A recent blog post by “Jester” David on problems with the Dungeons & Dragons hardcover adventures made me consider the use of sandboxes in D&D adventures. A lot of people talk fondly of sandboxes, but I’m unsure how many like the examples we’ve seen in official adventures. There’s a reason for that: Sandboxes, certainly in the context of a narrative story, are hard to design.

There’s probably no greater example of that that Princes of the Apocalypse. Let’s make this clear: I think Princes has a lot of brilliant design in it. However, as an adventure that is almost entirely sandbox, it runs into a lot of problems without the DM rewriting and reworking its material. The transition from the “finding the delegates” to “stopping the cults” should probably be explicit, but it’s an implicit assumption about how the players act. (And, of course, the delegate thread just keeps hanging around for entirely too long). Also, writing a sandbox for levels 3-15? That’s hard.

David calls out several adventures which have a sandbox that the players are expected to live in for several levels, marking out time until they’re ready for the next section.

In the case of Out of the Abyss, I find it hard to disagree with him. And this is entirely due to how this section ends: The DM decides you’re of the correct level, so the next escape attempt you make succeeds! I don’t have a problem with the sandbox as such. The players have a goal, and they explore the environment to discover how to complete that goal. And the exploration of the environment and the discoveries you make about the situation are key to the feel of the adventure. The Demon Princes are loose! It’s bad for everyone – what happens if they reach the surface! With that knowledge you’re primed for the next section of the adventure – stopping them.

However, because it doesn’t matter what the players do, as they can never escape early even with superior play, their accomplishments are undermined. That’s a type of design I dislike.

In the case of Curse of Strahd, I see the adventure in a different light. This is an example of a successful sandbox, although it’s entirely possible for the DM (or players) to get it wrong. In Curse of Strahd, the players are sent on a quest to find the items they’ll need to defeat Strahd. The range of the sandbox is levels 4-10, but here’s the trick: once the items are found, the players can defeat Strahd even if they’re not level 10. They have multiple quests, and so if one seems too difficult, they can try another. They can also find allies (apart from the items) to help them. How does the DM get running this one wrong? By not including the Tarokka reading. You need to give the players the quests.

This isn’t marking time. The PCs can confront Strahd at any time. They just should complete the quests first. I’m pretty sure a party of 6^th level characters can defeat Strahd with those items!

It’s not much of a secret that I don’t like the sandbox section of Tomb of Annihilation. It’s the favourite section of a lot of my friends, and I can see why. There’s lots of range to add your own material and build up what is suggested. Why don’t I like it? It’s because it’s very obscure as to the path taken through it. Your goal – find what’s causing the Death Curse and stop it – isn’t supported by most of the encounters, and, if you choose the wrong path, you don’t even get to meet the few people who know anything about it.

Is this marking time? Does the adventure not proceed until you reach the proper level, or does it just expect you to reach close to that level by the time you find Omu? I think the intention is the latter, but I think it’s likely that a lot of groups will feel lost in the investigation with no idea of how to proceed.

David isn’t much of a fan of the sandbox in Storm King’s Thunder. I think he mischaracterises it as lasting for three levels. In fact, it lasts a solitary level. You start SKT at level 5, and help defend a town that is attacked. You’re then given a few rewards – which are quest hooks – by the surviving NPCs. The sandbox is level 6, where you undertake those quests and discover that there are lots of giants about. At level 7, Harshnag turns up and takes you to the rest of the adventure.

I’ve seen a lot of DMs run SKT badly. It’s not helped by the first section of the adventure, “A Great Upheaval”. In my opinion, the real intention of the adventure is that you’re regular adventurers having adventures in the Sword Coast. In one of those adventures, you help defend a town from giants, and then become aware of the giant problem before you’re launched on the quest to stop it. “A Great Upheaval” introduces the giant threat too early, and makes it seem like you’re meant to stop it, before dropping you in the sandbox with no clues as to how to proceed against the giants. The structure of the adventure works better if it’s only when Harshnag appears that you’re given the quest to defeat the giants.

The big mistake is to drag out the exploration of the Sword Coast. This is something you can end at any time: you have Harshnag find the characters. What makes this different to the end of Out of the Abyss is that it’s a DM-directed hook to the next section, rather than the DM just saying “you finally succeed”.

Tyranny of Dragons isn’t written in a sandbox manner. (Its chapters give characters a lot of freedom of how to get to their goals, however). Waterdeep: Dragon Heist gives a brief nod to a sandbox, but it’s fairly tightly plotted. Waterdeep: Dungeon of the Mad Mage is one big sandbox, but it’s of the old megadungeon style, which needs no goals save what the players find there. They’re going to get treasure!

So, where does a sandbox work for me? Generally, it’s when it’s written where the players enter it with goals, and they can gain additional goals as they play through it, all of which lead towards the conclusion or next stage of the adventure.

There’s a couple of different forms of sandbox. I’d characterise Princes as starting with an investigative framework, the same way that Tomb does. Both of them flub it, however. Princes has the clues, but not a good transition towards the exploration of the temples and dungeons. Tomb fumbles the clues and the paths you can take to discover Omu.

Meanwhile, Storm King’s Thunder and Curse of Strahd are very firmly in the quest-based sandboxes. You have goals, let’s see how you complete them. Curse is, of course, a far more expansive sandbox. SKT instead has different quests and rewards for different groups. It’s why they’re more successful for me: the players always have an idea of what they want to do and (probably) what they need to achieve it. When the players have an obscure goal and no idea how to proceed, you get frustrated players.

That’s the sort of thing that happens when all the quest-giving NPCs die in SKT, and the DM doesn’t introduce Harshnag.

However, perhaps what’s missing from all these adventures is a proper explanation of how things are meant to work. Analysing adventure structure is a passion of mine, but it’s certainly not the passion of most DMs out there. Many pick up an adventure and expect it to work.

If you can explain to a DM how you expect the adventure to work, and the decisions they’ll need to make to enable it, you’ve gone a long way towards making adventures more accessible.

A sandbox, like all adventures, needs the players to have goals, and have an idea of how to complete those goals. In some sandboxes, the goals may be player-driven. But in the sandbox-as-tool of narrative design, it’s more likely the adventure sets the goals, so making them and their path clear is in everyone’s interest!