...or if he is, whether he's lying about what the PC thinks he is or covering up something else entirely unrelated.

[MENTION=996]Tony Vargas[/MENTION] is quite right in saying something like this in fact moves out of the realm of mechanics and into the realm of straight-up DM narration. On the player's "I want to see if I can tell if he's lying" you-as-DM have to quickly decide (if you haven't pre-determined this already) how good a liar the guard is - assuming he's lying at all - and either narrate whatever result you like or, if dice are the arbiter, snap-assign a DC and roll against it behind the screen again followed by narration of whatever the dice steer you toward. And in any case the player (in character) has no idea what they're up against, and nor should they.

Even then, with or without use of dice you can simply narrate "Well, the guard looks nervous; he's unsure of himself and keeps glancing towards the guardhouse while talking to you." Does that mean he's lying? Does it just mean it's his first day on the job and he doesn't want to mess up? Does it mean there's someone in the guardhouse ready to shoot him if he makes a wrong move? Or that he's got a paramour hidden in there? A narration like this opens the door for further investigation...maybe the party sneak goes and peeks in the guardhouse window to check it out, maybe Insight-guy calls the guard on his nervousness and asks what's wrong, and so forth.

I posit that it's a total crap shoot; the player's best strategy is to assume a 50/50 probability.

The player's best strategy is to assume nothing and carry on.

I do like keeping the player rolls in the open, but that's really a side issue. The main issue is that without any objective/neutral information for the player to process (that is, the sort that comes from dice rolls), the player only has the words, intonation, body language, past history of the DM to base a decision on. And part of what I'm suggesting is that DMs cannot reliably convey any probabilities more nuanced than 100%, 50%, 0%.

Base what decision on...the decision that the guard is lying, or the decision even to check for such?

Either way, the player (in character) can arbitrarily make those decisions at any time without resorting to any mechanics at all: "'You're lying!' says Terazon while gently grabbing the guard and pushing him up against the guardhouse wall.". The character has simply decided the guard's lying (maybe or maybe not based on any evidence at all) and it's up to the guard to bail himself out...or call for help...or admit that yes, he's been ordered to say things that aren't true...

Lan-"despite 3e's best attempts, mechanics are not the answer for everything"-efan

 

[Tony Vargas]

I posit that it's a total crap shoot; the player's best strategy is to assume a 50/50 probability.

It's like the one version of Augury that gave you an answer to a yes/no question that 50% likely to be correct.

I do like keeping the player rolls in the open, but that's really a side issue. The main issue is that without any objective/neutral information for the player to process (that is, the sort that comes from dice rolls), the player only has the words, intonation, body language, past history of the DM to base a decision on. And part of what I'm suggesting is that DMs cannot reliably convey any probabilities more nuanced than 100%, 50%, 0%.

I think a DM could certainly convey a lot of uncertainty, but for the player to translate that into a %, sure, there's no real way to do it. That'd be a benefit if you're going for immersion, I suppose (the PC has no real way of gauging things in that way, either).

[MENTION=996]Tony Vargas[/MENTION] is quite right in saying something like this in fact moves out of the realm of mechanics and into the realm of straight-up DM narration.

Darn, that was what I was saying. I guess I didn't put enough uncertainty into that post...



" .... Oh. We don't demand solid fact! What we demand is a total absence of solid facts! We demand rigidly defined areas of doubt and uncertainty! "

 

[Guest 6801328]

That's just my point - in a case like this the player flat-out shouldn't know their probability of success!

Do you feel the same way about combat? That the players should declare, "I attack the ogre with my sword" and not know their odds, even if they don't know the exact AC of the ogre? Do you let them roll their own dice? (For the record, I could understand and appreciate a game in which the answer was Yes. I just don't think that game is D&D 5e.)

So maybe we're only going to talk past each other because I think the player should know their (approximate) probability of success, even for things like detecting lies. They should at least have a belief about their probability, even if in some circumstances it changes (e.g., the guard is actually a grandmaster rogue and an expert in deception.)

@Tony Vargas is quite right in saying something like this in fact moves out of the realm of mechanics and into the realm of straight-up DM narration. On the player's "I want to see if I can tell if he's lying" you-as-DM have to quickly decide (if you haven't pre-determined this already) how good a liar the guard is - assuming he's lying at all - and either narrate whatever result you like or, if dice are the arbiter, snap-assign a DC and roll against it behind the screen again followed by narration of whatever the dice steer you toward. And in any case the player (in character) has no idea what they're up against, and nor should they.

Even then, with or without use of dice you can simply narrate "Well, the guard looks nervous; he's unsure of himself and keeps glancing towards the guardhouse while talking to you." Does that mean he's lying? Does it just mean it's his first day on the job and he doesn't want to mess up? Does it mean there's someone in the guardhouse ready to shoot him if he makes a wrong move? Or that he's got a paramour hidden in there? A narration like this opens the door for further investigation...maybe the party sneak goes and peeks in the guardhouse window to check it out, maybe Insight-guy calls the guard on his nervousness and asks what's wrong, and so forth.

This is where we start to disagree. I assert that at this point is becomes a matter of the DMs improv acting ability versus the player's "Insight" score, not the NPC versus the PC. And unless the DM is really bad at it, or the player knows his DM really well, the player has no information to go on; it's pure hunch. So he may as well flip a coin. (Ironically, the better the DM's acting ability, the less information the player has.)

I posit that it's total crap shoot; the player's best strategy is to assume a 50/50 probability.

The player's best strategy is to assume nothing and carry on.

Here we are saying essentially the same thing: by "assume 50/50" I meant "I got zero useful information from that, so it may as well be random."

Look, I can't disagree with most of what you are saying, but then what is the purpose of the Insight skill? Why have stats and dice at all? Ok, I could see how the DM might glance at the character stats and then decide what to reveal, but why have dice?

I know you all mean well, but my original question was about mechanics and probability. Some of the advice here sounds a lot like "You don't need mechanics, you just need to DM better." But I do not think it is good DMing, or good playing, to resolve deception/persuasion (and other) events through the improv abilities of the participants. I want the character, animated by the informed decisions of the player, to determine the outcomes. The DM should remain a neutral arbiter.

There's still lots of room for narration and roleplaying. But it's done to interpret the fall of the dice, not to determine modifiers to the dice, or to replace them.

Perhaps we just play the game differently.

 

[Guest 6801328]

That'd be a benefit if you're going for immersion, I suppose (the PC has no real way of gauging things in that way, either).

Oh, thank you for saying that!

Yes, that's exactly what I'm going for, and what I always mean by immersion: the player is experiencing the emotions (doubt, elation, fear, anger, etc.) of the character. My definition of immersion most explicitly has nothing to do with the acting ability of anybody at the table, or the consistency of the fiction, or whether anybody "metagames", or how accurate the weight of plate armor is.

I want the player to think, "Well I know my character is really, really good at detecting lies, so I'm pretty darned sure this guard is telling the truth. But I could be wrong..."

I've yet to see an explanation/mechanic for how that can be done, in a way that correlates to the ability scores of the character, solely through secret rolls and DM narration. Or through a single roll. I'd love to see a simpler, more elegant solution than the one I proposed.

 

[Tony Vargas]

Do you feel the same way about combat? That the players should declare, "I attack the ogre with my sword" and not know their odds, even if they don't know the exact AC of the ogre?

Arguably, they shouldn't know the exact AC of the ogre.

Do you let them roll their own dice? (For the record, I could understand and appreciate a game in which the answer was Yes. I just don't think that game is D&D 5e.)

I do let players roll their own dice, but I also think 5e is ideally suited for taking all resolution behind the screen. You can remove the telegraphing inherent in calling for a roll vs narrating a result (when you want to - when you want to telegraph 'that was easy/impossible' just don't roll, the silence of the dice speaks volumes).

So maybe we're only going to talk past each other because I think the player should know their (approximate) probability of success, even for things like detecting lies. They should at least have a belief about their probability, even if in some circumstances it changes (e.g., the guard is actually a grandmaster rogue and an expert in deception.)

They do get to know their own bonus, so they have an idea how good they are relative to eachother, and relative to the theoretical maximum possible bonus.

I assert that at this point is becomes a matter of the DMs improv acting ability versus the player's "Insight" score, not the NPC versus the PC.

I do not challenge that assertion.

And unless the DM is really bad at it, or the player knows his DM really well, the player has no information to go on; it's pure hunch. So he may as well flip a coin. (Ironically, the better the DM's acting ability, the less information the player has.)

That can all get very nuanced.

Here we are saying essentially the same thing: by "assume 50/50" I meant "I got zero useful information from that, so it may as well be random."

Yes. Thus 'uncertainty.' What you do know is how good you are, just not how well you rolled. If you're very good at something, you can, when presented with such uncertainty, probably assume you're right.

Look, I can't disagree with most of what you are saying, but then what is the purpose of the Insight skill? Why have stats and dice at all? Ok, I could see how the DM might glance at the character stats and then decide what to reveal, but why have dice?

You absolutely can use the dice. But, you don't absolutely need them. In theory, you could run 5e on pure DM narration. It wouldn't be against the rules - it'd be straining credulity that nothing's ever 'in question,' but not against the rules.

But I do not think it is good DMing, or good playing, to resolve deception/persuasion (and other) events through the improv abilities of the participants. I want the character, animated by the informed decisions of the player, to determine the outcomes.

Nod. That's the line I sometimes draw with "player as resolution system." I'm not suggesting that the DM just 'lie' (is it really a lie, it's all made up anyway?) in character and let the players try to figure it out.

I've yet to see an explanation/mechanic for how that can be done, in a way that correlates to the ability scores of the character, solely through secret rolls and DM narration. Or through a single roll. I'd love to see a simpler, more elegant solution than the one I proposed.

For the record, I do find your original idea quite reasonable (not at all in-elegant, either), for many of the same reasons I find taking resolution behind the screen entirely to be reasonable. (Hope that doesn't offend.)

BTW, I think it'd be nice to map confidence not to total bonus nor Expertise, but to proficiency bonus (level, I guess, that means, mainly). If you're untrained, you should have the least certainty about your hunch, however talented you may be. If you're trained you'll be more consistent. It may not be realistic, but I think it'd be good for keeping a sense of advancement and competence.

 

[Guest 6801328]

I'm not disagreeing with you at all, Tony. Your answers have very much acknowledged the validity of both approaches. And I do not mean to invalidate pure DM arbitration, either, in many situations. It just doesn't achieve what I want in this class of scenarios.

As an example of how DMing "lying" works just fine by me, in another thread long ago I argued that if an NPC succeeded at a deception roll to trick a PC into stepping onto a trap, I think it's acceptable to let the player decide where to move, and then to place the trap there. Some posters HOWLED over that one, I can tell you. I like it because the DM doesn't have to tell the player what he thinks or does, which for me is a sacred principle.