Dungeons & Dragons players don’t love sandbox adventures as much as they think, but such adventures can still offer fun. After I took aim at sandbox adventures, some dungeon masters running thriving open-world campaigns offered counterpoints. Michael S has run such a campaign for 9 years, outlasting the West Marches campaign that inspired him by 7 years.

Sandboxes can work. This post and my next will reveal the secrets. Once you read all these requirements, you may decide a fully-realized sandbox demands more time and energy than you can spare. Michael’s “crazy” world uses a database just to track NPCs and a “big” wiki for bookkeeping. (Of course, rather than filling your sandbox in advance, you can cheat. More on that next time.)

The hardcover adventures published for 5th edition have all tried to include some of the freedom of sandbox adventures. None of the hardcovers check all the requirements on this list. If the hardcovers proved frustrating to run, the missing requirements explain why. If they played fine for you, then you, as a good DM, improvised and reworked to fill in the gaps. You may have enjoyed the extra effort or, like me, you may have done it reluctantly. I have a D&D blog to write.

The sandbox archetype casts characters as explorers on the verge of a unmapped frontier, perhaps the shore of The Isle of Dread or descending from the Yawning Portal into the Undermountain megadungeon. When Out of the Abyss stranded the party in the Underdark, the hardcover followed the exploration model. The authors of the other hardcover adventures touted sandbox-like design, which ranged from player-led tours of the Sword Coast to simply including dungeons that could be visited in different orders.

My advice starts with this exploration model in mind, but these requirements apply to campaigns that stray from the prototype.

Set long-term aims

In the sandbox archetype, players start with a incomplete hex map like the ones in The Isle of Dread or Tomb of Annihilation and explore to fill the blanks. Such games can become unsatisfying grinds. The Angry GM explains, “The only rewarding part of the exploration is when you actually find anything. And the vast majority of the hexes have nothing to find. There is nothing interesting about yet another forest of hexes. And random encounters—dinosaurs waiting to jump out of the forest and kill you—are interesting, but they are the painful, dangerous kind of interesting. That’s not a reward.”

The 5E hardcovers all avoid this pitfall by setting a campaign objective from the start. In Tomb of Annihilation, players seek the source of a world-spanning curse. In Out of the Abyss, players explore to find an escape from the Underdark. The other adventures reveal some overarching menace in the opening scene.

Most players want to aim for an ultimate campaign goal. Throughout the game, they want clear options that take them closer to achieving their long-term aim. In Curse of Strahd, the players start far from defeating Strahd, but they can explore Barovia and gather the magic items they will need.

Not every campaign leads to some ultimate villain. Instead, for instance, a party could aim to find a lost heir and put her on the throne.

D&D adventures either set an ultimate goal for characters or assume they quest for treasure. (An adventure that omits any goal becomes a campaign setting.) Players seldom mind adopting a goal so long as they get to do D&D things like collecting treasure and smiting evil. When I ran Murder in Balder’s Gate though, my players rebelled. The adventure assumed characters would support one of three patrons who vied for power. The patrons start unsavory and, as they gain power, become worse. My players wanted no part of it. My Murder In Baldur’s Gate became something entirely different from the book.

If you introduce a long-term aim before character generation, your players can craft characters tied to a chosen aim. For example, Princes of the Apocalypse suggested players tie their character to an adventure-specific hook.

In a homebrew campaign, you start with the players’ characters and invent a long-term aim just for them. Perhaps a someone wants to play a member of a hiding royal family. Who can say how close she is to the lost heir?

Give enough information to start

Imagine a party shipwrecked on the coast of a lost continent. If they know nothing, they can only trudge inland and explore aimlessly. Perhaps a wandering monster will interrupt the drudgery.

Robert Conley’s Bat in the Attic blog focuses on sandbox play. He writes, “Picking one of the six surrounding blank hexes is not a choice with meaning. So work on the initial situation so that it is interesting and give the players enough information to make some valid decision of what to do.”

If that shipwrecked party has an explorer’s letter, a ciphered treasure map, rumors a lost city on a plateau, and so on, then the group knows enough to plot a course. Once you tie the lore into the characters’ goals, the clues will inspire action. Remember though, if you give one player a clue, they might not share it. Spread any secrets that you need players to share.

You can also launch a sandbox with a conventional adventure that starts with a patron in a bar. No one said your campaign requires a consistent model.

Out of the Abyss launched the characters’ explorations by pairing them with Underdark natives to serve as guides. See The surprising benefits of giving an adventuring party a guide.

Next: Running a sandbox campaign