- Gnome Stew - https://gnomestew.com -

Dynamic Sandboxes

Sandbox [1] games often get criticized for being static and devoid of any real plot. To use a stereotypical example, the player characters start in a town that’s surrounded by the Caves of Doom, the Tower of Death, and the Dungeon of Discomfort. It’s up to the PCs whether they want to explore each of these and in what order. Alternatively, the PCs may decide to hang around in town and harass the citizens, or they may simply wander around the map and tackle random encounters.

The locations themselves also tend to be static. The Caves of Doom always contain a clan of kobolds, the evil wizard of the Tower of Death continues to sip tea until the PCs arrive, and even the closest rooms to the entrance of the Dungeon of Discomfort are never cleared and looted until the PCs get there. In some cases, especially in class and level systems, the challenge level of a location and the PCs’ power level can be a jarring mismatch. A high-powered PC party could cake-walk through the Caves of Doom, but a low-powered PC party could get decimated by the statue guardian of Room #3 in the Dungeon of Discomfort, to say nothing of the avocado dragon that rules it.

This can also lead to “sandbox apathy,” where the players lose interest in the sandbox world as nothing changes. Sure, they cleared the Dungeon of Doom, but the evil wizard and the avocado dragon just yawn and sit in their strongholds until the PCs come after them. The PCs simply become game pieces that are moved around the board until there is nothing left to do.

One defense of this style is that “testing the waters” of each location would guarantee that the PCs will opt to clear the Caves of Doom first, then handle the middle-powered Tower of Death, until finally they tackle the Dungeon of Discomfort. If this is the case, it begs the question why the Game Master just didn’t run the adventures in order. Most sandboxes don’t contain enough nuanced locations so that a beginning PC party could tackle the hobgoblin-infested Lair of Destruction first, although they’d have an easier time of it had they hit the Caves of Doom beforehand. It also doesn’t address the related problem of what “fun” it’s going to be for the higher level PC party to torch the Caves of Doom for little effort and weak reward.

To be fair, there are a lot gamers that are perfectly content with this style, as they imagine their uber-level PCs sitting around a fire, telling tales of how they tore the Staff of High Wizardry from the dead wizard’s hands, and used the Blue Blade of Vengeance against the avocado dragon in order to gain their current power level, and they actually had a grand time showing those pathetic kobolds who was boss when they marched in, kicked tail, and took names. This article obviously isn’t for you :).

Once way to keep sandboxes fresh is to create dynamic locations. The world keeps moving as the PCs make decisions. Taking a page from linear adventures, you can craft a story around each location that unfolds as the PCs act, no matter where they are, and ensures that they get a different experience depending upon when they decide to visit that location or deal with a particular plot thread. The trick is to make it easy on you so you aren’t overburdened when preparing the campaign.

Here’s an example using a post-apocalyptic setting.

The Greenfield Campaign

The PCs are nomads traveling through the blasted landscape of the old “Heartland” when they discover Greenfield, a relatively intact walled town run by an elected Governor, whose power (the Greenfield Authority) extends throughout a relatively fertile valley. The PCs learn of four problems that need attending:

The PCs now have four immediate plot threads. They can follow any of these in any order they wish. That said the world keeps moving, so the GM has predetermined how each of these threads might evolve without PC interference.

Bomb Shelter Example

To use one location, the Bomb Shelter, as an example, it’s easy to see how the challenges change as the campaign progresses. At first, the PCs are facing small bands of ferals and their mutated hounds. If they clear the bomb shelter at that point they discover some unused pre-war foodstuffs and equipment, and information that this facility is more than a bomb shelter. They may then spend time dealing with the other problems or explore the “dungeon” that leads to the Vault, ultimately uncovering the war machine (which they may activate in the hopes of using it against Waterford – oops).

If the PCs wait and follow other leads, then they may learn of the Ragtag unit’s change of direction. The PCs may intercept them en route or attack them while they lay siege to the ferals. If they don’t then the Ragtag unit moves in. At this point the PCs may strike an alliance with displaced ferals that can help them infiltrate the shelter or simply attack a better-reinforced shelter. They may even try a diplomatic approach, perhaps convincing the Ragtag Colonel that it’s better for them to throw in with Greenfield instead of Waterford (the Captain would love this – she may even seek help from Waterford for her coup). If they still leave the Bomb Shelter alone, then the PCs have no chance of stopping the War Machine before it starts destroying everything in sight.


Dynamic sandboxes keep things moving and make the players feel like their choices matter without being constrained. It’s not so much that a sandbox doesn’t have a plot, but rather that the PCs are simply causing ripples in a larger pond of plots, the conclusion of which can be just as, if not more so, satisfying than a more linear campaign.

18 Comments (Open | Close)

18 Comments To "Dynamic Sandboxes"

#1 Comment By Noumenon On December 5, 2011 @ 4:18 am

I wish D&D could use the Dragon Age: Origins approach of saying “There’s four dungeons, the one you go to at level 1 has 1st level monsters, the one you go to at level 5 has 5th level monsters, and so on” — but D&D has that printed monster manual saying “No, there’s no such thing as CR 11 werewolves, forget about it!”

#2 Comment By Svengaard On December 5, 2011 @ 5:48 am

@Noumenon: What’s to say that can’t be done? I’m sure it can’t be hard to boost stats on kobolds in your campaign. In Walt’s situation, I bet having an army of kobolds invading the town and taking names will keep the players from dismissing a kobold threat ever again. I hardly ever crack open my Monster Manual anymore. Even if I bought all the Manuals and campaign specific monster guides I’d still feel they were restrictive.

#3 Comment By dyatil On December 5, 2011 @ 6:03 am

Excuse me, how do you write such a dynamic setup down? Is it a mindmap or a block-scheme or just long “if… then…” list?

#4 Comment By Tsenn On December 5, 2011 @ 9:02 am

Apocalypse World in particular is built with the dynamic sandbox idea front and centre. It uses ‘countdown clocks’ to help the GM pace the encroachment and advances of the various threats the characters may face. Leave a threat to fester, and something will happen by itself. However, the idea is to keep the characters in the spotlight, so these threat actions can’t simply occur in a vacuum. They impact on people the characters care about, work with, and rely upon (and the characters themselves where appropriate). Nobody may care when Raidertown burns Farmville if nobody’s been there, but when you realise Farmville provided most of the food for Characterburg, and now *your characters* are starving to death, that’s a motive to act. A good sandbox strongly relies on motivated characters, and should be capable of providing realistic motivations in a fashion that seems natural. Just my two cents.

#5 Comment By Tomcollective On December 5, 2011 @ 10:53 am

Call me a snob, but I’m not sure why this idea wasn’t obvious until recently.

Dynamic sandboxes are also, in my opinion, where rpgs shine. I love playing and running interactive stories.

As far as leveling monsters, I think there are rules for it. But either way, I say do it however you want! Its you’re game and the rules police wont get you.

Mapping such games is a simple matter of “what would the villain do?” You can’t if/then everything. But you can always react to your players. Just look at the situation from the villain’s perspective and the game writes itself.

#6 Comment By Walt Ciechanowski On December 5, 2011 @ 1:24 pm

[2] – First, I’d key the locations/situations as if the PCs were hitting them first (since you don’t know which way they are going to go), then I’d bullet-point the progression of each location, making modifications along the way based on the PC’s actions.

It’s easier to change keyed locations, since you’ve already statted out the NPCs. You also now have layers of history in your location (using the Greenfield Bomb Shelter example, there’d still be signs of the former Feral occupation after the Ragtag unit moved in. If nothing else, the ashes of funeral pyres would be outside).

#7 Comment By Walt Ciechanowski On December 5, 2011 @ 1:24 pm

[3] – You aren’t a snob, just an experienced GM.

#8 Comment By Troy E. Taylor On December 5, 2011 @ 1:36 pm

Ooooo, sandbox gaming. A favorite subject. As you say, having dynamic touchstones (encounters both interactive and combative) is the key.

#9 Comment By evil On December 5, 2011 @ 4:28 pm

[2] – I do a few things to set up my games.

First, I make a grid with the major enemy names across the top and phases of the game down the side. Then, I fill in the appropriate information.

I do the same for the major locations and plot points. After those are filled in, it’s just a matter of connecting the points and times in the game.

As you play, just cross off the areas of the grid that have been used or can no longer be used (for instance, if you wiped out Walt’s kobolds in the first example, just cross out that entire column).

#10 Comment By Kurt “Telas” Schneider On December 5, 2011 @ 5:02 pm

I smell another article about the easiest ways to prep the dynamic sandbox.

That, or my youngest daughter’s diaper needs changing.

#11 Comment By Martin Ralya On December 5, 2011 @ 7:59 pm

I’ll second what Kurt said: I’d love to see a follow-up about prepping for this style of game.

I don’t do the kind of prep that would be needed for this:

“The PCs now have four immediate plot threads. They can follow any of these in any order they wish. That said the world keeps moving, so the GM has predetermined how each of these threads might evolve without PC interference.”

…but it sounds like an amazing game to play, and not a hard one to prep if you’re better at prepping than I am.

#12 Comment By RevGunn On December 5, 2011 @ 10:18 pm

[4] – This is pretty much the style I’ve been GMing since I became an adult. I’ve got a Boot Hill 3rd Edition game going now that’s done this way.

Each major NPC has an agenda and an event timeline. All the minor NPCs have links to one or more major NPCs, either as allies or competitors. Or double agents. This style of game does intrigue really well.

I’ve got it all set in a gold mining town just about to boom. The City Council, miners, railroad, ranchers, outlaws, law-dogs, everyone has something going on. I’ve even got a newspaperman who started a Vigilante group. He’s a Marxist/Anarchist?Terrorist, but all the PCs think he’s an ally.

At this point, the Territorial government is getting involved as well.

Basically, the “web” of plots mentioned above is the key. Tomcollective is 150% correct. You simply play the NPCs as they would act, based on the character and it writes itself. Some sessions, its like I had nothing to do with it, stuff just happens.

Key out your locations, decide who lives where, and go.

#13 Comment By Bercilac On December 6, 2011 @ 7:04 pm

Two minor things:

“Alternatively, the PCs may decide to hang around in town and harass the citizens, or they may simply wander around the map and tackle random encounters.”

Yes please.

“scouting reports show that the soldier’s path won’t take them into Greenfield”


More useful things:

For the kind of faction-oriented sandbox that a few people have mentioned in the comments (@RevGunn and @evil) the GM advice in Dogs in the Vineyard is useful reading. What I like about it is that it ties all of the NPC plots together (linking them to a single, or small set, of underlying causes), but it refrains from saying what the main problem is, or how it should be solved: these issues are left up to player interpretation.

I think it’s useful to differentiate between two kinds of sandbox dynamic. Roughly:
-Dynamics which change the setting. The coup in Greenfield is an example. Yes, the setting changes, but the players could potentially ignore it (until some other set of objectives makes them cross paths).

-Dynamics which directly involve the players. You haven’t dwelt upon these so much, but the warmachine that destroys pretty much everything might be one. I’m thinking here more in terms of using PC background to create character-specific hooks. If the players are willing to ignore the evil necromancer, will they still be willing to when a PC’s cousin is captured for use in the wizard’s dark laboratory? I think of this as a sandbox full of bear traps. You can go where you will, but eventually something will grab hold of you, whether you like it or not.

#14 Comment By BryanB On December 7, 2011 @ 3:57 pm

I don’t tie things in my “Dynamic Sandboxes” down to locations. There might be locations of course on the map and I do know where the NPCs have their people and their stuff, but location is not the real anchor.

The anchors for me are the NPCs and their motivations. If you can figure out what the NPC wants and why they want it, then you can let the chips fall where they may. The players can become opposition or support for whatever the NPCs want. The players can also drive the motivation through their actions or inaction. In a dynamic sandbox, the players chosing not to do something might have just as much or more of an impact on the campaign than if they do something.

I tend to create a who’s who NPC matrix for the game. Who are the movers and shakers? What do they want? What are their goals? Are any of them enemies? Are any of them allies? Enter the PCs and their desires and see what happens. Perhaps the PCs will tip the balance of power one way or the other. Perhaps the PCs will just get rich of both sides. Or dead. 🙂

The interactions of PCs and NPCs and their conflicting or aligning goals and motivations are what makes the RPG engine purr along at full speed. The Dungeon of Despair, The Castle of Weeping Walls, or the town of Dragonthatch are for cool scenery and the potential locations for those conflicts.

#15 Comment By BryanB On December 7, 2011 @ 4:00 pm

Or in other words, I’m in line with what RevGunn & Tomcollective are on about. 😀

#16 Comment By RevGunn On December 7, 2011 @ 9:20 pm

[5] – “Sandbox full of beartraps” Brilliant! Its that exactly.

#17 Comment By Joah_from_Alberta On December 11, 2011 @ 12:02 pm

I really appreciate reading something like this. For *years I have been promoting openended-play and it’s nice to see how the community is becoming enlightened.

When using a blank canvas, people at first feel daunted by the blankness and don’t even get past the first glance. Once you do start, it is important to have a vision for where it’s going.

It is important to remember the three prime factors within theatre: stage, character, story (any order). To retained this trinity is imperative to the theatre (or RPG in this case) and I find it advantageous to picture a triangle and its three corners.

I believe factoring is crucial lest you grow burdenned by the details. By pairing down your system you can form equilaterals without any effort, they will simply be there at all times and brought out in the development when needed. There is none, there is one and there are many, this keeps to the trinity.

When you don’t know where to begin, you can ask questions, what is the motivation of the character? What are the needs of the villain? Once you perceive this then you can begin to see the stage, character and story for what it really is.

Anyhow, enough jedi. I have been developing a sandbox system based upon the concept of organizations in D&D. By having the organizations develop in a setting, a story forms based upon the interactions of the relevant organizations. It is a bit structured, but again I want to keep things factored as much as possible. This is why I have decided to work from “the ends shall justify the means”. By developing from the end, I can better fill in the values in the beginning (where the show starts). So my organizational paths will seem to permutate but in reality it is just a tangled traingle leading to one point.


#18 Comment By Dante Inferno On April 5, 2012 @ 10:41 am

This is always how I’ve run Sandbox games, but I called it a “living world”. I find it easy and intuitive.

I had no idea until recently people didn’t always do it this way. I can’t imagine playing in a static-MMO type world where NPCs just sit still until someone interacts with them. I always figured CRPGs more or less did that due to limits of the technology not because anyone LIKED it that way.

I personally don’t write very much down and I never any flowcharts or if-thens. I focus instead on the NPCs involved, their motives, capabilities and resources (including organizations they control, armies, money, whatever). Then I just have them act and follow things to their logical conclusion if the PCs don’t interfere.

I focus on NPCs, organizations, and the like rather than on locations or events. People are what make things interesting.

Done and done!