NPC Deception/Persuasion and player agency

[Bill Zebub]

Who's calling for a "default" here? We're all just discussing our preferences and why they work best for us (and defending them when others disagree).

Yeah, great point. Thank you.

 

[Bedrockgames]

Other than the part about symmetry (once again, if the DM is deciding on the DC in both directions then it's not really symmetric), I agree with all of the above.

For example, I think a great way to use a successful persuasion roll to try to convince a PC is for the DM to use their own knowledge about the player to make a better argument. E.g. "...and I'll let you copy (insert spell player has been wanting) from my spell book..."

In D&D that would be a factor for sure. In my game these skills are always rolled against Wits or Resolve, so it is always symmetrical in the sense that everyone is rolling against the same defense and having rolls made against the same defense (and these themselves can vary but the range is 6-10, and on a d10 dice pool roll)

I agree on the better argument thing. That is more how i would do it. Generally I only roll for these if there is doubt so most of the time, I just play the persuade rank of the character (if an NPC has three ranks in Persuade (ranks go from 0-3 in my game), I'm going to speak as persuasively as I can (usually this will be less about argument and more about making the NPC as personable as I can, as I find that works better)

 

[Micah Sweet]

I think things like Deception and Persuasion are always tricky in an RPG for this and other reasons. The most common thing I see is persuasions is something PCs use against NPCs but not vice versa. The way I do it in my own games is not have these skills impact actions directly. They simply tell the outside world how convincing, charming, etc the person is being (a sleazy sales person who shows up at your door and rolls a 20, maybe be charming but that doesn't mean you are automatically going to buy his vacuum cleaner because there are so many other things to consider, like whether you want or have the money for a vacuum cleaner). I find when skills like this are approached in this way, you can have them go in the direction or PCs or NPCs, without impacting the agency of either (because I think NPCs should also have agency)

Honestly, that's the kind of parity I can get behind.

 

[payn]

I'm having trouble wrapping my head around this, and would love some kind of illustrative example of what you mean.

If I map that to the example I used above, traps, my experience has been that with mechanics players tend to say, "I roll to detect to traps" or "I roll to disarm it". But without mechanics they try to engage and narrate what they are doing.

Yeap, thats an interesting dynamic you have pointed out. Skills have become a binary mechanic for the most part, though the social aspect has not been used or accepted as such. I've long thought there ought to be a separation of binary skills, like athletics for jumping pits, and something like the performance skill in which a PC is dancing or acting. One is a precise action, the other is a context driven one.

I know this is getting a little ahead of our current conversation, but my fix would be to binary one aspect, and contextualize the other. For example, spotting a trap or not is binary, disabling would become a more involved process of gethering info and taking actions instead of a single roll check. Alternatively, discovering the trigger or disable point could be the multistep investigation and then disarming (or triggering) would be the binary check.

In any event, the above should highlight my GM style in that there is usaully a context driven session (that may or may not involve skill checks) that leads to a binary moment where you get the resolution before starting the process over again. Something like jumping a pit is spared the process becasue it is unccesary the action type and resolution is obvious.

Can you describe how the opposite happens, although with social mechanics, in ways that you've seen?

Using the trap example, lets say the player knows its a blade trap. The player decides they are just going to walk face first through it because it wont kill their PC and they can just heal up afterwards. They are using meta knowledge to push past an obstacle instead of engaging it. They are ignoring the milieu of the game. I see ignoring social results without aknwoledging them as similar.

To be fair, I dont think a trap is comparable to a social situation. A trap doesnt change its position or offer up any defense based on PC actions, its simply static. My concern is the notion of asymetrical play you mentioned earlier in the discussion. If a PC can just yell, "diplomacy" and chuck a D20 and the NPC must abide by the result, I dont get why an NPC cant do the same? Mind you, im not saying thats how it should work at all. I think a social mechanic check is earned through role play and follows a logical set up from an exchange between players and GM. The reuslt informs the GM/player how the character ought to act to the situation moving forward. Ignoring the context of the situation and acting on player meta knwoledge instead, is out of bounds to me. Or has been described earlier in the discussion as being a poor sport. (I'll note, if folks want to simplify social resutls as binary results on way or another, I dont have a problem with that for them. I just dont want to play that way and expect a more interactive experience.)

For a social exchange example we can use the time tested getting past a guard. A PC decides to tell the guard they are the king's cousin and he is expecting them. They roll a bluff check or equivalent. Now, the GM plays the guard depending on thier context. Maybe this guard is firecely loyal and unmovable in their duty? The guard might believe the PC is a the King's cousin, but his orders are his orders. He would direct them to an authority to try their case there, but will not allow them to pass. Or, the guard might hate their duty, think the king is a jerk, and allow the PC to pass anyways. Depends also on how much prep the PC put into the situation. If this is an off the cuff response? Probably get the guard's former reponse. If they put in a disguise, learn local customs, play the bit up, then the latter might be likley from the guard.

Flip it to the PCs and lets use an interogation scene. The PCs are trying to get info from an NPC. The NPC makes a sucsessful decpetion check. If the PC (not the player) knows this NPC and their reputation, they might proceed with caution or even ignore the result. If they have no context of this NPC at all, they likley should act on the information. Whether its with suspicion and/or caution, the info should be considered. If its flat out ignored becasue the player knows of other avenues and options (that the PC is not aware of), I consider that out of bounds. Also, note how the NPC above gets to consider thier internal and external situation before acting, but always aknwoledges the play on the table.

 

[Bill Zebub]

I agree on the better argument thing. That is more how i would do it. Generally I only roll for these if there is doubt so most of the time, I just play the persuade rank of the character (if an NPC has three ranks in Persuade (ranks go from 0-3 in my game), I'm going to speak as persuasively as I can (usually this will be less about argument and more about making the NPC as personable as I can, as I find that works better)

This makes me realize that one argument I have NOT seen (which surprises me) is: "If social encounters are being resolved by the strength of the argument, without rolling, that penalizes players who have invested in Charisma and social skills."

When I'm GMing, I reward that investment in a couple of ways:

• I factor those mechanical bonuses into whether or not a plan works, or whether or not a roll is required.
• I don't like to gate information behind random rolls. If somebody has relevant skill/knowledge I just give it to them. Sometimes I'll have the whole table roll and give it to whoever rolls the highest.

I get that some people...especially if they are used to a different mode of play...will still feel shortchanged if they can't make a dice roll and add their bonus. What can I say? It's not how I like to run games. Fortunately I teach lots of new players, so I get to indoctrinate them however I like. Mwuhahahahahaha.....

 

[Aldarc]

No, definitely not. It does not have to be 1st person acting.

Maybe this will clarify (using an example from upthread):

1. "If you don't help us I will expose your affair with the Count's daughter to the whole court!"
2. I'll threaten to tell the court about the man's affair if he doesn't help us.
3. Can I roll Persuasion to see if he'll help us?

I'm just saying that I greatly prefer #1 or #2 (and, really, I prefer #2 myself) to #3.

I think that a pretty significant problem rests in presenting #1-2 as mutually exclusive with #3 or even framing #3 as the end of the conversation.

Yes. In my experience if there's a button on the character sheet players...especially new players, and I get a lot of those...will try to press that button.

EDIT: To further clarify, when faced with a challenge, I often see players looking over the list of skills on the character sheet to see which one they want to "use", instead of thinking about what the problem is and what some creative solutions might be. I see that occur a lot less frequently in Shadowdark, which does not have a list of skills on the sheet.

Players want to use what tools that they have at their disposal, and this just says that Shadowdark has less tools for players to use. I understand that this is a feature of OSR play, but I wish that it would stop negatively characterizing non-OSR play as button-mashing or roll-to-win.

Edit: FWIW, Vincent Baker developed Apocalypse World partially to design around what he perceived as "I roll for Perception" style play in 3e as well as his wife's desire for more free-form roleplaying. So rolls are not buttons that are pressed. Instead, they represent conditions being met in the fiction that mandate a roll.

 

[Aldarc]

This makes me realize that one argument I have NOT seen (which surprises me) is: "If social encounters are being resolved by the strength of the argument, without rolling, that penalizes players who have invested in Charisma and social skills."

When I'm GMing, I reward that investment in a couple of ways:

• I factor those mechanical bonuses into whether or not a plan works, or whether or not a roll is required.
• I don't like to gate information behind random rolls. If somebody has relevant skill/knowledge I just give it to them. Sometimes I'll have the whole table roll and give it to whoever rolls the highest.

I get that some people...especially if they are used to a different mode of play...will still feel shortchanged if they can't make a dice roll and add their bonus. What can I say? It's not how I like to run games. Fortunately I teach lots of new players, so I get to indoctrinate them however I like. Mwuhahahahahaha.....

You cannot be the GM for everyone. People are welcome to tell anecdotes of what they would do or how they run things, but that's not something that can be relied upon or can be reproduced consistently at tables in the same manner that systems or rules can. And one of the issue with free-form RP of this nature is that it makes things a LOT more heavily GM-dependent, which is something that I don't particularly enjoy as a player or as a GM.

 

[Bill Zebub]

Yeap, thats an interesting dynamic you have pointed out. Skills have become a binary mechanic for the most part, though the social aspect has not been used or accepted as such. I've long thought there ought to be a separation of binary skills, like athletics for jumping pits, and something like the performance skill in which a PC is dancing or acting. One is a precise action, the other is a context driven one.

I know this is getting a little ahead of our current conversation, but my fix would be to binary one aspect, and contextualize the other. For example, spotting a trap or not is binary, disabling would become a more involved process of gethering info and taking actions instead of a single roll check. Alternatively, discovering the trigger or disable point could be the multistep investigation and then disarming (or triggering) would be the binary check.

In any event, the above should highlight my GM style in that there is usaully a context driven session (that may or may not involve skill checks) that leads to a binary moment where you get the resolution before starting the process over again. Something like jumping a pit is spared the process becasue it is unccesary the action type and resolution is obvious.

Using the trap example, lets say the player knows its a blade trap. The player decides they are just going to walk face first through it because it wont kill their PC and they can just heal up afterwards. They are using meta knowledge to push past an obstacle instead of engaging it. They are ignoring the milieu of the game. I see ignoring social results without aknwoledging them as similar.

Interesting analogy. Are you saying that the player shouldn't be allowed to walk through the trap and take the damage? If you are, then I disagree. (After all, they probably don't know how much damage it is, and even if it's survivable, HP are a resource.)

So what's the equivalent in social interactions? A couple such scenarios have been offered up-thread, but in general the idea is that there is an in-game cost to ignoring the result of social rolls.

I'm fine with all of that. In fact I kind of love it. But while it's easy to think of scenarios where there is such a cost, in practice it is probably hard to figure out what that cost would be in all scenarios.

To be fair, I dont think a trap is comparable to a social situation. A trap doesnt change its position or offer up any defense based on PC actions, its simply static. My concern is the notion of asymetrical play you mentioned earlier in the discussion. If a PC can just yell, "diplomacy" and chuck a D20 and the NPC must abide by the result, I dont get why an NPC cant do the same? Mind you, im not saying thats how it should work at all. I think a social mechanic check is earned through role play and follows a logical set up from an exchange between players and GM. The reuslt informs the GM/player how the character ought to act to the situation moving forward. Ignoring the context of the situation and acting on player meta knwoledge instead, is out of bounds to me. Or has been described earlier in the discussion as being a poor sport. (I'll note, if folks want to simplify social resutls as binary results on way or another, I dont have a problem with that for them. I just dont want to play that way and expect a more interactive experience.)

For a social exchange example we can use the time tested getting past a guard. A PC decides to tell the guard they are the king's cousin and he is expecting them. They roll a bluff check or equivalent. Now, the GM plays the guard depending on thier context. Maybe this guard is firecely loyal and unmovable in their duty? The guard might believe the PC is a the King's cousin, but his orders are his orders. He would direct them to an authority to try their case there, but will not allow them to pass. Or, the guard might hate their duty, think the king is a jerk, and allow the PC to pass anyways. Depends also on how much prep the PC put into the situation. If this is an off the cuff response? Probably get the guard's former reponse. If they put in a disguise, learn local customs, play the bit up, then the latter might be likley from the guard.

Flip it to the PCs and lets use an interogation scene. The PCs are trying to get info from an NPC. The NPC makes a sucsessful decpetion check. If the PC (not the player) knows this NPC and their reputation, they might proceed with caution or even ignore the result. If they have no context of this NPC at all, they likley should act on the information. Whether its with suspicion and/or caution, the info should be considered. If its flat out ignored becasue the player knows of other avenues and options (that the PC is not aware of), I consider that out of bounds. Also, note how the NPC above gets to consider thier internal and external situation before acting, but always aknwoledges the play on the table.

Ok, but (again) who sets the DC when a PC is being deceived/persuaded/intimidated/etc.? I think we all agree that the DM decides how hard it is to deceive the guard. Who decides how hard it is to deceive the PC?

 

[Bill Zebub]

I think that a pretty significant problem rests in presenting #1-2 as mutually exclusive with #3 or even framing #3 as the end of the conversation.


Players want to use what tools that they have at their disposal, and this just says that Shadowdark has less tools for players to use. I understand that this is a feature of OSR play, but I wish that it would stop negatively characterizing non-OSR play as button-mashing or roll-to-win.

Fair. I am saying "button pushing" pejoratively. Can you suggest terminology for "invoking an ability listed on a character sheet instead of trying to think creatively about the problem"?

 

[Celebrim]

I think things like Deception and Persuasion are always tricky in an RPG for this and other reasons. The most common thing I see is persuasions is something PCs use against NPCs but not vice versa. The way I do it in my own games is not have these skills impact actions directly. They simply tell the outside world how convincing, charming, etc the person is being (a sleazy sales person who shows up at your door and rolls a 20, maybe be charming but that doesn't mean you are automatically going to buy his vacuum cleaner because there are so many other things to consider, like whether you want or have the money for a vacuum cleaner). I find when skills like this are approached in this way, you can have them go in the direction or PCs or NPCs, without impacting the agency of either (because I think NPCs should also have agency)

From my perspective, I think ultimately your complaint resolves down to how bad most social resolution systems are and how little attention they play to real world nuance - like getting someone to fall for a lie depends not just on how charming you are but also how good the lie is. The vast majority of cons depend not just on getting someone to believe a lie, but on exploiting weaknesses in the character of the mark - typically their greed. A lot of cons rely on getting the mark to believe they are the one tricking and exploiting you, which of course isn't going to work with someone who doesn't want to exploit or trick someone. So the character of the person you are trying to manipulate matters. It might be hard to exploit the greed of an honorable person, but easy to exploit their pity - for example. This is why I'm always taking the conversation and making a judgement about what circumstance modifiers apply to the particular conversation we've had when trying to figure out how difficult the social test is.

I liken this to a room with many walls of differing height and smoothness, and the player is choosing implicitly or explicitly which that they want to try to climb, knowing that some walls are more difficult than others and others lead to falls on more dangerous of surfaces. The conversational strategy is effectively trying to figure out which wall is easiest and less risky. You don't necessarily automatically succeed just because you have a great strategy - charisma isn't a dump stat - but a great strategy does help.

But, more than that, I would like to direct everyone's attention to what you call the "other reasons". Because the real reason that an NPC doesn't roll to convince or deceive a PC isn't volition and player agency, but rather because again - the mind of the player is always present in the game universe. There can be no consistent or effective or reasonable rules around the idea of "don't metagame". I have come to the conclusion that asking a player to not metagame is really impossible, since the player is asked to imagine what they would do if they didn't know something, and that is impossible to know. If you are presented with a puzzle or a riddle, but you have been told the answer to a puzzle or riddle, there is no way to approach that puzzle or riddle as if you didn't know the answer. Everything you do in that circumstance, whether solving the riddle or not solving the riddle is still metagaming.

So instead of trying to force the players to not metagame and try to force the players to act out as if they have been deceived when they haven't, I simply deceive them myself and don't leave it up to a die role. Their character may not be deceived, and if they are wise and fortunate I may tell them, "No, you are pretty sure the NPC is lying." But the reverse isn't really possible. I can't tell them, "No, your character believes them" if in fact the player doesn't, because then they have to metagame what their character would do and that's not really possible because in reality even if deceived they might take "trust but verify" steps or blindly do things that would reveal the deceit. After all, when the players themselves are deceived that is how they behave. So why should I ask them to try to imagine how they'd behave in the absence of knowing the answer to the riddle? They can't. It's not fair and it's really impossible. Ultimately that would end up being based on my opinion of how they would behave, and why should that rule the day?

So I don't have any "no metagaming" rules generally, not just in this situation.