Of Railroads, Sandboxes, and the Roleplaying Happening in between

(In this text and others, I use the term „Host“ for the position that is widely known as GM (game master or game moderator), and refer to any unspecified Host as she/her, obviously without implying that a Host is always female. In the same way, I will always refer to unspecified players as he/him for clarity. Generic examples for player characters (PCs) and Host characters (HCs) follow this pattern, too. For a longer explanation, please see this entry.)

This is another post which I wanted to write for a very long time (I have literally dozens of those…) It has been bumped up in my priority list recently, though, because of this blog entry by Xathrodox86. (But then again, my blogging for IMAGINATION went on a hiatus just when I was working on this, and later I got stuck, and then distracted by another blog entry, and distracted from that distraction by a third… so „recently“ is not at all true anymore.) I already wrote a lot on this topic in the comments to that blog entry, but unfortunately I could not directly reply at that time for technical reasons, so Xath had to copy and paste my words from direct messages, making them difficult to identify, and slightly garbling their formatting. I reiterate and refine a lot of what I said there here, but those replies also included a couple additional ideas and opinions, so if you happen to be interested in what I wrote there, I give you a somewhat cleared up version of it behind this link.

Ends of a spectrum, not categories

I have been unhappy for a long time with the inconsistent use of the terms „railroad“ and „sandbox“ in the roleplaying community, and especially if that use had been in conjunction with the connotations „railroad=bad“ and „sandbox=good“. I’m afraid that this terminology has been successfully hijacked by fanatics from the anti-storytelling camp. I hope I can clear up some of the misconceptions which have developed (and partly consciously introduced, I suspect) around it.

I admit I haven’t done any real research on the origins of those terms (and I don’t believe that would have been especially enlightening, although certainly interesting), but I suppose that „railroad“ has been seen negatively from its inception, while „sandbox“ has likely been a positive term from the beginning. The main association with „railroad“ is that you’re just a passive traveller with no real agency over the events you experience, while „sandbox“ evokes the idea of using your creativity to build something all by yourself from scratch. Over time, however, „railroad“ has come to describe all scenarios with a mostly linear structure, and for some even all plot-based roleplaying in general, while „sandbox“ has been used for all kinds of less linear approaches. Use and misuse of those words has reached the point where different people might categorize the same scenario with either – and yet, the terms have largely kept their original negative and positive associations. As a result, they tend to say less about the scenario or playing style that’s being discussed, and more about the preferences of the person using those terms.

Before I try to explain the variety of approaches in the spectrum between „railroad“ and „sandbox“ (I think I can get rid of those quotation marks now, can I?), I want to describe its extreme ends as concisely as possible:

In a pure railroad approach, the Host has complete control over the plot, and the players have none. Their relationship resembles that of a narrator and her audience.

In a pure sandbox approach, the players have all available (meaning all that is not determined by the game rules) agency over the plot, and the Host has none. Their relationship resembles that of actors and stagehand.

In between those extremes, the Host and her players cooperate in the endeavour which I know and enjoy as roleplaying.

The true railroad

Let me put this straight: A railroad in the original meaning of the term is certainly a bad thing! It puts the PCs in a narrative which their players cannot meaningfully influence. Many old scenarios effectively required them to make certain decisions exactly in a certain way, or the plot could not proceed. That extreme approach has become a lot less common nowadays, which may be another reason why the term has taken on a broader meaning, denoting certain scenarios and play styles even if those do not actually share the flaws of a true railroad. The prevalent reason for its use, however, is clearly the desire to slander an approach to roleplaying which the user of this term does not approve of.

Players who consciously follow a plot which has been prepared by the Host, having their PCs make decisions they might reasonably make due to incentives offered by her, are not necessarily being railroaded – they simply keep their characters in the story. The important question is: Do they get to make enough meaningful decisions to influence that story? If the answer to that is no, then they are in a true railroad. If the answer is yes, they are probably instead in the following kind of scenario:

The linear scenario (or, the cruise)

Why can linear scenarios be fun? Because they are less about choosing the path, and more about what to do along the way. Certain scenes or events may happen in a predefined order – but that does not mean that the players have no say in what happens! They may choose to focus on different activities; may or may not detect helpful information, discover useful objects, or acquire new skills; may or may not save Host characters from dangerous situations; may or may not deal weakening blows to opponents they will encounter again later; may find allies, make new friends or new enemies… Just like people on the same cruise may have totally different experiences, even though they travel on the same ship and visit the same places at the same time, PC groups can make a linear scenario their own; creating unique memories, tackling challenges with diverse approaches, setting themselves up for major showdowns in different ways (and with different chances of success), and exiting the scenario with varying senses of accomplishment and individual plans how to proceed further.

The matrix scenario (or, the hike)

This kind of scenario is also plot-based, and certain events are still going to happen at certain times, but their order is only partly predetermined. A simple example (which would be even easier to understand if I had found a way to produce a simple flowchart): At the beginning of a scenario, the PCs have a choice leading them to either event A or event B. Event C triggers off event A, but not necessarily immediately – maybe A makes C possible, but the PCs still need to initiate it themselves. In the same was, B enables D. The crucial event E will happen after either C or D, but again not necessarily immediately.

In this case, the scenario path leading to E may be either of the following:

The shortest paths, skipping one half of the matrix:
A > C > E
B > D > E
Incomplete paths, abandoning one way for another:
A > B > D > E
B > A > C > E
Incomplete paths, completing one way, then exploring but abandoning the other:
A > C > B > E
B > D > A > E
The longest paths, exploring the whole matrix, while staying on its component ways:
A > C > B > D > E
B > D > A > C > E
The longest paths, exploring the whole matrix, but interrupting one way for another:
A > B > D > C > E
B > A > C > D > E
The longest paths, exploring the whole matrix, but alternating between its component ways:
A > B > C > D > E
B > A > D > C > E

That is a whole lot of different ways to experience a really simple matrix scenario! And since running one – possibly using a flowchart – is a lot easier than trying to describe it without a flowchart, even a moderately complex matrix scenario will yield a large multitude of unique approaches for a PC group, based on a small number of crucial player decisions, and without putting much of a strain on the Host’s ability to develop her plot. Like on a hike, the players can largely decide which path they take (considering the environment, of course, represented by the connections between events as prepared by the Host), but will nonetheless reliably arrive at certain important landmarks, including their final destination.

The open scenario (or, the amusement park)

I still call this kind of scenario plot-based, although only barely, but many will consider it a sandbox instead (with which I disagree). It minimizes the amount of predetermined events, essentially only using a climax (or possibly one for each chapter, if it possesses a large scope), which may even come in several distinct possible incarnations, depending on the PCs’ approach and the development of events. While there is a plot in the background, it may be as vague as – for example – an impending invasion (with the PCs’ task to prepare a region for war); or the demise of a local crime lord sending ripples through a city’s underworld (with the PCs being members of a gang making a bid to fill the resulting power vacuum).

While the players will need to make their own plans how to approach an open scenario, creating most of its events through their PCs’ actions, those are still part of an overarching story designed by their Host, and the PCs may even encounter specific events and elements she prepared. However, she will routinely shift the exact placing of those around – both with regard to the timeline and their location – and adapt them to match the story’s developments. To a certain extent, a Host also does this in a matrix or even linear scenario (since players are unpredictable); but in an open scenario her preparation has been geared towards that principle. Additionally, she will often run a timeline of background events, which the PCs may or may not be aware of (or partly aware of), providing a framework for their efforts, and determining their success or progress.

Most importantly, though, the Host is still in control of the setting – she might relinquish that control for some of its aspects, and allow players to describe those, but that will be by her own choice. While the players have a lot more freedom determining their PCs’ actions, they are still immersed in a consistent game world, interacting with it via their characters as if it were real. This is why I suggest the term „amusement park“ to refer to that kind of scenario: The PCs can go whereever they want and visit its attractions in any order they choose, but those attractions are already established there (and while the Host may spontaneously place especially interesting ones exactly where the PCs are going, they will not know it; and ideally, their players won’t either).

This is certainly a very satisfying way to roleplay, and some players might not be willing to accept any more confining approaches, but let me strike a few blows for those:

1. Once you get rid of the notion that collaborating with your Host to keep your PC in the story is reducing your fun, you will probably find that tighter scenarios can also be extremely entertaining, and rewarding in their own way.

2. The ideal kind of scenario depends on the factors time, experience, and preference. Simply put, the less (real) time you have for the completion of a scenario, the more sense makes it to run a more linear one, while a long chronicle will on the other hand be ideally suited for an open scenario. As to experience and preference: Some players are simply uncomfortable without an obvious plot to follow. That is partly a function of roleplaying experience, and often starting with linear scenarios, and slowly making their way via matrix scenarios towards open scenarios will be a sensible progression for inexperienced players. But even some roleplaying veterans might still prefer tighter plots (see my point above).

3. You can mix and match for variety. A longer chronicle can basically be run as an open scenario, but feature short intermezzos of matrix or linear scenarios. In my experience, this is the most rewarding approach of all!

The collaborative scenario (or, the adventure playground)

(From Wikipedia: Adventure playgrounds can take many forms, ranging from “natural playgrounds” to “junk playgrounds”, and are typically defined by an ethos of unrestricted play, the presence of playworkers (or “wardens”), and the absence of adult-manufactured or rigid play-structures.)

This is a big step from the category above, shifting from immersive storytelling towards collaborative storytelling, which I consider fundamentally different. The crucial aspect is the Host no longer being in control of the story, sharing most of her responsibilities with her players – that’s why it’s collaborative – and thus elevating them to co-Hosts (or co-storytellers, if you will). While this might sound great coming from a vague democratic gut feeling, it comes with a profound cost: The immersive aspect of roleplaying gets lost. You simply cannot experience a story from the inside while also creating it from the outside. It’s like trying to tickle yourself – just as your body refuses to get fooled that way, so does your mind. I’m not at all saying that collaborative roleplaying cannot be fun: You just should be aware of what you give up, and what you get in return. It isn’t fun for me personally, though – at least not nearly as fun as I know immersive roleplaying to be.

With this approach, there is no larger plot anymore. Both players and Host introduce story elements, but there is no pattern to fit those in, with all structure coming from game rules (which can be gamistic, simulationist, narrativist, or a mix). Akin to a warden of an adventure playground, the Host may use her stake in the development of the story to steer the PCs towards potentially interesting characters and situations, but she does not actually prepare anything, completely relying on improvisation (which will often not even come from herself, but from her players). Instead of placing readily developed attractions, she at best hands raw materials to her players if those run out of ideas what to do. She is still the one responsible for keeping an eye on the bigger picture, though, but her goal is not to move the story towards any specific goal – just to make sure that it keeps moving at all. If she sheds that responsibility as well, however, we finally arrive at the other end of the spectrum:

The true sandbox

In a true sandbox, all narrative elements come either from player decisions, or from simulative rules (often simple random tables). The latter can be used to determine everything which in immersive storytelling would be determined by the Host (either when she prepares the scenario, or improvised during play); including capabilities, dispositions and goals of Host characters.

This means that the position of the Host is essentially gone – she is being relegated to a game rules proxy, identifying which rolls should be made when, and realising the results of rolls on randomizing tables. Since there is noone in control of any plot, or responsible for the consistency of the setting, these remaining Host duties could just be shared among players as well instead, though – there is not really a good reason (apart from player laziness) to burden a single person with that purely supportive role. I, at least, would certainly refuse to take this position; and I consider it a waste of the powerful resource which a true Host can be for a roleplaying group.

Instead of immersive storytelling or even collaborative storytelling, this leads to what I have seen being referred to as „emergent storytelling“, which in my opinion is no storytelling at all: You simply play to see what happens, and retroactively call the result of this a story. I do not find that enjoyable at all – certainly not as a proxy Host, but also not as a player. Some people seem to actually consider this the truest form of roleplaying, though, being freed from any constricting guidance by a Host. I, however, would rather follow the lead of any semi-competent Host than that of a set of randomizing tables, or a too-many-cooks group dynamic of a playgroup sharing or alternatingly possessing creative control – but de gustibus non est disputandum, right?

I just want to point out one thing: Although a player’s control over the actions of his character is maximized with this approach, he still needs to make his decisions with the entirety of his playgroup in mind, even if this only means that he must find ways to have his PC stay with the other PCs most of the time – because if he doesn’t, the other players likely will after a while simply commence play without him. Considering that the necessity of keeping the group of PCs together is only marginally less confining than the requirement to keep the members of that group in a plot prepared by their Host, I honestly cannot understand the zeal with which some roleplayers reject the latter. But then again, not everyone who expresses their preference for the sandbox style is really thinking of the purist approach of a true sandbox, or even of collaborative storytelling over immersive storytelling; and some may actually just have experienced bad or even abusive Hosting, and as a result lost the ability to trust their Host.

(to the IMAGINATION index page)