How to run a sandbox game

[tedcahill2]

If someone wants to point me to a thread that poses this question please do, I didn't find one.

What are your best tips and tricks for running a good sand box game?

My contribution: I have a pad of graph paper full of mini dungeon designs, warehouses, taverns, castles, etc., as well as individual puzzle rooms and trap rooms, that I can flip through and use on the fly.

Whenever I'm inspired and or have down time I try to add to the note book to keep expanding my options.

[Geddy2112]

As much as it sounds counter-intuitive, don't over-prepare. Have enough Tavern X Merchant Y random encounter Z to stall the party if they make a hard change in their goals mid session, giving you time to paint the scene and prepare for their next target. If you don't flesh out every minor detail of the world, you can have things appear when you need them. Players are on a quest to kill the evil dragon lord-good thing this cave of kobolds they are about to stumble into in the unexplored area will help them with that.

Second is to have a main plot and 2-4 notable subplots. Besthda games are an excellent example of this. Skyrim is a total sandbox, but there is the main plot about the dragons coming back/dragonborn, and the subplot of the Nords and Empire war, along with other minor subplots involving daedric gods, guilds, and the like. So your world is a sandbox, but there is a "main quest" and some lesser but big quests your players can go on. The key to a sandbox is that the players can ignore some or all of these and go fishing or whatever the heck they want.

On the pro level, your world should be living and breathing, meaning that your main quest and 2-4 subquests will advance even if the party does nothing. If the party does not choose a side in the war, one may eventually win. They might go fishing and come back to the port to find it burned to the ground by raiders, or taken over by another nation. Also their random actions should be able to influence these aspects of the world, but going out and sandboxing is not required.

[Malimar]

The main thing to consider is goals. Some parties are okay with bumbling around not trying to accomplish anything in particular, but most are not, and will fall apart without any clear goals. Either make each player declare one or more personal goals they want to work towards, or make outside quests easily accessible (my sandbox includes a bounty board in the Adventurer's Guild for an easy selection of quests), or both. (These goals don't have to be very related to the "main plot", if any. My sandbox has a "main plot" or two, but they're not as easily accessible, and need to be stumbled upon through the natural course of play; a couple of parties have indeed stumbled on the peripheries of these plots.)

[Red Fel]

Check your pants. Is the seat comfortable? Would you be okay flying by that seat for an extended period?

The best thing to have when running a sandbox game is an adaptable mentality. Have you ever done improv? Be ready to "Yes, and" a lot. When a player wants to do something, and offers a way in which he intends to try it, have an idea or two. Don't shut anything down automatically. Don't necessarily make everything easy, but make sure it's possible.

That's the core of an open world. Have you ever played Legend of Zelda: Breath of the Wild? It's kind of like that. If a player sees a mountain in the distance and wants to climb it, he can try to do that. If the party wants to raid the enemy camp, swords swinging, or sneak in under cover of darkness, or come up with an elaborate trick to distract and ensnare them, they can try to do that, to varying degrees of success. Be open to experimentation. Does a player want to try to smoke the Orcs out of the cavern instead of delving in, using firewood and a basic knowledge of physics? Let them try.

That said, Malimar is also correct. While it can be fun to wander from waypoint to waypoint, it helps to have goals. And while you don't have to assign those goals to the party, it helps if you have them operating in the background, or have the PCs come up with their own. For example, keep an idea of a handful of major world events going on - say, a political revolution, an apocalyptic cult, the awakening of an ancient Dragon, etc. - and keep track of where they are at any given time. The PCs may never interact with these events, or they may stick with one from beginning to end, but keep those plates spinning just in case they attract the players. Similarly, if one player says his goal is, say, discovering an ancient civilization, be sure to sneak ruins in there, with clues about more. If a player wants to find her PC's long-lost father, throw that in there, if she chooses to pursue it. And so on. Don't over-prepare too much, don't focus on tons of tiny details, but have a few big concepts brewing at all times, spinning around, if the players choose to tap into that.

[Aurosman]

I am actually creating my first one myself right now. I am having the first couple levels be about a certain plot and then i will open up the whole world as a sandbox. That way I should be able to figure out how the players want to play each character and can design encounters and lore for the future for each of them.

What you should do first is make a map, that way you an label cities, and where places are so you can be consistant. They have a basic world generator and it works great. I'm not that good at designing stuff like that and it has helped alot. Here is the link https://donjon.bin.sh/fantasy/world/

[Zanos]

Make it absolutely clear to your group that you aren't going to lead them by the hand, so it's important that their character have goals of their own, that are at least somewhat compatible with other people in the party, at the very least with a mutual assistance in personal goals being an understanding.

Provide possible hooks, but don't force the party down any of them. Don't get mad if the party doesn't decide to pursue one of your presented hooks.

I have played with people where an open world just does not work at all; if given a choice, they just clam up completely or argue about what to do for 2 hours.

[Quertus]

A lot of good advise has already been given, and more artfully than I could attempt.

The only thing I can think to add is, it helps when you know the world, the NPCs, etc. One common trick to that is to base them off people & places you know. The advantage being, when the PCs do something off the wall, like look for supplies to start a random fire, you know enough about the area to know what they'll find, and about the NPCs to know how they'll respond.

[daremetoidareyo]

Establish a home base by level 3. This will help players get a feel for whatever location they are in and give them some room to run their plots and personal stories as well. It also allows people to approach them.

[Quarian Rex]

Another thing to keep in mind as well is to have a lot of 'white space' on the map. Having every acre claimed by the crown (or what have you) can be stifling. There wasn't much that I liked from 4e, but the idea of a campaign based around 'points of light' has really grown on me. Also, try to portray the power structures of the various communities as being competent but stretched thin. Showing town walls being repaired after the last goblin/orc/owlbear/umberhulk attack can do a lot more to motivate players than a courtly decree, and them seeing competent town guard/patrols can go a long way to convince players to help rather than letting Darwin sort things out.

Be careful with the competence angle as well. A community capable of defending itself (though perhaps at growing cost) is a community worth trying to save. A town guard manned by hyper-competent badasses begs the question of why they aren't solving their problems themselves. If you want the PCs to respect the town guard, don't try to make them better than the PCs, just show them as trying (and mostly succeeding) to save their people from the savage wilderness filled with murderdeath that is the typical fantasy world.

A good sandbox world is one that needs the players, but will still exist without them.

[Eldariel]

Another thing to keep in mind as well is to have a lot of 'white space' on the map. Having every acre claimed by the crown (or what have you) can be stifling. There wasn't much that I liked from 4e, but the idea of a campaign based around 'points of light' has really grown on me. Also, try to portray the power structures of the various communities as being competent but stretched thin. Showing town walls being repaired after the last goblin/orc/owlbear/umberhulk attack can do a lot more to motivate players than a courtly decree, and them seeing competent town guard/patrols can go a long way to convince players to help rather than letting Darwin sort things out.

Be careful with the competence angle as well. A community capable of defending itself (though perhaps at growing cost) is a community worth trying to save. A town guard manned by hyper-competent badasses begs the question of why they aren't solving their problems themselves. If you want the PCs to respect the town guard, don't try to make them better than the PCs, just show them as trying (and mostly succeeding) to save their people from the savage wilderness filled with murderdeath that is the typical fantasy world.

This feels a bit black'n'white; I don't think sandbox is necessarily well-served by a simple world. I don't think players should automatically be necessary since sandbox entails the players having a complete freedom of what they do in any given circumstance. Thus there need to be power balances where the players can choose sides - but in any sort of humanoid communities, those tend to form automatically anyways, so it's just a matter of fleshing them out enough to play it. It's not like humans are all on the same side. Even if the townsguard is completely impenetrable to normal outside threats, there may be local power struggles and the townsguard itself may be oppressive to its people (think Goldcloaks from King's Landing in A Song of Fire and Ice for instance). Or there may be multiple factions that split the seemingly strong front. There may be power behind the throne that is driving the town towards ruin or using it for its own nefarious purposes, or perhaps the local monarch is a total monster (figuratively). Just because the townsguard is strong doesn't mean they aren't overburdened by internal strife, local/continental crime, etc. Remember, in a world with magic, particularly inherent á la sorcerers and races like Aranea, crime is much harder to control and much more dangerous. And most threats to communities are like to come from within - the community's good is not in everyone's interest and not everyone cares even if logically speaking it were in their interests. And with creatures like Doppelgangers and Hags, it is fully possible for monstrous factions to undermine a faction.

A good example is Baldur's Gate. A rather open campaign in which you can undertake any number of quests though the Bhaalspawn story is of course inevitable, but the undermining of certain factions in the Baldur's Gate coupled with the threat of Amnian war moves the focus squarely to the powers behind the walls. It's not a simple matter of fighting off a Barbarian tribe (orc or human, little difference); politics and real world-changing events are constantly taking place and there are dozens of power factions with uneasy alliances and quarrels within each rickety of walls. Thus seemingly solid power may be gnawed from the inside and unable to act. Red Hand of Doom is another such case; the Vale has many powers and powerful defenders but they have no history of working together and they all have their own best self-interest in mind. Thus, if the players want to save the Vale, rather than directly engage the enemy, they're best-served building alliances and negotiating truces and agreements between belligerent factions. There's always room for this sort of content in any sandbox, but the players can of course always choose which side to play or whether to found their own side and try to come out on top of it all. The last tends to lead to the most memorable campaigns in my experience; playing all the major powers against one another and working behind the scenes to engineer a takeover.

[Darth Ultron]

What are your best tips and tricks for running a good sand box game?


Don't do it. It's a Trap(with a Sarloc at the bottom).

It's a great ''cool buzz word'' to say my game is a Sandbox, but it generally makes for a bad game..unless of course the game is really not a sandbox and you just say it is one. After all just look at all the so called Sandbox advice that says ''have a plot'' and ''have plot threads'' and such. A pure Sandbox is a pile of random mess, and you don't want that right?

So to run a good Faux Sandbox game :

1)Make a setting and the local highlights. Make at least three generic things of each thing that might be encountered so you can drop them in as needed(For example make three taverns that are not located on the map)

2)Make Plot Hooks....and this is basically what is going on in the world. If you really must do a Faux Sandy game, then you will need to make a LOT of them, like at least 25, and then dangle each of them in front of the players. Hopefully they will pick one...if not, make up 25 more.

2A)If the Players refuse to pick any plot hooks, then your stuck doing a Second Life bland random game. Sit back, and wait for the players to do something. If the do something, react...if not, just sit there. A lot of what the players will randomly do will be dull and boring, and you need to accept that...but hey you have a Faux Sandy game!

3)Once the players pick a Plot Hook to follow, start a Normal Game.

[Quertus]

Don't do it. It's a Trap(with a Sarloc at the bottom).

It's a great ''cool buzz word'' to say my game is a Sandbox, but it generally makes for a bad game..unless of course the game is really not a sandbox and you just say it is one. After all just look at all the so called Sandbox advice that says ''have a plot'' and ''have plot threads'' and such. A pure Sandbox is a pile of random mess, and you don't want that right?

So to run a good Faux Sandbox game :

1)Make a setting and the local highlights. Make at least three generic things of each thing that might be encountered so you can drop them in as needed(For example make three taverns that are not located on the map)

2)Make Plot Hooks....and this is basically what is going on in the world. If you really must do a Faux Sandy game, then you will need to make a LOT of them, like at least 25, and then dangle each of them in front of the players. Hopefully they will pick one...if not, make up 25 more.

2A)If the Players refuse to pick any plot hooks, then your stuck doing a Second Life bland random game. Sit back, and wait for the players to do something. If the do something, react...if not, just sit there. A lot of what the players will randomly do will be dull and boring, and you need to accept that...but hey you have a Faux Sandy game!

3)Once the players pick a Plot Hook to follow, start a Normal Game.

Well, you're not completely wrong here. Even IRL, a physical sandbox isn't just a collection or random toys. No, the toys are placed with purpose, allowing and encouraging the user to tell certain types of stories.

IMO, a "true" or "pure" sandbox is, simply, a living world. But most sandboxes aren't that.

Most sandboxes have a(t least one) preceding adjective, like a political sandbox, or a war-era sandbox.

And, even then, the sandbox isn't completely random - and it shouldn't be. It should contain toys that the creator believes that the user will be likely to find engaging.

So, yes, a sandbox is a matter of asking what type of story the player would enjoy telling, and providing them with a rich, optimized environment in which to tell such a story.

Having "a plot" isn't, IMO, the stuff of sandboxes. The "create lots of plot hooks" mindset is much better. However, if you put 25 toys in the sandbox, and the party doesn't play with any of them, I'd suggest having an OOC discussion with the players about what you thought would make for fun toys vs what they were expecting when they signed on to a _____ sandbox, before wasting your time making 25 more toys.

But, remember, if you have a sandbox, it's the players who are playing with the toys, and telling the story. Let them tell whatever story they want. Don't pre-script the story.

[Hecuba]

If someone wants to point me to a thread that poses this question please do, I didn't find one.

What are your best tips and tricks for running a good sand box game?

My contribution: I have a pad of graph paper full of mini dungeon designs, warehouses, taverns, castles, etc., as well as individual puzzle rooms and trap rooms, that I can flip through and use on the fly.

Whenever I'm inspired and or have down time I try to add to the note book to keep expanding my options.

In my experience, running a sandbox is first and foremost an exercise in world-building. You need to build politics rather than plots. You need to give countries histories instead of hats. You need to give governments agendas. To generalize, you need to understand the motives of the actors that make up the setting.

And you need to, above all, be familiar enough with what is built to understand results. Because, above all, a good sandbox requires that the setting react back in a believable manner.

This is not to say that you cannot start with an existing setting. Indeed, for many settings you will find that some of the work is done for you in the supplemental books. But you will need a very good grasp of the setting.

To use an isolated example from Eberron, in a more conventional campaign House Orien can often just be treated as a taxi service for hire and occasional quest source. In a sandbox game, that probably won't fly. You need to know that they have interests in maintaining the availability of Aerenal soarwood - which means that they could bring their resources to bear if the party were to become disruptive in Aerenal.

[Quarian Rex]

This feels a bit black'n'white; I don't think sandbox is necessarily well-served by a simple world. I don't think players should automatically be necessary since sandbox entails the players having a complete freedom of what they do in any given circumstance.


Keep in mind that this was advice to the OP for starting his first sandbox. Having a rat's nest of political factions can be engaging as all hell but it is generally a poor place to start. Too much backstory, too much double-dealing, too much backstabbing, all being dumped on the PC's too fast. If you want to run a good sandbox you really need to establish some ties between the PCs and the world. Having them start in a place where their strengths are needed and their efforts (even as lowbies) can make a difference is critical for that.

Remember that while the characters grew up in this world/kingdom/hamlet/whatever and have all the emotional baggage that that entails, the players don't. In my experience, it doesn't matter how well written a characters backstory is, what happens in play trumps it. Engage the players on their level to make them care about the world. As the PCs gain power they can stumble on more involved plots, or come to the notice of some of the more involved movers and shakers of your world, but starting them there is usually a recipe for destruction. To continue with your Game of Thrones analogy, this can go about as well as it did for Sansa when she was dropped into King's Landing. She spent a looong time being nothing more than a manipulated tool for those of greater capability that her, incapable of even grasping the true extent her ignorance. While that sort of thing can make for a mildly entertaining read (Sansa's arc was no ones favorite for the majority of the series) it can be absolutely deadly for a campaign.

I'm currently playing in a sandbox-y campaign now that tried dropping us into the thick of it from the beginning. It has essentially turned my character into a murderhobo. I am not a murderhobo type player. This is the first that I have played, and not by choice. It was just the most logical course of action based on the world around us. That is the sort of thing that happens when you don't take the time to ground the PCs in the world.


I don't think sandbox is necessarily well-served by a simple world.

I don't think the world should be simple, but I think that it should (functionally) start simple. Simple problems can be handled by lowbies alone and lead to simple victories that have value and can shape characters views on the world. There is always more complexity behind the surface, I just don't think that's a great place to start.


I don't think players should automatically be necessary

Note that I said, "A good sandbox world is one that needs the players, but will still exist without them". The players are needed, not necessary. The world will continue without them in a sandbox, but there should always be places where their efforts will make a large difference. Showing that right from the beginning can shape the tone of the entire campaign. Showing them that they are ineffective and in over their heads from the very start can shape things as well, just not in the best of ways (murderhobos).

[Rizban]

The main thing to consider is goals. Some parties are okay with bumbling around not trying to accomplish anything in particular, but most are not, and will fall apart without any clear goals. Either make each player declare one or more personal goals they want to work towards, or make outside quests easily accessible (my sandbox includes a bounty board in the Adventurer's Guild for an easy selection of quests), or both. (These goals don't have to be very related to the "main plot", if any. My sandbox has a "main plot" or two, but they're not as easily accessible, and need to be stumbled upon through the natural course of play; a couple of parties have indeed stumbled on the peripheries of these plots.)


Make it absolutely clear to your group that you aren't going to lead them by the hand, so it's important that their character have goals of their own, that are at least somewhat compatible with other people in the party, at the very least with a mutual assistance in personal goals being an understanding.

Provide possible hooks, but don't force the party down any of them. Don't get mad if the party doesn't decide to pursue one of your presented hooks.

I have played with people where an open world just does not work at all; if given a choice, they just clam up completely or argue about what to do for 2 hours.

This is what I came here to say. Have clearly defined goals available in addition to those that you have the players come up with during character creation. Throwing out lots of plot hooks and letting the players go where they will is the key to running a good sandbox game. After they get a bit of momentum, they'll (typically) run with the freedom you give them and start coming up with their own plots and quests to pursue. Whatever they decide to pursue, go along for the ride. Have a selection of ready made events and NPCs to plop into the story as they go for your own convenience but otherwise just enjoy it by that point.

[Yahzi]

What are your best tips and tricks for running a good sand box game?
Spend two years writing an app that will generate an entire continent, complete with random encounters, kingdoms, rulers, political institutions, and army lists. Then, push a button, double-click on the map a few times to zoom in on a likely starting area, and roll with it.

Luckily, you can skip the first step, because I already did it: Sandbox World Generator (http://www.drivethrurpg.com/product/217951/Sandbox-World-Generator). It does use my weird XP system for calculating the levels of the rulers, but you can completely ignore that and just use the output as is, or edit it to your heart's content.

The next thing you do is whip up a few mini-adventures and keep them in your back pocket, to be deployed whenever the players seem insufficiently busy. Then you study the politics of the local area and make up a conflict, usually a war about to happen.

After that, it's just watching your players trying not to die.

[RoboEmperor]

Example1:
1. Have a bunch of CR appropriate encounters setup with different monsters.
2. Throw an encounter at the PCs with the correct type of monsters.

Example2:
1. Grab an adventure path.
2. Run the adventure path with the PCs
3. If the PCs start to deviate from the plot and perhaps even completely abandon it, grab another adventure path and see if the players get hooked into it.
4. Repeat 2 and 3 until PCs die.

If you are capable of creating your own adventures, then replace adventure paths with your own original adventures.

Sometimes a player starts an adventure by himself. Some examples include:
1. I want to punch Pazuzu in the face! At which point you just read Fiendish Codex I and setup encounters involving every demon in that book.
2. I want to hunt for a tarrasque so I can enslave it.
3. I want to start a revolution so I'm going to start recruiting and copy Amon from Avatar Korra.

etc. etc.

[Jay R]

I've never wanted to run a sandbox game. Back when I wanted to create a world, in which I pretty much knew what would be found pretty much anywhere, It made for a very successful sandbox.

Similarly, Tolkien wanted to invent languages. He eventually realized that they needed to be set in a history and culture to be complete. This turns out to be a great base for stories.

[atemu1234]

Honestly most people trying to run a sandbox are attempting to knee-jerk in response to hearing about a railroading DM.
Basically, keep this in mind: A little bit of railroading is okay, so long as the players don't see the rails.

[RoboEmperor]

Honestly most people trying to run a sandbox are attempting to knee-jerk in response to hearing about a railroading DM.
Basically, keep this in mind: A little bit of railroading is okay, so long as the players don't see the rails.

There's a difference between a DM trying to get the players into the adventures hes planned and a DM telling the players to shut the **** up and do as he says because he's the DM and he can kick you out of the party.

[Zanos]

Honestly most people trying to run a sandbox are attempting to knee-jerk in response to hearing about a railroading DM.
Basically, keep this in mind: A little bit of railroading is okay, so long as the players don't see the rails.
"Railroading" is fine as long as you're upfront about it. If the DM says he wants to run an adventure in the Kingdom of Ant People and the players sign up for that, and the first thing the players do is teleport away, that makes the players jerks, not the DM.