The Railroad-Quest-Sandbox Continuum

11 Apr, 2025

I find myself regularly in want of a more elaborate vocabulary and framework for discussing different styles of running a D&D campaign.¹ Working from Dungeonantology's INTERNAL LOGIC FOR YOUR CAMPAIGN SETTING as a base, this is what I have come up with, centered around this graph:

the units of the y-axis are left as an exercise for the reader

As you slide your campaign style from left to right, the roles of the GM and players change, along with the type of prep you tend to create. Different styles warrant different kinds of advice; what is good advice for a sandbox can be catastrophic to a railroad and vice-versa.

A GM asks: "How do I keep players from getting side-tracked?"

Respondent A: "Pause the session and tell your players out-of-game that they need to get back onto the story or else the game can't continue."

Respondent B: "There is no such thing as side-tracked, whatever the players are doing is the real story of the game."

Respondent C: "Try connecting your main quest to the characters' backstory to make it more enticing and relevant to them."

Who is correct? It depends on the style of game the GM is trying to run.

Definitions

I am co-opting some colloquial terms here and defining specific usages of them.

• "GM Responsibility for Fun²" is the overall control the GM has over the content of sessions and thus their quality. This is a proportional measure, so as it decreases, other factors gain prevalence in how much they impact the session. It never reaches 100% nor 0%.
• "Character-Centric Prep" measures how much the GM's prep work revolves around the party and the players, as opposed to prep-work that is made for its own sake regardless of the existence of the party. Level-scaling content to the party can be seen as a basic form of this, whereas as this nears maximum, a majority of the GM's prep is tailored to match the characters' individual motivations, backstories, builds, etc.
• "Player Agency" here means the amount of influence the players and their characters have on the direction of the story as played out at the table. It never reaches 100% nor 0%. How meta the players' agency is can vary, like players choosing an adventure module out-of-character vs the party making a decision within the fiction. Higher player agency tends to ask more effort of the players.
• In a railroad³, the GM sets a goal for the players and has a specific route they intend for the players to take to achieve it. The players might have some leeway during gameplay, but ultimately the path is set for them to follow.
• In a quest, the GM sets a goal for the players, but its up to the players to figure out how to accomplish it. The players are expected to stay pursuing this goal, to varying degrees. Unlike a railroad, there is no assumed path to completing the quest. The GM or module might have come up with some pre-planned possibilities, but the players are not restricted to following them.
• In a sandbox, the players define their own goals, and pursue them as they see fit. The players can end up flitting between different situations, dropping out of ongoing plotlines, or creating whole new ones. This requires active players willing and able to generate their own goals, as well as come up with viable ways to achieve them.

Railroad

In a pure railroad, there is a single route laid out that the characters follow, either implicitly or with the rails enforced by in-game or out-of-game barriers. There is little content "off the rails" for players to engage with, and "derailing" the game can be a source of table friction.

Since the players are essentially along for the ride, the GM has maximum (but not total) input on how fun the session ends up being. This also means the players have little agency on how the story unfolds. Most railroads are pre-planned without regard to the player characters, especially if it's a published adventure module. Usually they are at least balanced to the party's level.

An example is a totally linear adventure where an NPC guides the party through the adventure and prevents them from deviating. Many published adventures are railroads, especially one-shot modules.

The linear sequence of planes traversed by the players in Vecna: Eve of Ruin.

Pick a Rail

There are multiple railroads that players can choose between, either explicitly or as a result of "derailing". The paths they end up on are still static and there is still little content "off the rails".

The ability to choose a path is an uptick in player agency, but the GM still largely has control and responsibility over the game content. Prep is still basically agnostic to the characters, but the GM might need to adjust the adventures offered to the players based on level or party composition.

An example is a "branching path" type of structure, where the routes are pre-written but specific decision points have been designed to allow the party a choice of certain paths at those junctions. Some large published campaigns have this structure.

Quest

In a pure quest game, there is a single defined quest the players are expected to pursue, but they can accomplish it in whatever way they can come up with. There is little content outside the provided quest, and trying to run away and pursue some other quest would cause friction at the table.

Now that the players can choose their own approaches and strategies, they have gained more agency and more control over the fun-ness of the session. Their choices and behavior begin to shape the gameplay, possibly in ways the GM did not anticipate. Quests can tend to be character-agnostic, but level-scaling and sometimes adapting them to specific character types becomes more relevant.

An example is an adventure module like A Fistful of Feathers, where the players have the pre-defined goal of acquiring valuable goose feathers from the forest, but how to accomplish that is up to them.

Multiple Quests

There are several GM-defined quests placed in front of the players over the course of the campaign. Sometimes there is an overarching Main Quest that takes center stage, or the quests can be equal in importance. Players might be able to do some small activities outside these quests, but they will largely be a diversion and not have much importance to them.

With the ability to choose the quests themselves, the players gain even more control over what goes on at the table. While the GM is still designing the quests, the content is starting to be curated by the players. Players naturally tend to be attracted to quests that align more with their character concepts, and having several quests available makes adapting the content to the party even more relevant.

An example is giving the players a choice between several quest-style adventure modules to pursue at their leisure.

Quest Cards from Dragon of Icespire Peak, a very explicit example of Multiple Quests.

Character-Integrated Quests

There are multiple quests throughout the campaign of varying importance, but the GM also creates bespoke quests and narrative lines integrating the characters' backstories and intended narrative arcs, often collaborating with the players to do so. This investment of effort tends to focus the game towards these prepared quests, and ignoring them can be seen as disrespectful to the table.

This is the point at which GM and Player collaboration is at its highest. The players have greater responsibility to create interesting characters that the GM can then use to create interesting quests. This is a tipping point where if a player creates a one-dimensional character that doesn't give the GM much to work with, the resulting one-dimensional quest is at least as much the player's fault as the GM's. While the GM might still be designing the quests themselves, player agency has grown because they are now gaining direct input on the subject matter of the quests.

An example is a Critical Role campaign.

Character-Focused Sandbox

The players are placed into a world to explore and engage with. The sandbox is seeded with plot hooks tailored to the player characters, and the GM will sometimes fudge the world and situations in order to facilitate the party's activities. The GM might add new content, create dramatic timelines, or seed example hooks and goals for the players to grab onto, but there is no expectation that they are "meant" to do anything in particular. A small side activity by one character might develop into an entire plot arc, and that is equally as valid as any other gameplay the players might engage in⁴.

Entering the sandbox zone, now the players take on the responsibility of creating their own goals, often at the table. Much of the content in the character-focused sandbox is designed for the characters, so there should be plenty of hooks, but it's ultimately the players' job to seek it out. The GM might prod the players with a plot hook or dramatic moment if they get stuck, but the players now have the majority of control over the fun-ness of sessions.

An example would be creating a sandbox region with several competing factions, each of which ties into a character's backstory.

World-Focused Sandbox

The players are placed into a sandbox that is created without regard to their characters, and it's up to the players to engage with the world as presented. The GM may sometimes err on the side of genre or drama, but the world is largely operating on its own logic and own time. The players can pursue any activities they like, but those aren't guaranteed to end up interesting or important.

As players are now plopped into a world not made for them, the importance of their own agency grows. Now its up to them to make their characters engage with the setting and get involved in things, and turn those entanglements into fun gameplay. The GM might give them a nudge here or there if the context allows for it. Prep is largely made unrelated to the players, but some level-scaling might be employed.

An example would be plopping the players into a warzone with several competing groups and having the war progress on its own even if the players don't intervene.

Somewhere in this zone, the GM's response to stuck players changes from "give them a plot hook" to "fold your arms and wait". A threshold is crossed where now there is no expectation of the GM to push the game along.

Procedural Sandbox

The players are placed into pre-designed world or one created through random generation. The GM takes the role of an impartial referee and runs the game entirely based on the in-world logic or procedures used to create it. The GM defers to the already established rules of the world whenever possible, even to the detriment of genre or drama.

With maximum player agency, the players have the freedom to get themselves into no-win, undramatic, or boring situations. They should not expect the GM to bail them out, as the GM simply runs the world as defined in the procedures, leaving the fun-ness of the session entirely in the hands of the players and the rules. These kinds of random generation systems are usually entirely character-agnostic, often even eschewing level-scaling.

An example would be putting the players into a hexroll generated sandbox and asking them "what do you do?" without any specific plot hook. Hexcrawl modules are an archetypical form of procedural sandbox.

A randomly generated hexcrawl map from hexroll.

These categorizations are inherently fuzzy and can often bleed into eachother. A given campaign likely mixes two or three styles over the course of its length. A single pre-determined mission within a sandbox campaign can be considered a quest. A railroaded campaign might suddenly open up into a sandbox for a short time before returning to the railroad. A multiple-quests game might evolve into a character-centric quests game as the players become more confident in their roleplaying.

Each style can have significant variations in the absolute amount of GM effort and preparation required. Running a pre-made railroad adventure is a lot less work than writing an entire level 1-20 campaign from scratch. Equally, a randomly-generated sandbox could take 0 prep, while a fully-bespoke sandbox could take months to put together.

When I run an open table the sessions are straight-forward Quests. My home game style is something like a World-Focused Sandbox, but I usually start the game with a Quest as a more linear introduction. I also prefer to give the party an overarching purpose like "win the war" or "defeat the villain" to focus the game a bit more, which is at odds to the notion of a pure sandbox. I also enjoy putting together Character-Centric Quests when the players give me the material for it.

Each style has their own strengths and weaknesses, and different groups of players will prefer different types of campaign. Many players are much more comfortable in a railroad or quest game than being asked to drive a sandbox by their own volition. Other players will chafe at any structure and actively seek to break out of any pre-written plots set before them. Attempting to run a sandbox for passive players, or running a railroad for free-willed players, is a recipe for disaster. There is no "correct" way to run a D&D campaign, just different styles. Being cognizant of what game style you are running (or want to run) makes you a better GM, and makes you better at interacting with your fellow GMs.

#D&D #DMing #ttrpg