Grain size in your sandbox?

[virgil]

Say you're wanting a campaign setting that is more than Generic Fantasy B7 ("you are in the woods, there are orcs attacking"), and you want the adventures to be more involved than basic dungeon-delving and prince rescues. Altogether, a robust sandbox setting.

How much detail should you write out in the setting, and how much needs to be revealed up-front to the party? What elements need to be covered? How do those answers change if the setting is very similar to others (welcome, to Center-Earth!) as opposed to relatively unique ("in the boron jungles of Tryantheor, only the feathered starfish-men are so dishonorable as to start a fight with weapons you cannot match")?

[hogarth]

How long is a piece of string?

The amount of detail a GM needs to include depends almost entirely on the GM's ability to "wing it" convincingly and the players' tolerance for an improvised campaign.

[Dean]

Whatever you do you should follow the laws of conserving conceptual space. There is nothing gained by making your players be in the Tyantheoran Forest as opposed to just The Forest. Add details when they give your setting additional traction, add details that promote interactivity.

I got to watch a friend of mine become a much better DM over the last few years. He used to be an archetypal 2nd ed DM and when he ran a game he would create a campaign packet for the players to read. I gain great joy by looking at the packet he made for a game 3 years ago and the packet he made for our last game a few months ago after reading things around the Den and talking about them. 3 years ago his packet is 6 pages thick and is full of special names for everything with apostrophe's all over them. He has special species, proper names for famous elves and dwarves in history, proper names for dozens of regions and countries. It's dense and thick and there's nothing you can do with any of it. No one cares that the dwarf lands are called the Drumard and they were united 2000 years ago by Thrain Stonestone. The entire campaign never touched on any of the details in that packet because they were not made to be interacted with, they were just there. His latest packet was a page long. It had brief writeups of the locations around us that we could adventure in like the Land of Black Ice and the Hellmouth each with a little adventure seed sown into them. It told us who the King was but did so in an adventure hook about how he and his Vizier were in a power struggle that was locking other members of the city into sides. It was brief, full of details you could play with, and dripping with character. We interacted with that story from session one and continued doing so until the game ended.

[silva]

What people said above.

Also, a good way to get players involved is to make them participate in the creation process, even if its just small bits. So if instead of bringing a full pre-created setting, you work out parts of it with the players, ("Hey John, where do you come from ? How is this village ? Did they do some wrong to you ?" etc) in general the players get more involved and excited about it, feeling like the setting is theirs too. It helps them buy-in. And even with parts of the setting that are more well developed, consider hearing and accepting players suggestions about it too.

[Wiseman]

All good advice. In addition, the world shouldn't be static around the PCs. There should be consequences for their actions. Say if they happen to take out the hobgoblin bandits in the woods, trade routes open between two formerly separated towns.

If the PC's instead of fighting the bandits go raid a tomb full of necromancers, the bandits might get ambitious and attack one of the towns. If the PC's don't attack the tomb the necromancer might succeed in creating a particularly power undead or binding a threatening demon. However if the towns are restored to contact with each other, they can send gear/cohorts/weapons for the PCs to deal with the necromancers, perhaps clearing out the tomb of some revered king or religious leader, bringing another town to their side.

Inversely, if they attack the necromancers first, the bandits raid the town, but the PC's now have the necromancers magic items, plus whatever might have been hidden in the tomb, aiding them in freeing the town from the bandits.

Thus over time, if the PCs start bringing a formerly fractured land together, they start to be seen as leaders, bringing them into the high-power leagues.

EDIT: Ninja'd by K.

[K]

The problem with a lot of sandboxing is that a lot of it is not actual sandboxing. The encounter of the road with a troll merchant that the DM planned can take place anywhere in the kingdom and PCs will run into it on any road they take.

I think the easiest and truest way to sandbox is to simply run a story in the campaign and then let PCs figure out where they belong. Set them in a city that is being attacked because the King of Blackreach wants the Merchant Prince's gold and then let PCs figure out how they want to interact with that event.

To do that you set a script of NPCs and a roster of places that are related to that event. Set up a clan of dwarves who will help the Princes if you kill the dragon who is sitting in one of their mines. Write-up some NPCs that will annoy the PCs so that they have people they'll want to assassinate, then use urge to set them in the middle of that NPCs dungeon delve for a missing artifact.

The illusion of the sandbox is maintained by giving multiple potential outcomes for events based on PC action and by convincing PCs that they have performed actions that matter.

[hogarth]

Whatever you do you should follow the laws of conserving conceptual space. There is nothing gained by making your players be in the Tyantheoran Forest as opposed to just The Forest.[..]
No one cares that the dwarf lands are called the Drumard and they were united 2000 years ago by Thrain Stonestone. The entire campaign never touched on any of the details in that packet because they were not made to be interacted with, they were just there.

The world don't move to the beat of just one drum. What might be right for you may not be right for some.

For instance, if a player asks his GM "what's north of the forest?", one GM might be great at coming up with a sensible answer off the cuff, and another GM might get flustered by the question without preparing details about geography and history beforehand. So GM #1 might find a thick campaign binder useless and GM #2 might find it a godsend.

[Laertes]

My preferred beginning to a sandbox is to ask the players what sort of thing they want in the game. Ideally, you want to come up with a huge list of stuff which should go into it. Anything not mentioned, or mentioned only in passing, isn't that important.

Any player who can't state the things that are important to them in a game, or who believes that building the setting is the GM's task alone, is not cut out for a sandbox game.

[Blicero]

The world don't move to the beat of just one drum. What might be right for you may not be right for some.

For instance, if a player asks his GM "what's north of the forest?", one GM might be great at coming up with a sensible answer off the cuff, and another GM might get flustered by the question without preparing details about geography and history beforehand. So GM #1 might find a thick campaign binder useless and GM #2 might find it a godsend.

There's a big difference between "I have no idea what is in the world beyond the 1 square mile area the PC's are currently in" and "I know what the general lay of the land looks like, I just don't force my players to memorize a bunch of proper nouns before we start playing". The MC doesn't have to share all of their preparatory work and backstory stuff with the group. One of the options I've described is consistent with what dean said, and the other is not.

[Wiseman]

The world don't move to the beat of just one drum. What might be right for you may not be right for some.

For instance, if a player asks his GM "what's north of the forest?", one GM might be great at coming up with a sensible answer off the cuff, and another GM might get flustered by the question without preparing details about geography and history beforehand. So GM #1 might find a thick campaign binder useless and GM #2 might find it a godsend.

There's a big difference between "I have no idea what is in the world beyond the 1 square mile area the PC's are currently in" and "I know what the general lay of the land looks like, I just don't force my players to memorize a bunch of proper nouns before we start playing". The MC doesn't have to share all of their preparatory work and backstory stuff with the group.

I don't see the problem with giving places names. Just don't be too insistent on the PCs knowing or remembering them. Simply drop that the forest they're in is the Cith-Ralden forest once or twice and that more importantly elves live here and are highly aggressive towards non-elves. Giving a place a name, even if only dropped once makes the setting and game more memorable. However, overdo it and your players will hate you.

[Dean]

The reason I disagree with planning an entire geography and history and setting down names and dates and places is that it doesn't fucking matter, the players don't care, and it cramps the possibility of future storytelling. If I give my players a rough map of the region with some hotspots picked out that is a better map than one with every town and woods labelled and named. Exploration and discovery are thrilling sensations and taking that away from players benefits no one. Labeling every part of the map also means that next week you can't travel to The Obsidian Eyrie, a Demonologists fortress rising from an active volcano, because you didn't think of it a week ago when you made your map. D&D is a continuing story so there's no reason to constrain yourself to only the ideas you could come up with before the first session.

D&D is more like a TV serial than a fantasy novel and when Buffy travels around Sunnydale it continues to develop interesting places. Sure sometimes this means that Sunnydale will end up having 43 Churches, an airport, a military base and 2 different piers and only one high school but even that is a better than the alternative of having the entire show take place only in the locations written up for the airing of the pilot.

[Sakuya Izayoi]

My current campaign is sort of a hybrid. I drew up a map beforehand of a continent divided up into prefectures, Sengoku-era Japan style, with brief regional and cultural descriptions of the divisions. But the action stuff the players interact with, like towns and other locales, the dangers outside of them, and enemy factions, is based upon the plot threads that they've given me through their character biographies, or stuff in the world they've shown an interest in.

This has the advantage of letting me describe the general conceits I want (Asian influence, Tao and Shinto-influenced mysticism, No Elves, No Dwarves, Final Destination), and indulge in my cartography fetish, without having to write a novel, or prep stuff that won't get seen by the players.

[MGuy]

I think the easiest and truest way to sandbox is to simply run a story in the campaign and then let PCs figure out where they belong. Set them in a city that is being attacked because the King of Blackreach wants the Merchant Prince's gold and then let PCs figure out how they want to interact with that event.

To do that you set a script of NPCs and a roster of places that are related to that event. Set up a clan of dwarves who will help the Princes if you kill the dragon who is sitting in one of their mines. Write-up some NPCs that will annoy the PCs so that they have people they'll want to assassinate, then use urge to set them in the middle of that NPCs dungeon delve for a missing artifact.

The illusion of the sandbox is maintained by giving multiple potential outcomes for events based on PC action and by convincing PCs that they have performed actions that matter.

When people talk about sandboxing I always assume that this is what people are talking about because this is how I generally run it.

[silva]

I think the easiest and truest way to sandbox is to simply run a story in the campaign and then let PCs figure out where they belong. Set them in a city that is being attacked because the King of Blackreach wants the Merchant Prince's gold and then let PCs figure out how they want to interact with that event.

To do that you set a script of NPCs and a roster of places that are related to that event. Set up a clan of dwarves who will help the Princes if you kill the dragon who is sitting in one of their mines. Write-up some NPCs that will annoy the PCs so that they have people they'll want to assassinate, then use urge to set them in the middle of that NPCs dungeon delve for a missing artifact.

The illusion of the sandbox is maintained by giving multiple potential outcomes for events based on PC action and by convincing PCs that they have performed actions that matter.

Nah, I dont like this method. What if the players dont care for your story, or for political conflicts to begin with ? Then you have a setup that you find awesome yourself and will probrably have tons of fun running, but the same wont hold true for your players.

See, I think a very important thing for a sandbox game is having player input direct the game from the start, not only affecting a branching plot or pre-created setup, but dictate what form it will take and what direction it will go from there. This may be achieved through a frank conversation before the game begins (like Laertes said: "What do you want to play and how ?"), or through the concept of "flags" that some games use (see here for a brief explanation), which is just another way for the player to say "THIS is what I want the game to be about!". In fact I would say player input and clear comunication of interests is important for any game, but it is of paramount importance for a sandbox.

Thats my 2 cents, at least.

[Laertes]

Nah, I dont like this method. What if the players dont care for your story, or for political conflicts to begin with ? Then you have a setup that you find awesome yourself and will probrably have tons of fun running, but the same wont hold true for your players.

Then you change your story and your conflicts until you have a story and conflicts which the players are totally psyched about and can't wait to take part in.

Surely every GM in every game knows that?

[silva]

Then you change your story and your conflicts until you have a story and conflicts

Changing everyting midway may not be possible depending on the setup, or may be possible but at a cost of time some (or all) players may not have for giving another chance to the GM. I agree its important to constantly "measure" player buy-in and change / adapt things according to the different flags players go raising through the game. But the soon you align players expectations to the actual stuff in the game, the better. And theres no better method for this then having a frank conversation at the very beginning, or using games with built-in flags. It also has the benefit of making the GM not waste time creating/writing stuff that he wont use because of player-rejection.

surely every GM knows that ?

I wouldnt say so, based on the number of GMs I played with that brought their entire campaign pre-written for us to play, just to see it crumbling when we didnt follow the bread crumbs trail he prepared, And also on the number of books that explicitely advise this kind of thing (look at Vampire and Shadowrun, for example).

[Laertes]

But the soon you align players expectations to the actual stuff in the game, the better.

No. The stuff in the game needs to align with player expectations, not vice versa. If your players all want to sit around and basket weave, then they will have fun. If they want to optimise characters and play tactically, then they will have fun. You need to find out what they want, and then run that game, not try to shoehorn them into the thing you want to do.

If different players have different expectations to the point where you can't do this, then you need to abandon the game because the game is unworkable.

[silva]

No. The stuff in the game needs to align with player expectations,

Yup, this is what I wanted to say. Thanks.

You need to find out what they want, and then run that game, not try to shoehorn them into the thing you want to do.

This is exactly what Ive writen in my last 2 posts, only worded differently. We are in agreement in all our points.

[hogarth]

The reason I disagree with planning an entire geography and history and setting down names and dates and places is that it doesn't fucking matter, the players don't care, and it cramps the possibility of future storytelling.

I agree that players mostly don't care about history and dates. Players certainly do care about geography and places.

Exploration and discovery are thrilling sensations and taking that away from players benefits no one.

WTF? Players love exploration and discovery in a campaign that has no geography or places?

Labeling every part of the map also means that next week you can't travel to The Obsidian Eyrie, a Demonologists fortress rising from an active volcano, because you didn't think of it a week ago when you made your map.

Okay, now we get to the crux of it -- you're saying that you, specifically, are better at coming up with things on the fly than preparing them beforehand. See my previous post:

[..] one GM might be great at coming up with a sensible answer off the cuff, and another GM might get flustered by the question without preparing details about geography and history beforehand.

--

EDIT: This discussion is ultimately the cousin of everyone's old favourite discussion with Zachary Ess -- "rules-light games are awesome because every house rule that comes off the top my head is fried gold".

[silva]

EDIT: This discussion is ultimately the cousin of everyone's old favourite discussion with Zachary Ess -- "rules-light games are awesome because every house rule that comes off the top my head is fried gold".

Well, for those who likes rules-light games this saying would go like this:

"rules-light games are awesome because they provide a basic frame-work to free the players imagination, instead of rules-heavy games which rules feel like straightjackets for imagination"

And then it would be the cousin of te following:

"sandbox games are awesome because they provide a basic framework to unleash players imagination, instead of railroading games that straighjackets players to the GM story."

[virgil]

So a type of gazetteer is ideal for a campaign setting that will be more involved than slapping the party in front of a dungeon?

[hogarth]

So a type of gazetteer is ideal for a campaign setting that will be more involved than slapping the party in front of a dungeon?

It really depends on your players. If you think they'll want to do a lot of free-form exploring, then obviously you'll need to have a lot of things for them to explore (either written down on paper or in your head, it doesn't matter).

[silva]

So a type of gazetteer is ideal for a campaign setting that will be more involved than slapping the party in front of a dungeon?

In general yes, the most concise and useful the info the better.

But it also helps making sure what is there in the gazeteer actually interests your players, so having a frank conversation with them about what they want out of the game (like Laertes adviced above) or keeping a lookout on the flags they raise (if the rules have those) is a good idea too.

[Dean]

Labeling every part of the map also means that next week you can't travel to The Obsidian Eyrie, a Demonologists fortress rising from an active volcano, because you didn't think of it a week ago when you made your map.

Okay, now we get to the crux of it -- you're saying that you, specifically, are better at coming up with things on the fly than preparing them beforehand.

No, I'm not. I'm saying something incredibly obvious. That I could write more good ideas if given unlimited time than I could if given limited time. If you fill in your whole map before session one then you can't spend time after session one adding things to your map. As a result you should only make things that make good use of the conceptual space they occupy. Because everything you make takes the space of something you could potentially make in the future so you should only put things in your setting that add value to it. My contention has never been to prepare nothing my contention is that you should only include things that seem to have traction and the potential for interactivity. Adding good things to the setting is good and adding bad things to the setting is bad and that's obviously true. Where you're getting lost is that I am also telling you that adding neutral things, things of neither positive or negative value, is also bad. It is bad because it adds nothing and takes up space that could be better used by something of actual value later on so you shouldn't do it because no one benefits from it.

EDIT: This discussion is ultimately the cousin of everyone's old favourite discussion with Zachary Ess -- "rules-light games are awesome because every house rule that comes off the top my head is fried gold".

No, it's not. Maybe that's what Silva is saying. I don't know, I don't read his posts. I understand that he is supposedly agreeing with me but I'm sure he actually isn't and he's just saying his prepared lines and pretending it's related. Basically what's happening is I'm having a real conversation with you and he is standing next to me blathering about railroading while bits of bear meat drop from his slackened lips. I do not know that man and I did not come with him.

[Sakuya Izayoi]

It's important to talk to your players and get to understand the kind of stories they want to tell with the characters they've made and figure out how bears can fit into the picture.