(HB) Most Social Actions Should Not Target Characters.

[MGuy]

Is this useful?

It depends on what your design goals are. If simplicity is the only concern then this seems simple enough.

[Zinegata]

I've seen a bunch of discussion about social mechanics recently both on the Den and on various discords, and the conversations I've seen have focused mostly on systems for convincing people to do things or changing their opinions and attitudes. I think some people are hoping to get by with nothing but a system for changing people's minds, and others intend to have other mechanics but see a generic “persuasion” rule as a good place to start. I don't think it is. I think that you may end up needing a generic persuasion or attiude-adjusting mechanic, but you will definitely need other social rules and you should probably start with them and then work your way around to generic persuasion.

I think some of the basic principles of RPGs need to be reiterated to support this point.

Rules in an RPG are not the same as with most other games. This is why testing for their mechanical robustness had always been a lower priority in this design space.

Rules in an RPG are instead guidelines, which are meant to empower the DM (primarily) and the players to become co-designers. D&D, in itself, is not a complete game - because while it has systems, monsters, and such - none of it works until a DM puts together encounters and campaigns, and the players create characters.

The gap in "social encounter" rules isn't - and has never been - the lack of resolution mechanics. Folks have been doing arbitrary Cha checks since forever.

The gap is instead the lack of good tools and guidelines on how to create good social encounters.

To demonstrate: For combat, D&D has plenty of tools like map supplements, monster books, and environmental effects. They offer DM co-designers more interesting encounter and campaign design options to work from.

By contrast monster/NPC reactions to social encounters are often based on the DM's own design. And that's to a large extent because there is a "Monster Manual", but no "People to Persuade Manual".

[Foxwarrior]

Here, let me recount an anecdote that's relevant to the original topic: One time, I made a Shadowrun character who was all about talking really well. He could roll an absolutely preposterous number of dice on talky checks. However, in practice we tried not to show our faces at all to anyone except the Johnson until the grenade launchers came out, investigation of sites tended to involve only magical and hacking scouting missions, so I didn't get to converse that many NPCs into oblivion. (I'm actually sort of disappointed that the hackers in the party weren't that interested in providing me with a fake identity for Leverage style scams). Where the character actually ended up being important was as a Fixer, because he could use Charisma to buy things really fast. I imagine that "finding the right random NPC to persuade, and persuading them" would have been a genuinely useful thing for him to bring to the scouting phase if it had existed more.

[Orion]

Personally I'm not sure I want "Master Manipulators" to exist in my game on either end, so I think that just functionally makes most diplomacy systems bad for me.

I have never in my entire life met anyone who believes ~X and then was convinced to believe X in a time frame of a day.

I definitely sympathize with this perspective. I mean, I think in any game that tries to go up to level 20 D&D you definitely reach a point where that kind of thing becomes possible at some point (and in some low-power games it will be genre appropriate); I don't really want to go to level 20 anyway, though.

[Orion]

I think the reason conversations around social mechanics don't focus too much on things that aren't messing with people's feelings, desires, and actions is because that it 'is' the big deal thing where socializing is concerned. It's the most powerful/useful result you can get from any interaction with any person.

This comment is just weird to me. It's like saying "People don't really talk much about Summon Monster or Fear because Dominate is the best spell effect." There is no reason that abilities which don't work by making arbitrary line-edits to NPC motivations should be relegated to being "minor," "utility," "flavor," or "non-combat" or whatever. They can have big effects that decide adventures. For example, you don't have to write an ability that lets you target a specific NPC and make them an ally to give out the ability to bring in NPC allies. You can write all kinds of abilities that let you introduce new generic NPCs to the story who are already interested in fighting for you. The obvious ways to go involve membership in gangs or secret societies or whatever but you could even have abilities like "a really good damsel-in-distress routine" where given a couple of minutes in a crowded public space you can attract the attention of 1d3 gallant strangers who will fight to defend you. There are a variety of ways social skills can make people less likely to fight against you or more likely to stick with it. I personally am a big fan of morale rules and would like to see morale checks become a regular part of combat again. I think there's a lot of room for social ops to sap an organization's morale before an encounter begins, and for codified uses of both intimidation and diplomacy-style skills to penalize enemy combat morale.

[OgreBattle]

Most Social Actions Should Not Target People.
Persuasion Mechanics Are Incomplete.
Maybe You Can Have A Little Persuasion, As A Treat.

The gap is instead the lack of good tools and guidelines on how to create good social encounters.

To demonstrate: For combat, D&D has plenty of tools like map supplements, monster books, and environmental effects. They offer DM co-designers more interesting encounter and campaign design options to work from.

By contrast monster/NPC reactions to social encounters are often based on the DM's own design. And that's to a large extent because there is a "Monster Manual", but no "People to Persuade Manual".

Narrowing it to the scope of the game can help with how open ended this concept is.

D&D has sexy demons and nerdy wizards cast charm person on PC's, so a suave CHA diplomancer should get similar from their character level resource investment. It's mostly in the context of combat because D&D revolves around combat.

A detective noir story's use of Social Actions should revolve around advancing or halting the gumshoes from uncovering the mystery. So a suave playboy or gun moll can distract the PC from their stakeout, slip something into their drink, become a human shield that the PC must roll to shoot through and so on.

You can't have infinite inputs in social interactions, for the same reason you can't throw Pistol in a game of Rock-Paper-Scissors.

Factoid of the day: In the hand game Huntsman, Village Headmaster, Kitsune which predates RPS arriving on western shores, the symbol for huntsman is "holding up a rifle" to defeat the Kitsune which in turn throws up paw hands to bewitch the village headmaster via seduction and then the headmaster puts hands on knees like a grandpa figure to pull heirarchy on Huntsman.

[Orion]

I also don't think that in most HB games people talk about (at least not here but I've seen it in discords/message boards more focused on narrative games) rules where your dice rolls generate post hoc rationalizations for the mechanical effect they are set to achieve. Usually people aren't introducing the idea that a high roll for bribery 'means' that you get a guard that is now bribable. I think it's an interesting idea and K actually talked about it before here but it hasn't seemed to gain much traction.

Most DM already do this kind of thing in combat, even if they're not used to using it elsewhere. Like when narrating hits and misses. When I was young I used to look at the components of someone's AC and say "you dodge," "you block with your shield," or "it bounces off your armor" depending on the attack roll. But I think it's more common to just invoke new fictional details that don't correspond to any particular mechanic. "You're about to lunge when your back foot slips on a loose cobblestone and you lose your chance;" "You swing again and again but the bad guy deftly parries every blow" (even though he's not actually "fighting defensively").

I've been reading a lot of The Alexandrian's "Art of Rulings" essays and watching Matt Colville's DM advice and they both like to point out ways to use some of the mixed results, retroactive details, and other tricks lifted from the narrative/storytelling games movement to enrich D&D. In fact, I'm starting to think it's a actually necessary if you want to avoid what I used to call the Kangaroo Court problem.

[Stubbazubba]

I don't understand at all the desire to collapse social interaction into one roll instead of several, but then have potentially huge consequences to that one roll like selling drugs to an undercover cop. If this game is about street crime, then playing through the logistics of street crime, e.g. finding dealers/buyers you can never really trust, is just as important and tense as the actual meeting. If it's not about street crime, then having multiple discrete failure states that include "trigger an entirely new conflict between you and the authorities" seems like overkill. If something is going to have that kind of effect, it really needs a little bit more structure than one roll.

Kaelik is absolutely right that people don't change their mind on anything in an action. They can be nudged to believe something they're already contemplating, but no one has ever reversed an opinion that fast. They do sometimes act against their opinion that fast, though. I don't want to give strangers on the street money, but when a mugger pulls out a weapon, I act against what I want. Similarly, I can be deceived that something is what I want when in reality it's not; if I don't catch the deception I will happily give money to what I think is a real Nigerian prince. Similarly, I can be persuaded to take or refrain from taking an action in the moment that I wouldn't otherwise for a compelling reason. But never have I been talked into a new worldview I wasn't already predisposed to in one sitting.

Social checks should be about affecting what NPCs do in the moment, and learning what you need to know to do that well. A social system should let you gain influence with various factions/classes/groups and enable you to direct their power. You can't just whip up a rabble of peasants to overthrow the wicked Baron by standing on a soapbox for fifteen minutes, even if you are Talky McTalkface. Those seeds have to be planted before they can be harvested: his deeds have to be publicized, the people in charge of some group have to be convinced that he is opposed to their interests, small acts of rebellion have to be put down with disproportionate force, etc. And then you can incite a general revolt. Now, maybe you walk into that situation, a powder keg that just needs one final push. But sometimes you'll have to do the legwork yourself.

Social challenges are generally not something that happen within the time it takes to explore one room of a dungeon. It takes repeated interactions geared at creating the conditions for whatever change you are aiming for. Having a multitude of social skills and mechanical approaches is less important than being able to figure out the structure of the social framework you are trying to impact, allowing you to apply pressure in the right places to achieve what you want.

Even a relatively simple social challenge, like a small tribe of Goblins, has a structure that you can impact: the biggest, meanest Goblin is in charge and gets his way by beating/killing anyone who ignores him. The party that is captured can social their way out by sussing out who a potential rival is and making a deal with them; the rival lets the party out, and they'll help them take over the tribe by, e.g., ambushing the big Goblin in his sleep or by helping the rival Gob faction kill the big Gob in a violent takeover. Whatever.

Those structures need not be entirely predefined. MCs are level designers, they can put together a combat encounter, they can also put together social encounters. But having a list of off-the-shelf parts to put together, complete with ways of tracking progress (that may not be very visible to the players, but are to the MC) is completely doable, just like picking off-the-shelf monsters to put in your combat encounter.

The distinction between actions taken to discover this structure and actions taken to affect it is, I think, useful, as it allows the players to actually role-play what they would do once they see the playing field. If discovering the social structure and affecting it are one and the same action, then there's no room for players to approach the structure as their character would, they're just mashing social buttons until they either win or they get to a combat.

If you want to convince the king to form an alliance with the elves so that the elves won't form an alliance with the Dragonborn or whatever, you've got to figure out the king's decisionmaking process: who are his trusted advisors, and where do they stand on the issue? Are they persuadable? If yes, then how do we gain their trust? If no, how do we undermine them in court? Can we get an obstinate one removed or at least incapacitated while this decision is made? Can we become advisors? Who do we have to impress to do that? Once the king will listen to us, what are his concerns and how can we show the elf alliance is the best way to address them?

Answering those questions and taking those actions - some of which will be social actions, some not - is where individual creativity comes out, where the spectrum of personalities and aptitudes can be brought to bear and matter. Knowing that the king has three advisors he is closest to, e.g. the Grand Duchess, the Archbishop, and the Commanding General, enables you to inquire into their position on the issue. You learn that the Grand Duchess is persuadable, the Archbishop is for it, and the Commanding General is against it. Maybe you want to get in the good graces of the Archbishop, leverage your agreement there to gain an ally that can tell you more. Or maybe your party isn't church-approved and you think you might be better off feeling out the Grand Duchess' interests. Or maybe you're perfectly equipped for skullduggery and want to find some scandal on the General. Your abilities, interests, and previous connections will open up different options.

Now that sounds much more like the framework for an entire adventure, and that's a plausible criticism, but bottom line, my point is this: social actions should be used on characters to affect the actions they take in the moment (including spoken actions like sharing information, making a commitment, etc.), but social systems should just be about identifying the structures supporting social power and affecting them.

[Zinegata]

I don't understand at all the desire to collapse social interaction into one roll instead of several, but then have potentially huge consequences to that one roll like selling drugs to an undercover cop.

This is precisely because of what's been missing in Social Action discussion - there are basically no tools provided to help DMs create social encounters, much less an entire adventure based on social encounters. You're not wrong in thinking you have to make a framework for an entire adventure; indeed you have to look at the mechanics from this perspective in the first place.

To elaborate:

Combat isn't collapsed into one roll because the encounter components / challenges (primarily Monsters) have a variety of very clearly defined stats that must be overcome. In D&D, you need to reach the monster first via movement, then you roll-to-hit, then you roll damage if you succeed. It is not collapsed into one roll.

Thing is, this model could have been easily adapted into the Court Intrigue situation you proposed. Players first have to roll to "Discover" what motivates the various court officials (to-hit roll), and then if you succeed you roll to see how your argument impacted them (damage roll). Do enough "damage", and the court supports you by the strength of your combined arguments and research.

The thing is, having multiple rolls is more complicated than just a single roll. So the DM must either spend more time to properly stat out a Monster (give them AC, HP, etc), or they can pull out a pre-generated Monster from the Monster Manual.

By contrast, DMs don't have the "pull a pre-made monster out the Monster Manual" option for Social encounters. They almost always have to make one almost from scratch.

That's why social encounters are collapsed into a single roll. DMs don't have pre-made social encounters they can pull out of a book.

[MGuy]

I think the reason conversations around social mechanics don't focus too much on things that aren't messing with people's feelings, desires, and actions is because that it 'is' the big deal thing where socializing is concerned. It's the most powerful/useful result you can get from any interaction with any person.

This comment is just weird to me. It's like saying "People don't really talk much about Summon Monster or Fear because Dominate is the best spell effect." There is no reason that abilities which don't work by making arbitrary line-edits to NPC motivations should be relegated to being "minor," "utility," "flavor," or "non-combat" or whatever. They can have big effects that decide adventures. For example, you don't have to write an ability that lets you target a specific NPC and make them an ally to give out the ability to bring in NPC allies. You can write all kinds of abilities that let you introduce new generic NPCs to the story who are already interested in fighting for you. The obvious ways to go involve membership in gangs or secret societies or whatever but you could even have abilities like "a really good damsel-in-distress routine" where given a couple of minutes in a crowded public space you can attract the attention of 1d3 gallant strangers who will fight to defend you. There are a variety of ways social skills can make people less likely to fight against you or more likely to stick with it. I personally am a big fan of morale rules and would like to see morale checks become a regular part of combat again. I think there's a lot of room for social ops to sap an organization's morale before an encounter begins, and for codified uses of both intimidation and diplomacy-style skills to penalize enemy combat morale.

I'm only telling you my thoughts on why I think people dominate the discussion about it. What I meant there would be closer to saying "People are way more worried about rocket launcher tag in combat". Socializing isn't really exactly like combat and so there are other considerations. I'm not going to remake the post I already made where I talk about my thoughts on the differences there. You can find it in my thread. But the thing that seems to bother people the most, is influencing people in social interactions. Cowing people into submission, lying to people, etc are also THAT. So it's not as if I'm saying 'diplomacy is the only thing people tend to care about' but the result 'get people to do/believe what you want' is the thing people worry the most and why it gets more airtime. Compared to that result distracting, bribing, passing on a forgery does not seem to bother people as much.

It's also the same thing for people not talking about post hoc rationalizations. I'm not complaining about the mechanic mind. It's just why I think it hasn't gotten talked about. I'm not opposed to talking about it. I've read the Alexandrian myself which led me to take a look at Blades.

[Ice9]

I think that if you are going to do effect-based social rules, then the kind of abstraction proposed in the OP is the least jarring way to do it.

Like, "It's easier to convince a rich guy to cut out his own liver than to give you 5 gp (because in a setting without organ transplants, that liver is worth very little)" is so stupid it throws me out of the flow. But "It's easier to find someone who'll give you a questionable piece of meat than someone who'll give you 5 gp"? Sure, plausible enough. The abstraction is key.


As for hard-persuasion, make the roll and they're your slave? People want to have this ability, but nobody wants to be on the other side of it, including most GMs.

Oh sure, the GM has infinite NPCs and blah blah blah no that's dumb it sucks balls when every single one of them gets mind-fucked without fail. You notice how BBEG's tend to have high-ass saves where SoDs are unlikely to work? Yeah. The GM can more happily tolerate being steamrolled but there's still a limit.

So if you're going to use it, treat it like Dominate Person:
* Most people can't do this. It's not just a property of having high skill bonuses, you need a specific feat or class or something.
* Serious opposition has a decent chance to resist, and they do this by having a universal property like Will saves, not a particular skill they may or may not have.
* People can tell you did it and now maybe they're scared of you (or want to kill you) because it's fucking creepy.
* There are limits to it and things that block or undo it.

And indeed, "doesn't work on PCs" is a warning sign at the least.

[Orion]

I've been playing around with setting DCs for coercion-based rolls, and I'm not sure if this is a good idea or not.
I just said "fuck it" and made the Base DC for getting people to do what you want dependent on the objective outcome of the roll instead of its perceived outcome.

I don't think this is a good idea. I actually think it has two pretty bad failure points. The first problem is that there's going to be a blatant lack of psychological realism. Players will have social intuitions about what humans can and cannot be easily talked into and the outputs of your ruleset won't match up with those intuitions much at all. The second problem is that it will be unbalanced and exploitable. No matter how you structure your "objective outcome" tables there will be times when the context of a specific situation makes a low-DC outcome much more powerful and important than it was intended to be, creating game-warping loopholes. The only way around this is to make subjective ad hoc DC adjustments to prevent "exploits," which is basically just going full MTP with extra steps. Or to put it another way, PhoneLobster is basically right when he points out that being able to influence a higher-level character's behavior in any way, even a seemingly trivial one, will end up being very powerful because it creates exploitable information and enables Rube Goldberg schemes. I believe that in a level-based game coercion abilities need to be level-based, and if you have low-level people with significant wealth and power, they need to get special coercion resistance along with it. You should not be able to make Elminster do anything he doesn't already want to do.

For example, you give a homeless guy a suitcase full of cash and tell him it's totally not hot and people definitely are not looking for it. If that's actually true and this random guy isn't in much danger, then the Base DC is 3 and you shit out modifiers on top of that. If you're lying and there's a band of goons coming around the corner and they'll fucking kill whoever has the suitcase, then that's fucking suicidal and the Base DC for the check becomes 12 instead (a magnitude or two harder to achieve).
On one hand, this seems kind of weird and metagamey. On the other hand, it's effect-based.
Thoughts?

I don't understand why you would model this as a coercion situation in the first place. We don't actually care about the guy, right? He's just some guy. What's probably going on is that the PCs are being chased or tailed by cops or detectives or enemy spies and they need to lose the package before they get caught with it. What should probably happen here is that the players should be making an opposed roll against the people pursuing them, something like Deception or Sleight of Hand or Spycraft or Stealth, vs. Investigation or Hunting or Perception or Spycraft. They can specify their basic strategy before they roll, like "convince some random guy to hold onto it" or "find an out-of-the-way-nook to stash it in." If their roll succeeds, then that means they found a random guy who was willing to hold onto it. If the roll fails, they didn't find one. At no point does anyone need to be coerced. Indeed, attempting to coerce someone is potentially an auto-fail if it would create a loud or memorable scene that witnesses will be able to describe.

[deaddmwalking]

I think Orion's idea has legs. Instead of trying to convince an individual guard to betray his master, you're trying to find the guard that is willing to. There are a couple of places where I see a potential issue.

The first is when reasonably only a single person knows something (like the Lich's True Name or something). If you make the DC 50 Obscurity check, the DM has to figure out how something that was thought to be impossible isn't - but this is a minor thing. There could be some clues and he used an Anagram (like Voldemort was SUPPOSED to have done, but close enough). Explaining how the effect was achieved after you know the result shouldn't be too difficult, but it does put the GM on the spot.

The second potential issue is that there are times when you actually do want a specific individual to take an action. If the PCs have obtained a prisoner, for instance, they don't want to go make a gather information check to find someone that will tell them what they want to know - THEY HAVE SOMEONE and they need to convince that specific individual to talk. Some amount of intimidation, bribery, or kindness might be able to shift the alliance. Times when the PCs spared a goblin and he opts to devote himself to his new, kind masters are fun (if done sparingly). It seems that Orion's proposal is set up to find/obtain what you want, but doesn't seem to work quite as well if you found it but still need to obtain it.

[Harshax]

I’m still catching up ... what I was trying to do with the simple rules I posted yesterday was propose a framework for players to have agency in doing a social interaction and then flip the rules to the GM’s perspective and show how those same rules can set up encounters and fair TNs for players to overcome.

I’ll catch up later today or tomorrow.

[Orion]

I don't understand at all the desire to collapse social interaction into one roll instead of several

On reflection the actual number of rolls is not the main point. What I should be doing is advocating for what The Alexandrian calls "fortune-at-the-beginning" resolution for non-coercive social actions. Basically, in "fortune at the end" gameplay, you start with a detailed description of the fictional situation and what the character is doing, in which there remains some uncertainty about the outcome. Then you roll a die to find out what the outcome is. In "fortune at the beginning" gameplay you start with a vague sketch of a fictional situation and a player stating a desired outcome. You roll a die to determine the outcome first, and then you flesh out the fictional situation by inventing the details that caused the outcome and narrating exactly how the character acted on those opportunities.

The advantage of using fortune-at-the-beginning here is that we end up being able to kill off a lot of the modifiers that tend to populate the big lists of modifiers that people tend to right. Take the classic "racist against elves*" tag some NPCs have that gives elves a -3 to social skill rolls against them. To use that modifier we have to play fortune-at-the-end, start by working up a psychological profile of every clerk and guard the elf Bard interact with. We don't want to do that because it's tedious and because the player might get annoyed at us for dropping in too many racists. If we use fortune-at-the-beginning for non-coercive social interactions with unnamed NPCs, we don't need modifiers; we can wait for the elf to fail a Charisma check and then decide that it's because they're dealing with a racist.**

For non-coercive interactions with named NPCs, we use "fortune never." Either the DM just decides what the NPC does by fiat, or the NPC is governed by a more complex gameplay structure like a king's council set-up where his decisions are determined by a majority of the advisors. In the first case, the DM arbitrarily decides how racist he is and how he acts on it. In the latter case, the constraints of council politics outweigh whatever personal racism the king may or may not hold.

In both cases, coercive abilities also get to the skip the racism modifiers because being an elf-hater doesn't make it more difficult for elves to Hypnotise, Enthrall, or Dominate you. (Indeed the fear of elf-hypnosis keeps a lot of elf-haters awake at night)

*Treat this as a stand-in for all the personality trait and mood based modifiers.

**You might give a city "racist against elves" modifier that penalizes elves' social rolls due the risk of running into bigots, but that would be something you could put at the top of the location handout and everyone would know when it was or wasn't in play.

[The Adventurer's Almanac]

I don't think this is a good idea. I actually think it has two pretty bad failure points... I believe that in a level-based game coercion abilities need to be level-based, and if you have low-level people with significant wealth and power, they need to get special coercion resistance along with it. You should not be able to make Elminster do anything he doesn't already want to do.

Yeah, I ditched the idea pretty quickly. Good thing I didn't build anything off of it. I agree with you on this next point, though - charming people into doing anything you want is rather suspect. Charming people into wanting to help you, and THEN they decide to do what you want seems more reasonable and realistic, but perhaps I'm deluding myself.

What's probably going on is that the PCs are being chased or tailed by cops or detectives or enemy spies and they need to lose the package before they get caught with it. What should probably happen here is that the players should be making an opposed roll against the people pursuing them, something like Deception or Sleight of Hand or Spycraft or Stealth, vs. Investigation or Hunting or Perception or Spycraft. They can specify their basic strategy before they roll, like "convince some random guy to hold onto it" or "find an out-of-the-way-nook to stash it in." If their roll succeeds, then that means they found a random guy who was willing to hold onto it.

You know what's funny? I've actually got a way to find stashed goods written up... but no mechanical way of actually stashing shit. That's amusing and probably emblematic of my system at the moment.

Basically, in "fortune at the end" gameplay, you start with a detailed description of the fictional situation and what the character is doing, in which there remains some uncertainty about the outcome. Then you roll a die to find out what the outcome is. In "fortune at the beginning" gameplay you start with a vague sketch of a fictional situation and a player stating a desired outcome. You roll a die to determine the outcome first, and then you flesh out the fictional situation by inventing the details that caused the outcome and narrating exactly how the character acted on those opportunities.

Alright, now how does this interact with the scenario that Dead brought up? The party decides "we want this guy to tell us what he knows", they roll to determine how successful that is, and then they RP what happens based on if they passed or failed? And details about this particular NPC may or may not come into existence based on this roll?
Now that I actually type that out, the two don't sound that different in practice. Am I missing something critical here? I like where this is heading and want to make sure I actually fucking understand something that somebody posts for once.

[Orion]

Sure, in fortune-at-the-end you would start by describing the prisoner's demeanor and you'd set an interrogation DC. You might decide the prisoner is scared and open to talking and make it DC 5. You might decide the prisoner is brave and loyal and make it DC 15. You might decide he's loyal but not especially brave and make it DC 10. The player would describe the question they ask and then you'd roll a die to find out whether the prisoner answers them. The random element here is diegetic (exists in the fiction); the prisoner is a specific guy who might or might not feel like answering your question.

If you go for fortune-at-the-beginning you could instead reason that the bandit lord probably has some loyalists who would die for him and some opportunists who would eagerly sell him out. If they players captured a dozen bandits you wouldn't even make them roll, you'd just assume someone talks. In this case, though, they only got one bandit. You assign the bandits as a group a "Loyalty DC 10". The randomness here is not diagetic; they either got a guy who wants to talk or a guy who doesn't but we're rolling to find out which is which. Let's say that the PC in question has a +5 Intimidation. If they roll a 1-4 you tell them "bad news, this guy is ride-or-die for the bandit king. You pull out all your tricks but he's prepared to take his secrets to the grave." If they roll a 10 or higher you say "he's eager to talk." If they roll a 5-9 you say "he's reluctant at first but you wear down his resolve and eventually he opens up."

Fortunate-at-the-beginning is a pretty flexible technique that you can do a lot of different things with, this is just one example. Some people like using a die roll to determine what their PC actually does. Like instead of delivering a stirring and eloquent speech and then rolling a die to find out if the audience buys in, you could roll a die to find out whether your PC actually gave a stirring speech or whether they got flustered and then committed a social faux pas. Basically you just want to think carefully about what accounts for the random variance in outcomes. Is it inside or outside the fiction? Is the effort the PC puts forth random, is the effect on the recipient random, or is the identity of the recipient random?

EDIT: I've mentioned The Alexandrian's "art of rulings" series a few times so here's a link to his social skills article. You may want to start at the beginning though, the whole series is good. https://thealexandrian.net/wordpress/43 ... ial-skills

[Orion]

Yeah, I ditched the idea pretty quickly. Good thing I didn't build anything off of it. I agree with you on this next point, though - charming people into doing anything you want is rather suspect. Charming people into wanting to help you, and THEN they decide to do what you want seems more reasonable and realistic, but perhaps I'm deluding myself.

More realistic, perhaps, but still something you want to be vigilant about. The actual Charm Person spells in D&D has at minimum the following constraints:[*]It can be resisted using a defense that is (supposed to) scale up with level[*]If you try it and it doesn't work, the target knows you try to do something to them.[*]If you do it in front a third party, the third party probably knows you did something to your victim.[*]People in general will think ill of you if they suspect you of doing it to people.[*](In some editions/interpretations)It will eventually wear off and the target may feel used afterwards.

If your D&D-style game has social skills that can replicate the effect of Charm Person, then they should also operate under the same constraints as Charm Person, or you will quickly no longer have a D&D-like game. Making anyone you want into your friend is one of those abilities that is so powerful that if you put it into a game, that's now what you game is about.

[Stubbazubba]

I've mentioned The Alexandrian's "art of rulings" series a few times so here's a link to his social skills article. You may want to start at the beginning though, the whole series is good. https://thealexandrian.net/wordpress/43 ... ial-skills

This is a good article but it didn't explain fortune-in-the-beginning or -end. That article is here.

And in fact, the example the first article uses in discussing fortune-in-the-middle...isn't actually that at all, it's a single check that fails and then Erik throws the ballistics back at the ships, which is what he was about to do anyway.

I'm wary of fortune-in-the-beginning as a model for a roleplaying game. The choice of approach is as central or even more central to the concept of roleplaying as the choice of desired outcome. The approach you use may very well have consequences beyond success/failure of your immediate goal, e.g. if you persuade the elven captain to release the prisoners of war vs. threaten her to do so, your relationship with her going forward is different, even though in both cases she releases the prisoners. If how high you roll determines success/failure, and then you induct the approach afterwards, the character's ideals and motivations are taking a backseat to an improv prompt from an RNG.

There are some actions where the approach truly won't matter, but I just don't see a compelling reason to have an entirely different mechanical order of operations. "Ascertain intent and approach, roll, narrate outcome" is super effective at centering player choice, the key to an RPG.

TBF, that's not what you describe. You describe fortune-in-the-beginning here as entirely on the DM side of the screen; the PCs' approach and intent is decided by the players, the difficulty is not based on the individual target NPC but based on something else (could be based on the group, could be universal difficulties, etc.), the roll is made, and then the narration justifies the result by inventing specific characteristics of the target NPC. From the player's perspective, this is identical to fortune-in-the-end.

Some people like using a die roll to determine what their PC actually does. Like instead of delivering a stirring and eloquent speech and then rolling a die to find out if the audience buys in, you could roll a die to find out whether your PC actually gave a stirring speech or whether they got flustered and then committed a social faux pas.

This, again, is not what Justin Alexander describes as fortune-in-the-beginning. The PC decided to try an eloquent speech to accomplish whatever intent they had, rolled a failure, and then narrated a failed speech. The approach (and an assumed intent) was set before the roll, so this is fortune-in-the-end.

JA's point about not describing the speech before the roll is not about fortune positioning, it's about where description comes in the mechanical order of operations. In bog standard fortune-in-the-end checks, the roll should still come before the description. This is often not what happens in social situations, and he insists (rightly) that it should be corrected. But that is not fortune positioning.

[Orion]

On reflection, you're right. What I'm talking about doesn't actually line up very well with Alexander's model. Maybe I get to coin a new phrase actually? Let's talk about the Locus of Uncertainty. If you're making a die roll to resolve something, you must be uncertain about the outcome, but why are you uncertain? What part of the process are you uncertain about? The fictional element or elements you are randomizing is/are the Locus/Loci of Uncertainty. This is going to be fuzzy but I like dividing up the possible loci into basically three types. You can have uncertainty of action where the question is "can the character execute the intended approach?" You can have uncertainty of impact, where the question is "assuming they execute their action correctly, will it in fact have the desired effect?" And you can have uncertainty of circumstances, where the question is "will as-yet-undescribed circumstances obstruct or facilitate this approach?"

Let's take an Elf Bard and look at examples of how each approach to uncertainty could apply to Perform, Jump, or Archery. Warning: some of these examples will be very weird things that you wouldn't want to do in an actual game, or that would only make sense in a very weird game. That's what comes of forcing examples to fit a schematic.

Perform Examples

Locus: Actor/Execution. The bard has impersonated an understundy to sneak around backstage at the opera house but then a freak accident occurs and she has to perform. She's heard this aria once before and had 15 minutes to to look over the sheet music. Can she sing it? A high roll hits all the notes, a medium roll muddles through all the lines, a low roll cracks under pressure.

Locus: Impact/Target. The bard is performing some songs from her new album for concert hall packed with her die-hard fans. She's warmed up and knows these songs cold so she's basically guaranteed to perform them correctly. The issue is that we don't know what kind of mood the crowd is in and whether they'll like the new stuff as much as the old stuff. A high roll indicates that they're in a good mood and they warm to the materil, a low roll indicates that someone shouts "play the hits."

Locus: Circumstantial. The bard walks into a random bar in a strange city and belts out a country ballad. Once again we can assume that her rendition is good, but this time we have no idea who's even in the bar or how receptive they'll be. A high roll indicates that the bar is full of country boys staring silently into their beer. A low roll indicates that it's full of metalheads busy brawling with Nazi metalheads.

Jump Examples

Locus: Actor/Execution. There is a 15-foot-wide chasm. The bard wants to jump across it, but in this system her long jump distance is randomized every time. She rolls to find out whether she makes it across the 15-foot gap.

Locus: Impact/target. We are playing a game system where each character has a fixed long jump distance that never changes. There is a 10-foot wide chasm with loose and crumbly soil near the edge on the far side. Player characters in this system always have at least a 10 foot long jump, so any PC can jump the gap, but they have to roll to find out if they land on solid ground or in the crumbling soil. You add your jump distance as a bonus to the roll because the ground is firmer further away from the edge.

Locus: Circumstantial. Again we're in a fixed jump distance system where the bard know she can always jump exactly 15 feet. Some spells or monster ability has just created cracks in the earth splitting a building in two, but the description doesn't specify how big the cracks are. The DM decides to roll a d20 and let that be the width of the chasm. If the bard were temporarily blinded or something, then the DM might require the player to commit to attempting the jump before the DM determines how big the gap actually is, creating a pseudo-jump-check where what's being randomized is the size of the gap instead of the distance traveled.

Longbow Attack Rolls

Locus: Actor/Execution. Do you successfully nock, draw, and loose the arrow so that it goes into the target you want it to go into?

Locus: Impact/Target. Does the target block with a shield, or dodge? Does the arrowhead actually penetrate their padding?

Locus: Circumstantial. Does a sudden gust of wind spoil your shot? Was the fletching on this arrow damaged when you fell down that pit trap earlier? Do you notice that your bowstring is frayed and you have to make some kind of spot-fix? We can imagine weird story games where like you staked out a spot to snipe the evil baron with an arrow and we roll a die to find out whether he's wearing armor under his shirt.

I'm basically proposing that a circumstantial locus of uncertainty is a really good fit for a lot of social actions.

[PhoneLobster]

While I entirely agree that describing a success on a success, and describing a failure on a failure is a thing, and a thing you should do.

When did it deserved this much description and discussion? This is bog standard RPG 101 and has been for decades.

OK, so maybe some people need a reminder on it for the social rules, because they think they should describe in advance and if they describe a success big enough should be given a mechanical success for that, but I don't think it's the basic concept of how things are normally described in every other action in RPGs that those people are missing.

They believe, or claim to believe, that social rules are an exception to that, and pretty much all the other, normal elements of design, balance, and at the table execution.

[Ice9]

Depends on what you're comparing it to. If a social action is like "an attack roll" then you roll and you get what you get.

But on the other hand, if a social action is like "Get inside that palisade which is guarded by Gnoll mercenaries", then the approach does change the difficulty and even what checks you're rolling:
* Teleport Inside - no roll needed, but you will be arriving blind
* Fly over the Wall - no roll needed, but you'll get shot at on the way in
* Climb over the Wall - climb check, not that difficult except you're being shot at.
* Disguise as Mercenaries - disguise and bluff checks, may or may not even be possible depending on whether reinforcements are plausible.
* Sneak in with the Food Shipment - stealth, and either more stealth or bribery or something else to get inside the cart.
* Bust Through the Wall - strength check, very difficult, or doing a large amount of damage.
* Tunnel under the Wall - probably no check but takes a long time, might be discovered.
* Fight the Entire Mercenary Company - combat, so attack rolls and such.

And so forth. So "different difficulty (up to and including auto-success and auto-failure) depending on your approach" isn't necessarily crazy-town.
And yes, making a social action be like an entire area means a fair amount of work. Very overkill for "get a 5% discount on these potions", but reasonable enough for "convince the king to declare war".