I wrote before about why a game with a narrow premise is a good idea, which included some reasons why a sandbox game is problematic. They can be overwhelming for both players and GMs, paralyzing them with possibilities. That’s not to say sandbox games should never be played, which is what this post is about: what I think makes a good sandbox game.

What is a sandbox?

First, we need to figure out what a sandbox game is. There are a few variations on the definition, but for a tabletop RPG it boils down to a few elements:

• The world is large and detailed enough to allow free (or realistically limited) exploration and interaction;
• The characters have a firm grounding in the world; and
• The story is guided entirely (or almost entirely) by the players’ decisions and actions.

Really, those elements are good for any game, but they’re of particular importance to sandbox play. The emphasis is on the world as a mutable, changing thing in which actions and consequences are carefully linked and in which the players can feel like they have a real effect on how the world changes. It’s the opposite of a railroad game, where players can make small decisions that affect the immediate outcome, but the larger narrative doesn’t shift accordingly.

So, why play (or run) in a sandbox?

Player agency is king.

Part of the appeal of tabletop RPGs compared to their electronic counterparts is the sense that you can go anywhere and do anything. It’s the reason I agreed to try D&D 3rd Edition back in middle school: I pictured my hero wandering an endless world, doing good deeds and making a difference. The first session didn’t disappoint, with the fighter Gray Althen boarding a ship in port with a motley crew and journeying to a far-distant castle to retake it from its orc invaders.

Granted, he immediately got hit by a ballista from a pirate ship and almost died, but that was part of it. The world was big and dangerous and challenging, and I wasn’t always going to succeed. What made it okay to fail was the fact that I was failing based on my own choices.

In the sandbox, you can have a brief non-player-driven story at the beginning to get the characters introduced to each other, maybe introduce the world as well if nobody is familiar with it. After that, it should be up to the PCs where they go and what they do. Their choices, and the consequences of those choices, guide the narrative.

The downside of this is that such absolute control can be intimidating, especially for new players who are used to some structure and might be paralyzed by indecision. A good way to limit that is to give the players a premise for being together: part of a band of mercenaries, interstellar traders, anything that gives them cohesion and direction without also forcing some expectations on them. They shouldn’t play as part of a military organization where they have to follow orders, for example; it’s better if they’re a group of bounty hunters who can choose what jobs they take and why.

This can also be intimidating for the GM, but we’ll discuss that in a little bit.

Backstories can guide the narrative.

If you’re going to do a sandbox, you need to ask for some kind of backstory from your players. They might be understandably gun-shy about doing so, since plenty of GMs either ignore their backstories outright and invalidate the work they’ve done, or else use it as so much rope to hang them with.

But in a sandbox, the PCs can’t exist in a vacuum. A proper backstory introduces plenty of hooks for the GM to craft engaging NPCs as allies and rivals, as well as settings for the world. To help, please allow me to share my handy rules for backstories:

1. You are not an orphan. Even if all of your parents/siblings/family pets have been brutally murdered or lost to the mists of time, you’ll still have a support network in the form of friends, mentors, and allies. Think about how they helped you, made you the way you are, and what they’re up to these days.
2. You have a home and probably a job. Even if it’s not where you grew up, there’s still a place you consider your hearth, where you’re welcome anytime you care to show up. You might even have responsibilities and relationships there, which means you have to go back semi-regularly, and you’ll be missed if you’re gone for too long.
3. You are not a rank novice. Even if you’re a level 1 Fighter, you had to be trained. Sorcerers go to school, bounty hunters need to qualify for their licenses, shadowrunners have to build up reputations. Maybe it hasn’t been world-shaking or galaxy-changing, but you’ve already been on some adventures and proven your skills, at least to yourself if not to the people around you.
4. Your experiences are unique. Not everyone who has walked the path of a ranger has started the same way. You’ve seen things no-one else has, been to planets and talked to people your fellow PCs don’t even know about. The experience might have changed you, or it might have reinforced your life choices. It might have been a chance exposure to a god walking the earth, or maybe it was just a kindly bar patron. Whatever it was, it was meaningful to you.

I can’t stress enough that if you’re running a sandbox game, or even planning on introducing elements of sandbox gameplay, you should run a Session Zero. I’ll do a post about Sessions Zero in the future, but the idea of creating your characters, a handful of NPCs, and one or two important locations together as a group makes the world much more personal for the players.

The story is what they make it.

All right. Let’s get something out of the way: just because the players are the ones responsible for deciding what they do and where they go doesn’t mean the GM can get away with doing less (or no) work. If anything, a sandbox means you need to do more work for the sake of the story while also being mentally prepared to drop a plot-thread idea at a moment’s notice.

If you don’t have a reasonable expectation of where the players are going next, then you need a stable of location-agnostic ideas, and probably a handful that are location-specific, just in case. If you play the Star Wars RPG, a quick and dirty but also fun solution is to hit up @SWRPGAdventures on Twitter. Otherwise, it’s a good idea to mine your favorite TV shows (individual episodes) and written form entertainment (short stories and individual comic book stories) for ideas. Movies can be problematic, since their underlying premises can be the stuff of entire campaigns.

This list of ideas acts as your toolbox, so that no matter where the players go you have an idea what can happen. Make no mistake: the players will often be driving the plot when you’re playing in a sandbox. Your job will mostly be to react to their actions. However, that doesn’t mean they have control over everything that happens, which gives you the chance to call upon your knowledge of pop culture and/or storytelling in order to keep things happening around them.

The world is a character, too.

It goes without saying that any sandbox-style world will need to be full of characters. They don’t all need to be interesting, but you need a handful in most locations that are compelling, that can help move the plot along when the PCs come knocking.

However, the key character in a sandbox game is the world itself. It needs to react to the actions of the PCs as well as be proactive, as if it existed independent of the people playing the game. It needs to celebrate, mourn, grow, break, rise, fall, thrive, dream, strive, and fail. It needs to do these things when the PCs prompt it and even when they don’t. It can be as intimate or as expansive as you need it to be.

There’s only one thing it can’t do: succeed without the PCs.

There cannot be way for the corrupt king to be dethroned except by the actions of the PCs. If there’s a key to destroying the galactic superweapon, you need to get it into the PCs’ hands. Economies can collapse, stars can blink out, allies and mentors can die with or without PC intervention, but you should never, ever allow for a path to victory in which the players do nothing.

Remember: the PCs are always the protagonists, and the protagonists are always necessary to success. Your world should reflect, support, and challenge that core assumption, but it can never refute it. Because, in the end, whether you’re playing in a sandbox, on a railroad, or somewhere in between, it’s the characters who make the story happen. The GM’s job is just to make sure that there is a story that can happen.