Why Dungeons & Dragons Players Don’t Love Sandboxes as Much as They Think

Many role-playing gamers set sandbox adventures as an ideal. We all agree that railroads make bad adventures, so do sandboxes offer all the virtues that railroads lack?

In role-playing adventures, sandboxes and railroads fall on ends of a spectrum. Railroads offer players no options. Sandboxes allow complete freedom, including freedom to choose a goal. If a character favors a bartending in Barovia over vampire hunting, they still get a place in the campaign.

Boxes of sand let kids choose their own goals. They can make sand castles, bake sand cakes, anything. And when they grow up, they can stage miniature battles.

Some games deliver all the freedom of a box of sand. Minecraft lets you play a survival game, but it owes its success to all the other things you can do: Some players build forts or replicas of the seven wonders. Some create a circuits from redstone. Players make their own goal.

D&D used to force a goal on characters

Original Dungeons & Dragons never started as pure sandbox, because the rules included a goal: Take treasure from dungeons and the wilderness. By rule, characters who won treasure gained experience and power. They won D&D. See The fun and realism of unrealistically awarding experience points for gold.

When the original D&D characters reached high, name level, the game turned into a sandbox where players chose a new goal for their characters. Stronghold building offered fighting men an obvious goal, but some other classes lacked anything as clear. What do you want for your bard or druid? Should a wizard build a tower or start a school? Apparently, many high-level wizards go mad and build dungeons. Where else could the living-chess puzzles and reverse-gravity rooms come from? Endless possibilities await!

Instead of embracing the freedom of a high-level sandbox, players returned to dungeons.

Sandboxes can overwhelm players with choices

In Why Dungeons & Dragons (and Role Playing) Took Years to Leave the Dungeon and How the dungeon crawl’s advantages propelled Dungeons & Dragons to success, I described the attraction of dungeons. Among other advantages, dungeons limit the characters’ options. This doesn’t just help dungeon masters prepare, it helps players.

Common wisdom suggests there is no such thing as too many choices, but psychologists conclude that people flooded with options become paralyzed by them.

When dungeon masters offer a true sandbox and come willing to improvise any course their players choose, they confound players. Once the players stop wondering what they’re supposed to do, they struggle to choose from boundless possibilities. Whatever they finally decide, they leave the table with a nagging feeling that they chose wrong.

The value of limited options

In D&D, dungeons, patrons, and hooks all limit the options that players’ face. Such tropes give players direction. A little direction improves the game.

Make no mistake. Players still want options. Every game session should leave players wondering what might have happened if they followed a different course. Mike “Sly Flourish” Shea recommends that giving players three plus infinite choices. DMs should offer three known options that take characters closer to their goal, while being open to anything players want to try.

Many sources of DM advice suggest seeding a sandbox setting with hooks—opportunities for players to land in stories of their choosing. Exactly. Those hooks help players narrow all the options of an open world to a sweet spot of three plus infinite choices. They nudge the game a bit closer to the railroad end of the dial. Some railroad-phobics might even argue that such hooks show a DM working too hard to push players through a story. Their ideal game only works with perfectly spherical, frictionless players. The real players at your table want hooks.

The sandbox dungeon

D&D’s mega-dungeons limited players’ choices, but many fans still tout multi-level dungeons as sandboxes. Sure, characters need to adopt the goal of seeking treasure, but they never need to dutifully follow a story arc planned by a DM. Plus, players could chose a difficulty level by deciding how deep they dared to delve. A few D&D players still favor that style of play.

Embracing story and fewer options

Despite the freedom of a dungeon sandbox, most D&D players craved story and deeper motivations. The D&D game changed to provide. When Tracy and Laura Hickman penned a series of classic modules including Ravenloft and the Desert of Desolation trilogy, they led the change. Their introduction to a self-published version of Pharoah gives D&D adventures four, new requirements:

  1. A player objective more worthwhile than pillaging and killing.
  2. An intriguing story that is intricately woven into the play itself.
  3. Dungeons with some sort of architectural sense.
  4. An attainable and honorable end within one or two sessions of playing time.

When characters explore Castle Ravenloft, they quest for more than loot. They aim to free the land from the menace of Lord Strahd. Adopting the goal of a story takes a measure of freedom from players. Now the their options narrow to the choices that lead to the magic items that will help defeat Strahd. Few players mind. They see clear options that take them closer to achieving their characters’ aims. As the adventure progresses, the players’ paths narrow to a railroad that leads to a final confrontation.

Of course, at any time, the characters could leave the railroad and open a tavern in Barovia, but that never happens. Partly because D&D players like doing D&D things such as smiting evil and winning treasure. Partly because players follow D&D’s social contract by honoring the DM’s preparation. Mostly because players enjoy stories in D&D and they willingly abandon the freedom of a sandbox to foster them.

Too often, D&D fans tout sandboxes as the pinnacle of adventure design. Dungeon masters and adventure authors aim for the freedom of a sandbox, but just leave players feeling adrift. Players enjoy D&D most when they see a few, clear options that take them closer to achieving their characters’ aims.

40 thoughts on “Why Dungeons & Dragons Players Don’t Love Sandboxes as Much as They Think

    1. James

      That’s why I am leaning towards a “West Marches” style of wilderness adventure: the players decide where they’re going to go or what they’re going to do between sessions, so that the session itself can be devoted to the agreed upon goal. That said, the sandboxy nature of the campaign certainly means they could stumble on something else they would rather check out, but that is a different matter entirely from wasting half the session searching for adventure hooks.

      1. Douglas Cole

        My difficulty with “full sandbox” is quite simply as a GM, it can be very very challenging based on time constraints. When I can (say) only game once a month, or every other week, then having the players go off in a completely unexpected direction, no matter how cool, basically puts the game on hold unless the setting, system, and personalities involved are really amenable to winging it.

        that’s me, though – many folks are quite comfy in that space.

        I do like the three plus infinity comment, though. That’s a good summary.

  1. Logan Bryce

    Excellent article. I used to be a huge fan of sandbox until I listened to TAZ and it blew my mind what could be done with a fairly linear storyline. Can’t wait to start using some of those ideas in my game.

  2. Andrew Six

    Love the article!

    I’d say the problem with the railroad is less about the endgame and more about the path players take there. Knowing that there are four Macguffins you need to find before confronting the big bad – who is usually stationary – does not itself a railroad make; because, with that alone, the PCs have infinite freedom in which order they find those macguffins, and how exactly you go about obtaining them. And while dungeons are undeniably railroady (there’s a 10-foot-thick stone wall preventing you from moving off the DM’s preferred path, after all), there’s a certain logic to them: it makes *sense* that a dungeon would be railroady, because it has walls.

    Narrative arcs, however, do not have walls.The worst kind of railroad, in my opinion, is one where the PCs bump from questgiver to questgiver until they eventually stumble their way into the climax of the adventure of no volition of their own. Ones where, in the second or third session, the players are confronted with some looming threat that – for some reason or another – only they can stop, and the flowchart for the rest of the campaign is basically a line of questgivers telling the PCs what they need to do next to stop said threat (or something equally obvious, like finding a signed note in a chest at the end of a dungeon) until, at the end of the line, the party stops the threat. It might have a fork here or an optional mission there, but those only serve to highlight how linear the campaign is; even railroads branch occasionally.

    I might draw an analogy to Legend of Zelda: Breath of the Wild, which has influenced my campaign design considerably. The Divine Beast quests are an example of bad campaign design (though not necessarily bad video game design): there’s a linear quest to enter each dungeon, each with a single NPC you *have* to befriend culminating in a set-piece encounter. But Ganon’s castle? There’s a dozen ways to get in and a thousand ways to scale it. There are no arbitrary requirements, no NPCs you have to befriend, no paths you have to follow. You can run to the castle at the start of the game, climb the walls all the way to the top, and fight (and probably lose to) Ganon without any artificial barriers standing in your way. The rest of the game, then, is just an exploration-based sandbox of sidequests designed to make that final confrontation easier.

  3. Vase Odin Crisan

    “We all agree that railroads make bad adventures”. That is a very general statement that is very misleading, as almost every player I know personally (over 50 total) agrees that “railroad” and “sandbox” are BOTH kind of necessary and too much of either is bad. I actually don’t know ANY players who specifically prefer sandbox.

  4. Michael 'Chgowiz'

    David, I love your stuff, but I think I’ve got some rebuttals to make. Rather than write blog posts in your comments, I wrote them in two parts – part 1 is here. Hope you’ll stop by! http://chgowiz-games.blogspot.com/2017/09/sandboxes-and-players-they-make-story.html

    First point- I don’t think that’s an accurate assessment of what OD&D/AD&D were meant to do. History itself, through the published campaigns of the time, rebuts that.

    Second point – it’s a common fallacy that sandboxes are supposed to be passive only. Sandboxes are meant to be extremely vibrant and active!

    Third point – it’s another fallacy that sandboxes can’t have walls or limits. Of course they can! What they don’t do is remove players’ agency.

    Fourth point – The Hickman modules are about as railroady as they get. If you’re going for extremes of sandbox and railroads, you’ve got them both! 🙂 I think the truth and sweet spot is more towards the middle.

  5. Damian Williams

    I am a very new DM (about 2 years) I can’t manage sandbox well and neither can the players. When I build for the week, I build for 3 or 4 things to do with a fast or slow path to each goal and let them pick where they want to spend their time. Remaining uncompleted quests remain open, get recycled or resolved by NPC and affect the story arc later.

    When we first started playing as a group I actually built around the mechanics, it was definitely railroad but not a narrative straight line there was something new every week and the story changed even though they had limited options. I only taught the mechanic as it was played. It went amazing and locked us together.

    As I was learning “how to” I was introduced to a lot of personalities, thinkers and shapers of DnD. They all had wonderful and varied advice on the game. I appreciated all the time and energy they put into teaching and sharing but ultimately the table has been my best teacher. Sandbox can be fun at the higher levels to be sure but until the players find their footing in the campaign free agency can be rather underwhelming and we all feel better pushing each other in directions.

    Honestly that’s been the most exciting part, having multiple paths laid out and watching the party debate how to spend the time is like it’s own game in and of itself and provides ample drama.

  6. Chris Hinson

    Nice Article DMDavid, I enjoyed the read. You seemed to make a good point that some guidance is good, and allowing players to make choices within those rails is best. We strive for exactly this with the RPGCrate adventures. Each card provides guidelines, a problem to solve, and a story. But, the choices of why, how, and when to deal with those problems is typically left up to the player – within the guidelines of the world rules. It’s perfect that you use the term sandbox and not just sandpile. Everyone knows that a sandbox also has a BOX around it for guidelines to keep the sand inside. If you just create a pile of sand, with no borders, it eventually devolves, just as the game will devolve, into dirt and grass – no one knowing where to “Play”.

      1. Michael 'Chgowiz'

        It’s not been my experience that player “skill” at sandbox is necessary. I have had many different skill levels of players in my 9 year sandbox, some familiar with the concept, others not at all.

        It is all in the presentation. In how you open up the world, how you show what is going on. it could also be the West Marches style where the players do a good job of sharing information and passing it on as well, but they had to be taught/shown that too.

        And I think that’s another myth – that there’s no “structure” to sandboxes. There might not be as clear of a PLOT POINT A to PLOT POINT B but again, this is presentation. If I can present an interaction or provide clues, that is no different from one to the other.

  7. Ray

    I played in a sandbox game and none of the other players were to keen on doing the original hook and with no prompt we just ended aimlessly wandering. I’m a firm believe in the story having a beginning middle and end that leads somewhere. Provide options and be flexible as a DM but have those hooks and an overall what is going on in the world. When I DM I have an outline of a story and then remain flexible for shenanigans.

  8. C.H. Peterson

    Ultimately, you need to know your audience. Generalizing and saying that railroad or sandbox is superior to the other is selling your players short. No single gaming group is the same as another.

  9. Steve Kent

    Nice! Sandboxing and railroading (at least as we’ve defined them here) are two extremes that most adventures fall somewhere in between. My guiding principle to keep me from getting too far one way or the other is “players should be able to do whatever they want, but there should always be something for players to do.”

    Right now my favorite approach to open-world adventures is to make a wilderness to act as the overworld, then make a town and some “dungeons,” which are just more detailed areas with encounters, secrets, a boss and treasure

  10. alphastream

    A good example of sandbox play is a big open wilderness you can explore. A bunch of empty hexes and you slowly fill them in. Because you are free to go anywhere, the world isn’t expecting you to go one way or another. You might end up at the goblins, or at the evil sorcerer. The Caves of Chaos are a bit like that. I think that lack of motivation or direction is a key that the Hickmans helped change in adventure writing.

    You can add some constructs and still have it feel very open. The game Fiasco is like that. You come up together with various setting bits: the treasure vault, the guards, the corrupt king. You build relationships between players and setting. Then you take turns deciding what will happen and everyone adds to that. It’s a great game.

    I’ve had that kind of fun a few times during convention games where players abandoned the plot and just made their own fun. I’ve had my players do it to me once. They were good friends and could basically self-entertain themselves with all they knew about each other’s characters. Eventually we got back on track.

    I think D&D isn’t a particularly good game to try to emulate the Fiasco model or something like it. Just exploring random wilderness or caves with no actual causal relationships between the things we find is going to be dull over time in a campaign or series of sessions. It no longer is considered acceptable adventure writing. We could develop some Fiasco-style rules, but now we are adding a layer onto the game. Or, we have to truly work very hard to put a lot of pieces in the adventure/campaign that keep our interest high and are flexible enough to interact with all the other pieces. That’s hard.

    I tried to do it to a certain extent in The Howling Void, by having more encounters than you could visit, no set order in which to explore, events that take place at specific times, and two sets of foes moving around locations on a set schedule. I think it works well, but it requires a well-prepared DM. And, players often say they wish they could play the whole thing. (In home play, they often do). It’s not truly a sandbox (you have a clear goal to stop a cataclysm), but even here it is hard to give a large range of options and have it pay off well for all types of players.

    1. alphastream

      And, in case it isn’t clear, I generally agree with David. I think sandbox has great appeal, and it sounds effortless, but is actually very hard to pull off well without additional subsystems and/or the right group. I think our modern focus on plot and motivation is a great advancement in adventure design, even while it can be fun to draw from sandbox elements.

      1. Beoric

        Its easy to screw up either extreme. From what I see of published modules, the modern focus on plot and motivation is also hard to do without ruining player agency. At least, a lot of adventure writers seem to have that issue.

  11. simontmn

    “Players enjoy D&D most when they see a few, clear options that take them closer to achieving their characters’ aims.”

    I think the reality is that players like to have both clear options AND the freedom to proactively set their own goals, And they definitely value the ‘open world’ feel of something like Skyrim. Also there is a time & place for both. I find that in my sandboxy games, what works best is 2-3 clear options at the start of the campaign, set in an open world where the players then have the opportunity to set their own goals. It should be a ‘world in motion’ where NPC acrtivity also generates new clear options. And it’s ok to plonk down the occasional ‘set quest’, but too much of that makes it a linear campaign.

  12. mschellman

    I agree. I used to play AD&D and now 5e. Homebrewed adventures were mostly just elaborate maps full of monsters and the goal was to clear the out as much treasure as possible. Better DMs had some kind of story The goal was to present the illusion of choice but all the while gently nudging the players through towards a predetermined goal – helping them to have fun by continuing to make progress. I like the more open plan of 5e, but I can see how it could be frustrating if players insist on wandering into undeveloped areas – or spend their time with no inspiring goal. I like your blog a lot – keep up the good work.

  13. Pingback: Sandboxes? | Cirsova

  14. Pingback: Creating a Player-Focused Tomb of Annihilation Campaign | Alphastream

  15. Robert Conley

    I wrote a few things on this topic. 🙂

    There are two recognized potential issues with sandbox campaigns that you touch on in your post. One the overwhelming number of choices is dealt with by establishing what I call a good Initial Context. The problem isn’t that there are too many choices, the problem is that there not enough information to decide which of them are either interesting or which will serve the player’s goals for his character the best. Hence an initial context.

    Limiting options doesn’t solve this issue as every one of those options can be found uninteresting to the player.

    Second the lack of story. A story in part is an account of a sequence of events. Following a story is antithetical to a tabletop roleplayiing campaign as it robs the players of meaningful choices. But a story is not the only way to convey the sense that there a deeper meaning to what going on.

    With tabletop rpgs a superior method is to show that meaning by what the NPCs do and don’t do. To have the players experience the events. As opposed to passive experience of being told about the events.

    But it more than that. Because as free agents the players as their character can change the outcome of events in ways that you the referee can’t foresee. For tabletop roleplaying any planned sequence of events should be treated like a real life plan of battle. Something that organizes and focuses the direction of the campaign, but is ultimately changed in light of new circumstances. In this case the circumstance being what the PCs choose or not choose to do.

    Myself and other call this a World of Motion. A way of bringing the setting of the campaign and its inhabitants to life. That it isn’t some static place that only unfolds when the player arrive at a locale. But something dynamic with a life of its own.

    I recently wrote a summary of this in a post

    I talk about a World in Motion here

    and here

    I talk about the Initial Context here and includes an example from actual play.

    and here

  16. Jeff

    No group game can effectively be a complete sandbox.

    If you have more than 1 player at the table, the group has to come to a consensus about what overall direction to take in the story, or they cease being a group. The larger the group, given a true “sandbox” setting, the more likely at least one member will decide they want to take a radically different direction than the group. While that can work for 1 or 2 game seasons, more than that either results in two games, or that player leaving. So the “sandbox” is self limiting, if you want the game group to continue.

  17. ripx187

    I think that you are confused over what a sandbox is, it isn’t just a set which players run around and do whatever in, the DM is playing the game WITH the players. The DM is causing movement through NPCs, and the sandbox is a game of problem solving and consequences. There is still a story going on here, but it isn’t static, it is dynamic.

    The Dungeon is also a sandbox. If you make too much noise in a room, the creatures that can hear you will react and make decisions. There should be a history to the dungeon, as well as politics and ecology. If it has none of this, and you can fight 4 orcs standing guard in a hall, then open a door and still find 12 orcs in there playing dice and gambling in a 3×3 room, this lacks logic and is static and boring.

    A villain whose plans are flexible, who causes movement but also reacts and changes plans according to logic and the movements of the PCs is interesting and dynamic. The villain has goals, objectives, and his own set of victory conditions, we as the DM are playing him. He also has limitations and must follow the same rules as the players, if they can get ahead of him or force him to move in a predictable manner, they can catch him.

    Instead of prepping every session, we prep once, and only what we need for the next game. A good sandbox, once the initial prep is over can be used for many many games with no more prep involved beyond making simple adjustments here and there. The castle doesn’t move, but the people in and around it do.

    If I were to write what happened during my games so that you can recreate these events at your table, you have a module just as static as “Ravenloft” even though my game itself never felt this way.

  18. Pingback: How to Start a Sandbox Campaign | DMDavid

  19. Justin Alexander

    “In role-playing adventures, sandboxes and railroads fall on ends of a spectrum. ”

    Common misconception, but not actually the case.

    This kind of sets the tone for the article. You claim that default goals aren’t part of a sandbox (even though they’re actually integral to sandbox play and always have been), and then conclude that sandboxes don’t work because players don’t have a default goal. That’s like arguing that cars don’t work because the “true automobile” doesn’t include wheels.

    Then you claim that “true sandboxes” don’t have hooks. Again, you’re flying in the face of reality here: Since 1974, sandbox settings have featured baked-in hooks.

    Then you claim that a hook “nudge the game a bit closer to the railroad end of the dial”. And that doesn’t make any sense at all. You’re claiming that someone saying, “Hey! The Coliseum is really cool!” is attempting to kidnap you and force you to go to Rome against you will.

  20. Jim Inverness

    The term sandbox here is a misnomer. The opposite of a railroad is free, abundant choice. A “sandbox” is just an isolated, confined play area. Think Ravenloft or Averoigne.

  21. Tag

    (ok, this turned out long, so…
    tl:dr, stating that a ‘sandbox style’ game would give players too many options to be enjoyable is kind of a false-equivalence. Psychology studies suggest that more GIVEN options tend to lead to less satisfaction than for those with less GIVEN options. ‘Sandbox’ doesn’t necessarily mean ‘more given options,’ though.)

    I know the psychology article/concept you referenced. I think. Maybe. I mean, I don’t wanna read it cuz I know the concept rather well (just a quick skim)…
    Forgive me if this is left-field,
    But the idea was between a smaller vs a larger set of GIVEN options. So yeah, if a DM were to choose between presenting to their players a list of 3 quests and a list of 7, your average player would prefer 3 (the theory is they feel a sense of loss for the options they didn’t choose – aka the “opportunity cost”). And it does tend to make people hesitate more when faced with an option with more given choices.

    [Easy illustration of the concept (plus our common work-around): a restaurant that offers nine types of sodas vs. three. The consumer figures there isn’t a need to order more than one soda, so they should choose just one…but has to decide between caffeine, diet, and flavor options, and begins second-guessing themselves about ‘maybe I do need caffeine’ and ‘if I get root-beer then I miss out on the cherry-coke option.’ And of course, people have already solved that by deciding their favorite ahead of time. Brand loyalty.]

    That’s the thing with a sandbox though: it doesn’t list the choices for you. What I’m saying is, it’s the sense of having a list of choices/options that has the supposed aforementioned effect. A sandbox game (by definition) has limitless options, but they aren’t usually listed to the player. And as long as the chooser doesn’t feel like they are losing out on options B and C (or B through G), they notice less potential-loss for those choices they hadn’t selected. (But yes, as mentioned, options may still be explicitly given to the player, such as via questhooks, but in a sandbox game it should be understood that while the questhooks represent opportunities, they aren’t the players’ only choices.)

    The concepts are related of course. Maybe (1) players in a sandbox game tend to have more Given Options than those in a more railroad game. Or even, (2) players may tend to have a more negative experience in a sandbox game, similarly to those who are given a list of more options (due to the vagueness and/or imagined number of possibilities in a sandbox world, and the stress of losing out on those options he/she doesn’t choose? I guess?). But as to the first point, I think a good DM would simplify as-needed for their players (like the 3 + infinity point you/DMDavid referenced – I like that idea). And as to the second, I don’t think psychological studies have supported this conclusion yet. Perhaps even, the fact that one’s choices in a sandbox game AREN’T limited may negate the stress from more (limited) options.

  22. Pingback: Running a Sandbox D&D 5e Game - Ambition vs. Possibility - Dungeon Solvers

  23. Pingback: How to Start a Sandbox Campaign | DMDavid

  24. Cheimison

    Yeah I fucking hate Dragonlance modules, I have no idea what you’re talking about. Having a theme and a mode of play is not the same as a railroad, which is what most of the modules in the last forty years are.

  25. Pingback: How to create a freeform (aka. open world, sandbox) campaigns and keep your players entertained for years? – Fanatic5 Solutions

  26. artedetimo

    Good discussion going, but a lot of straw men used in arguing all this. Neither railroad or sandbox seem like how a social game can be played, if it is lasting.

    As an analogy, a sandbox for example can be just sand, which ends up being dug up and thrown, then left behind in tears.


    A sandbox with shovels and beach toys will go one way, and one with action figures and dinosaurs another. Add a cardboard box or a flowing water hose and you got a world. Let little 2yo brother stomp through and you have a villain. If the bucket of toys dumped in is engaging enough, whatever type they are, everyone usually has fun playing that style of game.

    Its not the sandbox or railroad in the end, but what’s in it and who’s playing.

    I haven’t been a DM long, but i’ve felt like my role is to bring the raw materials, and some manipulatives, then invite the PCs to bring their own imagination and creativity to build something out of it. Neither a sandbox or a railroad, and more like a craft table.

  27. Pingback: Running a Sandbox D&D 5e Game: Ambition vs. Possibility - DungeonSolvers

  28. Pingback: Running a Sandbox D&D 5e Game: Ambition vs. Possibility – DungeonSolvers

Leave a Reply to C.H. PetersonCancel reply