Delve design challenge: non-mechanical interrogations
Fictional example of play, drawn from friends' accounts, vague memories of years past, and stuff that's occasionally come up in less egregious fashion in Delve:
The village elder says old man Donald's been acting strange. PCs Vandr, Mar and Mengot seek out Donald and have a quick conversation with him where he claims he knows nothing, but is shifty-eyed and suspicious. The PCs go one to do other stuff, and Donald's name repeatedly comes up. The players have various theories on what Donald's role could be, but they all agree he's important, and they should question him harder. Vandr in particular thinks Donald is up to no good, and declares he'll yell at him and even deliver a beating if he has to.
Mengot offers to guard the door. Mar doesn't want to be part of beating up an old man, so he waits with Mengot. Vandr goes in and yells at Donald menacingly. Donald seems scared but repeats his earlier claim that he knows nothing. Vandr yells some more. Donald repeats himself some more. Vandr resorts to his plan of a beating. Donald takes the beating and cries.
Vandr's player is at his wits' end. He's sure Donald has info, but doesn't know what else to try to get it out of him. Torture would go against his character vision, plus likely erode the goodwill of the villagers, and he has no particular reason to think it would work.
The GM is also at his wits' end. He knows Donald will cooperate if offered a chance to avenge his daughter, but he didn't plan any other options, and has no idea if any of the chatting, yelling, and beating that's been roleplayed thus far ought to work.
Here are the only good parts:
1) The first paragraph. Get info -> form theory -> form plan = perfect.
2) Vandr goes in and yells at Donald menacingly. Following the plan = perfect.
3) Torture would go against his character vision, plus likely erode the goodwill of the villagers. At least Vandr's player is playing Delve in the proper spirit! Playing a fairly regular person means having fairly normal morals and empathizing with other regular people. Treating the world as a real place means thinking about consequences beyond the immediate.
4) He knows Donald will cooperate if offered a chance to avenge his daughter. Good option to have.
The rest is awful. Here are the fixes:
Mengot offers to guard the door. Mar doesn't want to be part of beating up an old man, so he waits with Mengot.
No no no. Limiting the characters present means limiting the players who can contribute means limiting your resources to interpret Donald and try new strategies. As with every challenge in Delve, three heads are better than one. Guard the door from inside the damn cabin.
Rule: if you split the party, go into Summary mode. Don't split the party to do anything you want to play through in Real-Time. If it's important, it's worth having all the players' brains in on it. The one allowed Real-Time split is if the characters are synchronizing time-pressured, challenging tasks.
Donald seems scared but repeats his earlier claim that he knows nothing.
There's no new information there! That's a "nothing happens" result! Bad bad bad! Give the players something! Further, give them something that matters! "Donald stares you down; this old guy is tougher than he looks," would be okay. "Donald clenches his fist and snarls, 'After what I've been through, you can't scare me'," is better. They both give new data to inform new strategies, but the latter is a hint toward something the GM has already determined would work (angry + torn up over daughter = revenge!).
Rule: never say "nothing happens" when the players try something new. Always give new feedback. The one exception is if the players try random stuff with no goal in mind -- in that case, don't give feedback, just demand that they pick a purpose or do Summary mode.
Guideline: GM, all your problems want to be solved. They want to help the players solve them. Something is preventing them from doing this. Have them indicate what's blocking them, and suggest ways around it.
Guideline: GM, drop hints and clues like crazy. Basically a different phrasing of the above.
Example: the greedy NPC. "Gee, I want to help you guys, but I need to look for another odd job because money's tight right now." He's telling the players the obstacle to his cooperation (priority is money) and suggesting a way around it (get him money). It should now be on the PCs' radar that they could either bribe this guy or use their clout to get him work. Of course, they might rather admonish him for being greedy and pull out their swords. That would be a judgment that a few pence isn't worth getting hurt over, and the GM ought to agree unless there's a specific reason otherwise.
Vandr yells some more.
Now the player is being as dumb as the GM. Try something new. This is also easier and more fun with the whole party.
Guideline: players, if a first attempt fails, try something new. If you can, base it on the new info you just received. If you have an idea, communicate it in character, even if another character is taking point on this task.
Vandr resorts to his plan of a beating.
This could be good or bad depending on whether it follows or ignores the previous guideline.
Donald takes the beating and cries.
This is technically new, but it's exactly what would be expected of a normal person in this circumstance; thus the "newness" fails to qualify as new information. And, of course, there isn't any useful, actionable info here either.
Guideline: expected outcomes that change nothing don't count as "new feedback".
Vandr's player is at his wits' end. He's sure Donald has info, but doesn't know what else to try to get it out of him.
If this ever comes up, the first person who notices it needs to grab the pacing dial instantly. "Okay, guys, we're not gonna waste play time spinning our wheels. Declare what Vandr does until he gives up, so we can go deal with one of our other leads instead. Maybe we'll learn something elsewhere that'll help us with Donald. In fact, I'd say that Vandr should leave Donald in a shape where we could try again later."
If the experience is simply boring, Vandr's player should be the first to notice, "Hey, I'm bored," and grab the dial. Waiting for someone else to notice he's bored is a punk move.
If the experience is frustrating, and Vandr's player is focused on the info he's not getting, someone else will probably notice first. If all the players are in the same boat, that leaves the GM. So, GM, when the players are spinning their wheels, and not latching onto what you give them, and you don't have any more hints you can instantly drop -- end the damn scene! Either crank the dial to Summary or interrupt the characters with the next of your Steps of Doom.
Rule: if you ever get bored, grab the pacing dial and speed up. You are responsible for not letting yourself get bored.
Rule: if you ever get truly stuck on something, give up and try something else. Vow to come back later if that helps.
Rule: if you ever see another player get truly stuck on something, either go to Summary mode or interrupt them with new information.
Torture would go against his character vision . . . and he has no particular reason to think it would work.
Would torture work? You'd have to discuss it as a group to establish the characters' best guess. The answer might turn out to be, "You have no idea," but don't assume that prematurely if it's something you're really interested in (not that this example qualifies).
Guideline: when you don't know something, but your character might, ask!
The GM is also at his wits' end. He knows Donald will cooperate if offered a chance to avenge his daughter, but he didn't plan any other options, and has no idea if any of the chatting, yelling, and beating that's been roleplayed thus far ought to work.
Not planning other options is fine. Not having other options is generally not -- it means you're insufficiently prepared to improv.
Also, sweating whether a given technique "ought" to work is the wrong mindset. You should go with what feels right to you, both socially and in the fiction, erring on the side of giving the players what they want, limited only by what's plausible.
There's no formula for "what feels right". It's a product of taste in sights, sounds, characters and stories; it's a response to which of your friends seem to be digging what, right now; it's an instinct about where we are in the adventure, near the slow beginning or the frantic climax; it's a nod to whether you need to break or wrap up soon. So don't over-analyze, don't try to be perfect; just try to have fun.
Rule: before you decide what to do, get all the information you need from the fiction to make that decision. For players, this means asking what your character sees, hears, or (occasionally) knows. For the GM, this means forming an understanding of the feature of the world you are playing, whether it's an NPC, a magic item, or a creaky manor. You can do this in prep or as ad-lib right before you narrate, depending on what works best for you. If you need a moment to ad-lib, you can stall by taking a bathroom break, flipping through your notes, or asking the players questions about their current position and direction ("You moving more quickly or quietly? You think you'd be able to hold onto that torch while using your shield in combat?"). You can even ask them questions about their expectations for the feature you're conceiving! "This guy just lost his daughter; you really think he'll be in the mood to help?"
Rule: GM, don't waste time on accuracy; plausibility is enough. Precision is great, but it's not worth stalling the game or stressing out over. Remember, the whole group is looking out for the gameworld's integrity, and you're not expected to be perfect. Being corrected is fine.
Rules: As part of your adventure scenario prep, define 7 ORDs. Define the Obstacle to the players getting the reward; the Reward itself, even if that's just the form the directions are delivered in; and one or more Directions, reasons for the players to go encounter other ORDs.
Develop each ORD's identity to the point where you know it well enough to play it naturally -- your NPCs' names, wants, demeanors; your puzzles' pieces and properties; your forts' entrances, guards and traps.
Write down at least one way to overcome each obstacle.
Record your ORDs on the scenario flow chart. Imagine this is a game that has already been played. Your goal is not to reproduce that game, but to use it as inspiration to jointly craft a new game with the players. In play, don't even look at your flow chart unless you have to.
Guideline: in play, improv more Directions onto your ORDs. Do this most early in play; late in an adventure, when the players have already parsed a lot of info and formed some plans, singular Directions are fine. When you do improv new Directions, they can point to other prepped ORDs or to new ones you invent on the fly. Don't invent more than you can handle playing, though! If you feel the urge to reward the players for something other than overcoming an Obstacle, improv a "freebie" Direction. I'll often do this for good teamwork on engaging, colorful character portrayals that, while not plausibly addressing an NPC's Obstacle, seem like they ought to get some compliance.
Guideline: vary your NPCs. See the templates for inspiration.
The village elder says old man Donald's been acting strange. PCs Vandr, Mar and Mengot seek out Donald and have a quick conversation with him where he claims he knows nothing, but is shifty-eyed and suspicious. The PCs go one to do other stuff, and Donald's name repeatedly comes up. The players have various theories on what Donald's role could be, but they all agree he's important, and they should question him harder. Vandr in particular thinks Donald is up to no good, and declares he'll yell at him and even deliver a beating if he has to.
Mengot offers to guard the door. Mar doesn't want to be part of beating up an old man, so he waits with Mengot. Vandr goes in and yells at Donald menacingly. Donald seems scared but repeats his earlier claim that he knows nothing. Vandr yells some more. Donald repeats himself some more. Vandr resorts to his plan of a beating. Donald takes the beating and cries.
Vandr's player is at his wits' end. He's sure Donald has info, but doesn't know what else to try to get it out of him. Torture would go against his character vision, plus likely erode the goodwill of the villagers, and he has no particular reason to think it would work.
The GM is also at his wits' end. He knows Donald will cooperate if offered a chance to avenge his daughter, but he didn't plan any other options, and has no idea if any of the chatting, yelling, and beating that's been roleplayed thus far ought to work.
Here are the only good parts:
1) The first paragraph. Get info -> form theory -> form plan = perfect.
2) Vandr goes in and yells at Donald menacingly. Following the plan = perfect.
3) Torture would go against his character vision, plus likely erode the goodwill of the villagers. At least Vandr's player is playing Delve in the proper spirit! Playing a fairly regular person means having fairly normal morals and empathizing with other regular people. Treating the world as a real place means thinking about consequences beyond the immediate.
4) He knows Donald will cooperate if offered a chance to avenge his daughter. Good option to have.
The rest is awful. Here are the fixes:
Mengot offers to guard the door. Mar doesn't want to be part of beating up an old man, so he waits with Mengot.
No no no. Limiting the characters present means limiting the players who can contribute means limiting your resources to interpret Donald and try new strategies. As with every challenge in Delve, three heads are better than one. Guard the door from inside the damn cabin.
Rule: if you split the party, go into Summary mode. Don't split the party to do anything you want to play through in Real-Time. If it's important, it's worth having all the players' brains in on it. The one allowed Real-Time split is if the characters are synchronizing time-pressured, challenging tasks.
Donald seems scared but repeats his earlier claim that he knows nothing.
There's no new information there! That's a "nothing happens" result! Bad bad bad! Give the players something! Further, give them something that matters! "Donald stares you down; this old guy is tougher than he looks," would be okay. "Donald clenches his fist and snarls, 'After what I've been through, you can't scare me'," is better. They both give new data to inform new strategies, but the latter is a hint toward something the GM has already determined would work (angry + torn up over daughter = revenge!).
Rule: never say "nothing happens" when the players try something new. Always give new feedback. The one exception is if the players try random stuff with no goal in mind -- in that case, don't give feedback, just demand that they pick a purpose or do Summary mode.
Guideline: GM, all your problems want to be solved. They want to help the players solve them. Something is preventing them from doing this. Have them indicate what's blocking them, and suggest ways around it.
Guideline: GM, drop hints and clues like crazy. Basically a different phrasing of the above.
Example: the greedy NPC. "Gee, I want to help you guys, but I need to look for another odd job because money's tight right now." He's telling the players the obstacle to his cooperation (priority is money) and suggesting a way around it (get him money). It should now be on the PCs' radar that they could either bribe this guy or use their clout to get him work. Of course, they might rather admonish him for being greedy and pull out their swords. That would be a judgment that a few pence isn't worth getting hurt over, and the GM ought to agree unless there's a specific reason otherwise.
Vandr yells some more.
Now the player is being as dumb as the GM. Try something new. This is also easier and more fun with the whole party.
Guideline: players, if a first attempt fails, try something new. If you can, base it on the new info you just received. If you have an idea, communicate it in character, even if another character is taking point on this task.
Vandr resorts to his plan of a beating.
This could be good or bad depending on whether it follows or ignores the previous guideline.
Donald takes the beating and cries.
This is technically new, but it's exactly what would be expected of a normal person in this circumstance; thus the "newness" fails to qualify as new information. And, of course, there isn't any useful, actionable info here either.
Guideline: expected outcomes that change nothing don't count as "new feedback".
Vandr's player is at his wits' end. He's sure Donald has info, but doesn't know what else to try to get it out of him.
If this ever comes up, the first person who notices it needs to grab the pacing dial instantly. "Okay, guys, we're not gonna waste play time spinning our wheels. Declare what Vandr does until he gives up, so we can go deal with one of our other leads instead. Maybe we'll learn something elsewhere that'll help us with Donald. In fact, I'd say that Vandr should leave Donald in a shape where we could try again later."
If the experience is simply boring, Vandr's player should be the first to notice, "Hey, I'm bored," and grab the dial. Waiting for someone else to notice he's bored is a punk move.
If the experience is frustrating, and Vandr's player is focused on the info he's not getting, someone else will probably notice first. If all the players are in the same boat, that leaves the GM. So, GM, when the players are spinning their wheels, and not latching onto what you give them, and you don't have any more hints you can instantly drop -- end the damn scene! Either crank the dial to Summary or interrupt the characters with the next of your Steps of Doom.
Rule: if you ever get bored, grab the pacing dial and speed up. You are responsible for not letting yourself get bored.
Rule: if you ever get truly stuck on something, give up and try something else. Vow to come back later if that helps.
Rule: if you ever see another player get truly stuck on something, either go to Summary mode or interrupt them with new information.
Torture would go against his character vision . . . and he has no particular reason to think it would work.
Would torture work? You'd have to discuss it as a group to establish the characters' best guess. The answer might turn out to be, "You have no idea," but don't assume that prematurely if it's something you're really interested in (not that this example qualifies).
Guideline: when you don't know something, but your character might, ask!
The GM is also at his wits' end. He knows Donald will cooperate if offered a chance to avenge his daughter, but he didn't plan any other options, and has no idea if any of the chatting, yelling, and beating that's been roleplayed thus far ought to work.
Not planning other options is fine. Not having other options is generally not -- it means you're insufficiently prepared to improv.
Also, sweating whether a given technique "ought" to work is the wrong mindset. You should go with what feels right to you, both socially and in the fiction, erring on the side of giving the players what they want, limited only by what's plausible.
There's no formula for "what feels right". It's a product of taste in sights, sounds, characters and stories; it's a response to which of your friends seem to be digging what, right now; it's an instinct about where we are in the adventure, near the slow beginning or the frantic climax; it's a nod to whether you need to break or wrap up soon. So don't over-analyze, don't try to be perfect; just try to have fun.
Rule: before you decide what to do, get all the information you need from the fiction to make that decision. For players, this means asking what your character sees, hears, or (occasionally) knows. For the GM, this means forming an understanding of the feature of the world you are playing, whether it's an NPC, a magic item, or a creaky manor. You can do this in prep or as ad-lib right before you narrate, depending on what works best for you. If you need a moment to ad-lib, you can stall by taking a bathroom break, flipping through your notes, or asking the players questions about their current position and direction ("You moving more quickly or quietly? You think you'd be able to hold onto that torch while using your shield in combat?"). You can even ask them questions about their expectations for the feature you're conceiving! "This guy just lost his daughter; you really think he'll be in the mood to help?"
Rule: GM, don't waste time on accuracy; plausibility is enough. Precision is great, but it's not worth stalling the game or stressing out over. Remember, the whole group is looking out for the gameworld's integrity, and you're not expected to be perfect. Being corrected is fine.
Rules: As part of your adventure scenario prep, define 7 ORDs. Define the Obstacle to the players getting the reward; the Reward itself, even if that's just the form the directions are delivered in; and one or more Directions, reasons for the players to go encounter other ORDs.
Develop each ORD's identity to the point where you know it well enough to play it naturally -- your NPCs' names, wants, demeanors; your puzzles' pieces and properties; your forts' entrances, guards and traps.
Write down at least one way to overcome each obstacle.
Record your ORDs on the scenario flow chart. Imagine this is a game that has already been played. Your goal is not to reproduce that game, but to use it as inspiration to jointly craft a new game with the players. In play, don't even look at your flow chart unless you have to.
Guideline: in play, improv more Directions onto your ORDs. Do this most early in play; late in an adventure, when the players have already parsed a lot of info and formed some plans, singular Directions are fine. When you do improv new Directions, they can point to other prepped ORDs or to new ones you invent on the fly. Don't invent more than you can handle playing, though! If you feel the urge to reward the players for something other than overcoming an Obstacle, improv a "freebie" Direction. I'll often do this for good teamwork on engaging, colorful character portrayals that, while not plausibly addressing an NPC's Obstacle, seem like they ought to get some compliance.
Guideline: vary your NPCs. See the templates for inspiration.