log in or register to remove this ad

 

D&D General How do you sandbox ?

This is a thread for posting about how you run and prepare for sandbox games, what style GMing you bring to the table, etc. If you have experience and thoughts on sandbox campaigns, post away!
 

log in or register to remove this ad

payn

Adventurer
I dont know where I heard this, or who came up with it, but I like the onion model. You start by detailing a well defined center. This is the starting place of the sandbox and has a lot of info the players start. Many elements are within the PC's reach. Then, you begin to add layers that are less dense and as you move away from the center. Basic ideas that become fuller and more dense like the center as the additional layers are explored. That's how I start and progress in my sandboxes.
 

Oofta

Title? I don't need no stinkin' title.
This was pretty much covered recently by this thread, but it depends on your definition of sandbox. I do sort-of-sandbox in that the PCs are always free to do whatever they want. On the other hand, I always give them a set of options before they move on to the next arc (at the end of a session) on what they want to do next so that I can prepare those scenarios.

However, I keep the scenarios pretty open. I know who the actors are, what their goal are, what the high level conflicts are and so on. Then if the PCs go off the rails I just improvise and either add a new actor or pull one in the PCs may or may not know.

So at the end of session, I let them know they have choices A, B, C and mystery bag whatever they want to do that I didn't think to include. So I prep for that. If, during session where they tell me they wanted to do B and then they go off the rails completely, I'll revise pre-planned encounters (I always do high level prep on an extra encounter or two) and go with it.

For example, I had the PCs travelling from one city to another and they had stopped in kind of a shady town, reputedly run by criminals bandits. I hadn't really fleshed out the town because they were just supposed to be there overnight, the description came from some general notes that I just threw in for flavor. Well, the rogue decided to try to make a bit of profit in town without telling anyone else, the entire party was soon involve due to some bad rolls. The next thing I know it's a battle in the streets with burning warehouses, breaking down the walls into a brothel to escape the flames and the group is taking on one of the crime bosses. Totally unplanned.

I handled it by changing the fluff on some of the monsters so they were human, using my handy list of randomly pre-generated names I always have handy and a lot of improv. Then I added the criminal organization to my "actors" log with a note of enmity to the group.

So that's how I handle it. Most of the time I ask what direction they want to go next with enough time for me to do prep work but I'm also willing to improvise and make things up on the fly. More difficult to do online than in person, but the same concept applies.

*Actors is general. Could be an individual, an organization, a political faction or even a nation. Human or monster, doesn't matter.
 

iserith

Magic Wordsmith
I start with determining how much real time I expect we'll play in the sandbox. Usually that's 20 to 30 four-hour sessions, tops, which equates to about 1 to 1.5 years. This helps me determine the size and scope of what I need to prepare.

Then I think about the theme and genre, soliciting player input on this to ensure that they are interested in the concept. This will inform what sort of rules variants or house rules I want to employ, plus character creation rules. It's important that the rules and options all play into the central theme and genre so it's all sort of self-reinforcing.

I will typically use a hex crawl or point crawl design, depending on whether it's pure wilderness exploration or settled lands (or mix). I create lots of tables for random generation of encounters and a process by which that is resolved during play, plus a number of fixed, dynamic adventure locations and factions. I will put all of this into a VTT (maps, handouts, monsters, and macros) to make it all easy to manage.

Now the players create characters and set about playing, picking up on hooks as it makes sense to reveal them or discovering things just by exploring. It pretty much runs itself. As we approach the real time limit of the campaign, hopefully things have escalated enough where there is some kind of climactic drama and a good resolution for everyone.
 
Last edited:

overgeeked

B/X Known World
Start with a bang and do targeted prep for a few of the various options they might immediately pursue, but be ready for them to wander wherever they want. That's the point. They wander into some entirely too high level area for their current level, they can bug out or die fighting. Seed the game with hooks they can pick up on but don't force anything. If they can't follow any lead (or no leads) you present, then it's not a true sandbox. Ask them where they're planning to go next after each session and do the same targeted prep for various options in the direction they indicated. It's good to begin with a fairly well fleshed out starting area, like a small town or village. Matt Colville and DungeonCraft have good videos on running sandbox games. Sly Flourish's Return of the Lazy Dungeon Master has a lot of really great advice, too.
 

DeviousQuail

Explorer
This was pretty much covered recently by this thread, but it depends on your definition of sandbox. I do sort-of-sandbox in that the PCs are always free to do whatever they want. On the other hand, I always give them a set of options before they move on to the next arc (at the end of a session) on what they want to do next so that I can prepare those scenarios.

However, I keep the scenarios pretty open. I know who the actors are, what their goal are, what the high level conflicts are and so on. Then if the PCs go off the rails I just improvise and either add a new actor or pull one in the PCs may or may not know.

So at the end of session, I let them know they have choices A, B, C and mystery bag whatever they want to do that I didn't think to include. So I prep for that. If, during session where they tell me they wanted to do B and then they go off the rails completely, I'll revise pre-planned encounters (I always do high level prep on an extra encounter or two) and go with it.

For example, I had the PCs travelling from one city to another and they had stopped in kind of a shady town, reputedly run by criminals bandits. I hadn't really fleshed out the town because they were just supposed to be there overnight, the description came from some general notes that I just threw in for flavor. Well, the rogue decided to try to make a bit of profit in town without telling anyone else, the entire party was soon involve due to some bad rolls. The next thing I know it's a battle in the streets with burning warehouses, breaking down the walls into a brothel to escape the flames and the group is taking on one of the crime bosses. Totally unplanned.

I handled it by changing the fluff on some of the monsters so they were human, using my handy list of randomly pre-generated names I always have handy and a lot of improv. Then I added the criminal organization to my "actors" log with a note of enmity to the group.

So that's how I handle it. Most of the time I ask what direction they want to go next with enough time for me to do prep work but I'm also willing to improvise and make things up on the fly. More difficult to do online than in person, but the same concept applies.

*Actors is general. Could be an individual, an organization, a political faction or even a nation. Human or monster, doesn't matter.
This is my general method as well. As long as I know what needs preparing for the next session or two then the players can choose to do almost whatever they want.

There are a ton of sources for sandbox design so I'll just say this; don't do too much ahead of time. This isn't homework, it's a fun game. Prep what you need, have some contingencies, and then let things develop as you go.
 

Ath-kethin

Elder Thing
I string together short(er) modules as a rule, which gives me a lot of flexibility in terms of what we are doing next and how it's happening. I work in NPCs that are relevant to the party and always keep the background of the world moving around and behind everything that's going on.

That way, I can adapt to what the players want to do and how they want their PCs to develop while still benefiting from the time-saving nature of running prewritten adventures.

Now, I don't really think of this approach as a sandbox exactly, but many of my players have referred to it that way.
 

jgsugden

Legend
I build a foundation. Then I talk to the player about what they want to play. Then I weave.

My foundation will have two to five storylines. At least one of them will run from levels 1 to 20. The others will run for just a few levels, and might be planned to start at first level, or much higher levels. The key features of these storylines is that they are externally driven. There are things already in motion that will spring events into play, regardless of what the PCs do. There might be a Demon Lord locked behind a door that will open at a specific date. There might be a natural disaster that will bring about a conflict between two nations. Some of these are more of a railroad, but others just move around all the puzzle pieces and create a new normal for the PCs to explore.

Then I talk to the players about the PCs they want to play. I ask them about their favorite stories. I collect raw elements from them that I can either interweave into my foundations, or weave into additional stories that will develop out of their contributions. For these, I don't usually have a conclusion in mind. Instead, I let the story tell itself. Usually, I take the elements the players give me, and I twist them in an unexpected way so that the story tends towards a direction they do not expect. However, sometimes I dive right into what they set up and run it straight at the PCs just as they'd expect.

Often, the start of a sandbox is a railroad. That railroad will take the PCs to a place they do not know, leaving them in unfamiliar lands, with no contacts and emerging objectives. Then, the floor drops out and they get to decide what options they wish to explore. There is no way they have time to do everything. Some things they don't touch go bad and feed future stories. Others get addressed by other hero types. Then, when the time is right, we upset the applecart with those later developing foundation storylines. By the time we get to ~17th level, the field of options begins to narrow again until the ultimate foundational storyline that has been foreshadowed, developed and threatened for the entire campaign busts open and carries them to the finish line.
 

I set the parameters of the sandbox, usually focusing on a home base and surrounding region. I detail the major players in the town, plus any major locations I want. Then I design about a dozen or so adventure hooks to drop to the party. The plot hooks not taken will resolve as they would without player interference, showing the players that their actions and inaction has consequences. As the campaign progresses, I'll expand the size of the sandbox to include more adventure opportunities. I generally don't have any overarching storyline, unless one evolves over the course of play.
 

kenada

Hero
Supporter
I’m running a sandbox *crawl. The Alexandrian has been influential in how I go about managing that, though Apocalypse World has also been influential in inspiring me to be mindful of my agenda and principles. When we started our current campaign, I pitched an idea (PCs are part of an expedition exploring an uncharted region) and then built up what I needed around that.

My two primary documents are my hex map and its key. I’m in the process of transitioning to a new scale (from 12 miles to 6 miles), so the key is getting reworked. I originally tried to key every hex, but I’m only keying a subset because keying 484 hexes is too onerous. The PCs have a primary base of operations, and they set out from there to explore. I also have documents on settlements and dungeons, which get detailed as necessary.

I didn’t have any particular plots in mind when we started, and I still don’t (about 30 sessions into the campaign). Everything that’s happened (from finding their initial base camp, to moving it to a new location, and what and where to explore, etc) has happened at the PCs’ instigation. I keep a timeline of events and as well as a list of things that will seek out or interact with the PCs to change the status quo (“bangs”). Even if the PCs ignore what’s happening, the world carries on without them (e.g., they’re opted not to help with a bulette problem, so the town is trying to handle it, and that’ll result in shifts in power as factions come together against a common foe).

One thing I try to do (and this is part of the Apocalypse World influence) is name every NPC and try to create relationships between them and the PCs. When the PCs went to recruit retainers a few sessions ago, they all had details and personalities. Alas, the PCs’ subsequent mission ended in a rout, so they’ve decided to part ways with the surviving retainers. However, they’re part of the world, and I’m sure they’ll show up at times in the future.

Because of the sheer scope (especially since this is a homebrew setting), a lot of my prep is just-in-time. I feel pretty comfortable improvising though, and if something does get too screwed up, we can fix (retcon) it. That’s not really been necessary though. The biggest challenge has been getting concrete plans out of my PCs. They’ll say they want to do one thing and then go do the opposite. Last session, they had a great plan for attacking a pack of ghouls, which they abandoned (causing the previously mentioned rout).
 


77IM

Explorer!!!
Supporter
  1. I try to steal as much as I can (monsters, encounters, adventures, areas, NPCs, etc.) so that I don't have to spend much time on prep.
  2. I throw as much stuff at the wall as I can, and see what sticks. "Stuff" = plot hooks, NPCs, factions, areas, dungeons, treasures, etc. "Sticks" = gets the PCs' attention and interest. Then I put that sticky stuff front-and-center, embellish it, connect it to other sticky stuff, etc. (Part of my gaming philosophy is: Don't try to make the game fun. Instead, make the game interesting, and then the players will make it fun.)
  3. Secrets! I'm always creating secrets. For every stuff, there's something to be discovered about it. But, and this is important, DON'T hoard your secrets -- let the players figure them out at a steady pace. Any action that might reasonbly turn up a secret, should turn up a secret. And it's fine because there are always more secrets to be found. This is actually something I am trying to get better at; I tend to guard my secrets too closely, hoping for a "big reveal," but it usually works out better to just let players in on the secret.
  4. I highly recommend that anyone who wants to level-up as a DM should run a campaign of Apocalypse World. It is a real eye-opener, and you can take techniques learned there and apply them in other games, even much crunchier games (like D&D). But for sandboxes specifically, another great training tool is Mutant: Year Zero. The whole thing is basically a sandbox hexcrawl that you randomly generate as you go. It's kind of brilliant. The game provides you with a lot of cool elements and then you roll them randomly and have to figure out how they all fit together, right at the very moment the PCs step into the hex! (well they use a grid rather than a hex but it's the same idea.)
  5. Take good notes. I've used campaign-logger.com in the past and it has worked very well.
 

My sandbox campaigns are confined to a specific area. They are not the entire world. This area can be a region of the world, a country or even as small as one city, or one building. But within this space the players are free to go as they like.

I present multiple clues and plothooks to my players, which often share some plot connections or are part of the main plot in some way. The main plot is usually in the background, but occasionally parts of it will be brought into the light. The players are not obligated to chase any of the hooks, but the plot responds to their actions, even if that action is ignoring the plot. The plot can continue without the players, creating the feeling of a living breathing world.

The plot does not rely on any singular villain. And so I can introduce many disposable villains, without fear of their lives ending suddenly. I keep the main threat off the stage until I feel the end of the campaign is near.

I use dozens of random tables to inject a large variety of encounters into the campaign, based on their location and the type of environment they are in. I have city encounters, coastal, jungle, sea and underwater encounters. I even have interdimensional encounters.

Exploration is very important to me, so among the random tables are also discoveries. These include obstacles, geographical features, ruins, settlements, dungeons and landmarks.

I also roll randomly to determine the weather and wind direction, plus the movements of sea monsters.

These sort of campaigns require a ton of prep before we start playing, but once prepared I am able to run most of it without further prep. I do need to make maps for upcoming battles or new locations my players visit, which means also prepping monsters and npc's. All in all it is a lot of work, but very fulfilling.
 

Li Shenron

Legend
I don't exactly sandbox but for me the ideal is to have a lot of stuff up for grab for the PCs: plots to investigate, locales to explore, quests to start while having others still ongoing, and plenty of connections between everything...

One thing that helps a lot is having a stash of maps at different scale. For me an essential part of playing in a sandbox is the ability to say "we go there" while choosing a point of interest on the map, and most likely find something to do, whether it's a combat or social encounter, or an item to acquire or a piece of knowledge to discover.

For the same reason, good ready-to-use lists of everything: monsters (but the MM books are already that), NPCs, items, rumours, unusual locations, mini-adventures, and so on.

If I have time, I can just collect a bunch of each category and decide their placement on the map, but the whole thing can also be improvised and the players won't notice a difference between the randomness of your decisions during the game and the randomness of your decisions before the game. It helps a lot if you're the kind of DM who doesn't have a God complex (i.e. wants to fully design the universe top-down in advance) and you're ok with the fantasy world to come alive on its own.

A key feature of sandboxing is that the fantasy world is not tailored to the PCs, particularly in terms of difficulty level: monsters are as difficult as they are, not magically adapting to the PC's current level, so start a fight at your own risk!

However, personally I am not afraid of flat-out telling the players what is a monster's CR, or at least whether a certain fight or other kind of challenge is a deadly risk. That's because I never have a whole group of players all of which are ok with their PC dying, there's always someone who wants the opportunity to create and develop a character long-term and would be disappointed to be forced cut short.

For the purpose of making this feel more in-character, I use the concept of "zoning" i.e. the PCs have in-character knowledge of which regions are safe and which are infamously dangerous, so at least the extreme CR-PC level differences can be easily avoided.
 

aco175

Legend
Would Icespire Peak box set count as a sandbox? I tend to think so and some may depend on how I run it. This is my current campaign which I have the 'quests' that are given to me in the box and a few more that I made along the way that connect if the players want to explore more. There is some railroading in the session 0 where the players and PCs agree to explore the region and not suddenly say that they all want to go to Baldur's Gate or such. They all created PCs with ties to the region and even Phandalin where they meet.

The starter adventure I made introduces some new NPCs and businesses since the box set does little for the town and you really need the Lost Mines box as well. Follow-on adventure involves taking one of the quests or furthering the starter adventure. Higher level quests and areas are unfolded along the way. Maybe it is more a structured sandbox. My players are fine with me presenting a few options for them and they go along with it over trying to forces something else like going to someplace for no reason. They may want to go to Leilon or Neverwinter, which they can, but there is not much planned yet for those places and it may run boring for a while.

I also like to get an idea from the players on where they are planning to go at the end of each adventure, so I can plan some encounters and events for next week.
 


Emerikol

Adventurer
I think the onion approach mentioned above is totally valid and is the original way it was done I believe.

First, I create a map. Typically at least a continent sized one. Think of the Forgotten Realms, Greyhawk, etc... I love maps and this truly is a labor of love but it is a labor and is often as much work as many of the succeeding steps combined. Right now I use worldographer. If I have educated people in my group, I will eventually give them a map though it may not be as detailed.

If I am going to deviate from the vanilla understanding of races, I do it here. I don't always. Last time I made the elves psionic instead of magical. I tend not to use the newer races much. So elves and dwarves are in all my campaigns but the rest are sprinkled in for variety. I also tend to make some of the humanoid races into full blown national actors. So you might see a hobgoblin trader in port. The race might tend evil but they are always killed on sight. I vary which ones.

I often develop my cosmology and religions at this time. I develop the various priesthoods and how they operate. I love this aspect of the game as well. Creating truly distinct religions is a lot of fun.

Then I work on the nations. Typically I will figure out some history going back a few hundred years. This is a medieval style world so much farther is probably overboard. I just use broad strokes back from there. Typically I detail out the economic details at this time in a broad way. What does the nation produce, who does it trade with, how strong is it, how is it ruled, etc.... Think of the Greyhawk gazetteer but in my own style which perhaps in some ways is more verbose but less polished in the writing.

I define the cities with names, populations. I create a rogues gallery of notable personalities. This would include notable monsters like a dragon. I tend to play dragons differently than D&D. They are practically walking gods who typically dominate a region for some time. Only the very highest level PCs dare go after a dragon and when they do they put their affairs in order before setting out to do it. So when I say cities, I do not mean fully detailed cities.

Now all of the above can be reused so I might run several campaigns in a world before making another.

Then I create my first sandbox. Typically it is on some frontier area. It is rarely near a big city. As cities are a major creative effort. When I do create a really big city, it is often the sandbox. I will often start working on the details of the cities though while the group is mucking about in the city. These details might be nothing more than maps. I sometimes use maps of cities from third parties.

Inside the sandbox, I create appropriate wandering monsters tables. I create the NPCs of the local village/town. Typically in a good bit of detail.

I then begin stocking the area with adventures and places of interest. Most of them are small one session shots. Others are fully developed dungeons. I tend to vary things up a bit so I have the lairs of the major humanoid enemies. I have forgotten tombs, etc... I just vary things. I tend to make stuff closer to the base easier and the farther out you get the more dangerous it gets.

Another alternative is to create a single large multi-layer dungeon like Gygax did with Castle Greyhawk. I find this to be a good change of pace but not something I'd do all the time for sure. The players kind of have to buy into this idea in advance of the campaign.

I also prepare a lot of names, NPCs,etc... for those situations where I might be forced to ad lib it as mentioned above. I try to avoid it but you can't always. If you have enough of these ready to go, you can always use the ones that don't get used this campaign in the next. These are just ready combat encounters. I also like to use some kind of personality building tool for the NPCs. I have over the years tried a variety of different things. Even psychology tests. :). I want my NPCs to seem different and not just a monolithic version of me.
 

GlassJaw

Hero
I love sandbox-style campaigns. I can't imagine running a hack & slash railroad campaign again. No thanks.

That said, the best lesson I learned about running a sandbox campaign is that the "sand" is just as much about clues, hooks, leads, and rumors as it is locations. Actually, I'd argue that they are more important than the locations themselves. Whereas a traditional hexcrawl is solely location-based, I like to run mystery sandboxes, where the players have to choose which clue to follow, not which point on a map.

In my experience, it's woefully inefficient to begin designing a sandbox campaign by creating a massive area map and filling in locations. That comes much later, and can even wait.

I start by brainstorming about what's generally going on in the region. Who are the major players and factions? What are their motivations and goals? What are the primary conflicts or sources of tension in the region? Who will the players interact with on a regular basis? Are there are any special or timed events that are going to occur in the near future? What are the rumors going being talked about locally?

Once that starts to take shape, I'll dive a little deeper and add some detail to the major NPCs. I'll start connecting the dots on their plans and how the NPCs might be connected to each other. It's especially helpful to think about what would happen if the PCs never get throw into the mix. After that, I'll try to come up with a secret for every NPC, major or minor. It's also helpful if the players do some of the "work" themselves and give you some basic background on their characters. Those can be mined for plot hooks as well.

During the brainstorm, I'll also come up with some unconnected leads and one-offs (side quests!) and locations too. Everything doesn't have to be connected, and the players figuring out what's what is part of what makes a sandbox fun. And yes, part of the initial setup is a map of the starting area and having some locations prepared. But again, I don't make that the focus at the start.

What's going to grab players immediately is choices. And I don't mean "left or right?" or "pick a hex". Choices will ultimately lead to a location, but the location itself isn't the driver (at least not all the time). When the players are unleashed into the campaign, whenever they talk to an NPC or discover a clue, it will lead to another NPC/clue/location, etc, etc. After a couple of sessions, ideally they'll have a few breadcrumbs to follow. Now you've got your sandbox!

Once I have a good idea of the goings-on behind the scenes, I come up with an initial hook that ties the players in. I try to make it have some personal connection to one or all of the characters. You want the players to have immediate buy-in so the upcoming choices feel important. It can be as simple as a letter from someone that needs help (family member?) or huge and sudden event that the players are pulled into (like a coup attempt!).

I'll also add that complications and events are great for mystery sandboxes. There's nothing like throwing a wrench in the players' plans and forcing them to think quickly and then make a difficult decision about what to do next, especially if they are on the clock.

It goes without saying that I borrow heavily from published sources, and the easiest things to repurpose are maps. That's one of the reasons why I don't spend much time up front fleshing out the details of locations.

Finally, I always refer people to this module: The Beast of Graenseskov, written by ENWorld's own Quickleaf. It's an amazing example of a mystery sandbox. It has a massive amount of content (maybe too much!) within a relatively small area. The players are quickly thrown into a mystery and get a ton of clues right out of the gate. Following up on those clues leads to others and to new locations, etc, etc. It's a textbook study in creating a sandbox campaign.
 

Emerikol

Adventurer
I love sandbox-style campaigns. I can't imagine running a hack & slash railroad campaign again. No thanks.

That said, the best lesson I learned about running a sandbox campaign is that the "sand" is just as much about clues, hooks, leads, and rumors as it is locations. Actually, I'd argue that they are more important than the locations themselves. Whereas a traditional hexcrawl is solely location-based, I like to run mystery sandboxes, where the players have to choose which clue to follow, not which point on a map.
I would agree that these are import. I call them plot threads. So and so is up to no good and here is what he is doing. Here is how he operates. Here is what he is planning. I develops tons of NPCs who are up to no good and others who are trying to do good.

The locations of which I speak though are the places where these plot threads are operating. If someone is plotting to murder a neighbor that neighbor and his home are defined. The lair of the gnoll warchief who is plotting to destroy the local village by allying all the gnoll tribes for one massive attack for example.

So the PCs engaging the world is all about picking up on these plot threads. I don't introduce a thread I'm not willing to go to full fruition. So if the group ignores the gnoll warchief then he will ultimately destroy the village assuming there aren't other actors involved. If there are then it would involve dice rolls to determine how it goes without PC engagement.
 

GlassJaw

Hero
I call them plot threads.
Great post, I entirely agree. I wanted to emphasize that it's the clues and hooks that get the players to the locations, not the locations themselves. I mean you can run a simple hex-to-hex sandbox but I find the breadcrumb style way more interesting.

One nitpick is the word "plot". I try to avoid using it. A DM shouldn't create plot. The plot is created when the players interact with the events and happenings that you've set in motion. It's semantics I know but it helps me get in the headspace of what my job as a DM is, especially in a sandbox.

It really helps me present things as a neutral adjudicator and then get out of the players' way. It's my job to present the world as it is and let the players decide how to interact with it. It's not my job to present only the story I want to tell.
 

An Advertisement

Advertisement4

Top