Sunday, April 1, 2012

Avoiding the RPG Railroad

Going off the rails on a crazy train

You keep using that word. I do not think it means what you think it means.
--Inigo Montoya

Earlier this week, Drance asked the question, "Can one merge the techniques behind railroading and the sandbox style of roleplaying?"  I quipped off an answer, something like "there is no style of railroad campaign, railroading is a technique, not a campaign style".  I've often observed the term bandied about to describe Paizo's adventure paths, or any investigative scenario, as if an entire class or style of gaming is railroad after railroad.  "We don't do adventure paths, because we don't do railroads".

The problem is one person's "railroad" is the other guy's great night of adventuring.  Lots of people buy and play those Paizo adventure paths and seem to enjoy them, so something seems fishy denigrating the whole lot of them.  I had some guy telling me recently the AD&D Ravenloft module was the worst railroad ever written.  I don't know which Ravenloft he read; the one we played was a 6 month guerilla war waged against Strahd and his coven of vampires, where every session tested the player's planning and resource management.  It was an amazing sandbox game.

This goes beyond nomenclature and definitions -  the whole issue is ridiculously subjective.  There is a sliding scale around constraining player choice, with varying shades of grey, and no clear demarcation point where diminished choices become an actual railroad.  Back when a few of us were discussing quantum ogres and whether the DM could fudge dice like Al Pacino, my concern with Illusionism was that if it was well done, the players wouldn't know the difference - so where do you draw the line?

Sidebar:  Illusionism is the technique of giving the players the appearance of choice, but the DM foists his or her own plan on the situation regardless.  The idea behind the quantum ogre involved moving the DM's precious ogre encounter into whatever woods the players entered first.  If you're Al Pacino, you can get away with it, and the players are none the wiser.  But Illusionism is the Railroad's cousin.  If I had to settle on a simple definition for Railroad, it'd be something like this, "imposing a predetermined outcome on the game's events" - at least to get past the semantics.

To give you an idea of the problems with subjectivity, consider these scenarios:  Is a trap with only one solution, a railroad?  How about if the town guard is after the party, and there's only one safe way out of town?  Does your answer change if they robbed someone, precipitating the man hunt?  How about when one NPC after another keeps giving them the same, tired plot hook, leading towards the DM's prepared adventure site? Is a linear dungeon a railroad?

My own solution to the problem of subjectivity was to look at things from the game master's side of the table, and to apply the objective test based on my definition of railroad:  Is the DM predetermining the game's outcomes?  Looking at the situations above, each one of them can be construed as a railroad if the DM has predetermined a singular solution or a particular outcome - if there's truly only one way to beat the trap, to get out of town, where to adventure, or how to proceed.

The interesting thing about blogging, you get exposed to new ideas, you get the chance to reflect and mature your own thinking.  These days I don't believe the objective test goes far enough - the opposite of the railroad is agency and choice, which means the players have explicit influence on the actual events.  The test of whether player choice matters is subjective from the player's perspective.  It's not enough for the DM to do the right thing from the DM's side of the screen, but he or she needs to let the players know they're playing it straight, too.

Consider this pair of railroad examples to illustrate the concern around influence and choice:

  • No matter what the players do, the villain is going to escape - invisibility, fly spell, teleport - the villain will pull something out of the hat and escape to return later in the adventure.
  • No matter what the players say to the duke, they're not going to change the duke's mind.

We've all seen adventures from the 80's that required a villain to make an appearance, do some bad things, and then escape to return later.  (Poor Dragonlance, always taken to the wood shed).  When the players slap a Silence 15' spell on the evil wizard, and he uses a teleport spell to escape anyway, that kind of apparent cheating seems to cross the line - it's like the classic dice fudging problem, not letting the players win when they've played well and should have won.

But consider how the answer changes if the players subsequently learn the big bad guy wasn't actually a wizard at all, but a demon in human guise using spell-like abilities to teleport (thus circumventing the Silence spell).  Likewise, the duke has been mind controlled by his wicked vizier and the players have no chance of convincing him of their innocence because of the enchantment.  It's only when they're surreptitiously visited by the previous adviser to the Duke that knowledge about the charm comes to light and they understand why things unfolded poorly for them.  In both cases, the player's lack of influence can be mitigated when they get the rest of the information - which is why I'm putting an emphasis now on going the extra mile in terms of player-facing information.

Here’s the summation.  The appearance of being railroaded is going to come up in any game - it's subjective, because there's a sliding scale of constrained choice.  The scale starts with resource constraints and mild nudging on the one end of the spectrum, and goes all the way to the DM's heavy hand forcing the action with Mary Sue NPCs, pixelbitching traps,  and undefeatable villains on the other side of the spectrum.  It's important for the DM to avoid predetermined outcomes, but it's just as important to allow the players to influence events through their choices, and ensure that they see that they have influence, too.  The TL;DR answer to the railroad comes back to the simple principle of "saying Yes" during gaming and ensuring that choices matter.

Getting back to Drance's question, my answer stands - there's nothing inherent about a sandbox game that makes it immune to the railroad, because the railroad is a bad technique that pops up whenever the DM starts reducing choice and influence; it's not a campaign style.  There are plenty of opportunities for the DM to introduce antagonist reactions and chokepoints in a sandbox that force the direction of the game, even if the sandbox started fairly wide open and player-driven.  I tend to think what he's really asking is whether the DM can introduce events, antagonists, plot hooks, and other elements of story-driven games into the sandbox, and still give the players a full range of agency.  The answer is YES, but use these things judiciously.

There are campaign styles that are more susceptible to the railroad technique than traditional site-based locations, which is how the adventure paths and story games get implicated.  I'm doing a lot of prep for our upcoming Cthulhu games, and the investigative genre is fraught with peril.  Impactful horror games rely on the big reveal, and if you're not careful, you can find yourself taking shortcuts through nudging and leading and railroading to ensure the players get to the payoff moment.  That's a big enough topic - avoiding the railroad during horror gaming - to make it a separate, upcoming post.

*The picture is Blaine the Mono, from the cover of Stephen King's The Wastelands


  1. Good stuff.

    I've learned to be self-vigilant in making sure the PCs always have choices, even if sometimes there are obviously better and worse ones... or sometimes only bad ones.
    Something I've noticed, though, is that a lot of times players seem to be looking for clues about what the GM wants them to do... like there is a correct path through 'the story'... and that can be a bit tricky. It's like they're expecting a railroad and looking for the rails so they can follow them.
    I've played in groups where my suggestions were shot down by the other players because they perceived a desire by the GM to have us go a different way... and that was kind of annoying.

    1. Something I've noticed, though, is that a lot of times players seem to be looking for clues about what the GM wants them to do... like there is a correct path through 'the story'... and that can be a bit tricky. It's like they're expecting a railroad and looking for the rails so they can follow them.

      I have the same problem sometimes with my players. My solution is to have several obvious hooks available and then to let them choose. If they choose the one that they think I wanted them to choose (whether or not I did) that works just fine.

  2. Thanks for that post. Perhaps it's worth bearing in mind that paths can be strayed from, and players can go their own way if they want, but the path still appears the obvious, often simplest, route to get somewhere - usually where the DM wants you to go. On a railroad there is no choice. You are stuck on it.
    I'm all for player choice (player agency is the phrase of the month), but I get the impression from some posts on the blogosphere that any sort of outside suggestions or encouragement constitutes railroading. Nonsense. Thank you for setting that straight.

  3. Like this a lot.

    Knobgobbler raises a very good point though. A lot of players to try to find the "right" way to deal with and adventure. My solution is exactly the same as it is in improv drama classes. Whatever choice the players make is the right one. It will have consquences, sure, but effectively they write their own adventure. I just say "yes" and roll with it.

    This works surprisingly well. It may mean more backstage work for the GM, but it puts the adventure firmly where it belongs - with the players.

  4. The railroading problem came up in a recent Star Trek-style game I ran. It was an impossible task for the GM(me) to manage a sandbox game in a galaxy with 1000's of unexplored and unique worlds. The players wanted to travel to whatever planet they wished and expected a fully fleshed out environment everywhere they went. My solution for next session will be a compromise. I'll prep a small number of generic planets and wherever the PCs go, it will be one of the prepped planets.

  5. Amanda's advice is real good - if you can make the player's solution the right solution, it takes care of a lot of problems.

    @HoldFast: Trek gaming seems like a tough choice for the sandbox, because the genre implies (to me) "episodic mission of the week from Starfleet" and that's a bit different than free-wheeling, go anywhere exploration. The trick to making "mission of the week" fun is making the missions challenging with free-form approaches to planning and solving them, keeping on with the theme of this post.

    The trick for running the free-wheeling space exploration is to have good supporting tools. That idea of prepping some generic planets is a good approach; one supplement that has solid systems for creating interesting planets for a space sandbox is Stars Without Number. (And it's free). You could ignore the rules and just use the star system generation bits.

    Another good idea is having a conversation at the end of the game session about next week's excursion; it gives them some freedom to pick their destination, and gives you the time you need to prepare ahead of time.

  6. Personally I don't have any issue with linear adventures vs. sandboxes. I don't think either is inherently a railroad or not a railroad.

    I figure it doesn't really matter how many choices of adventures or places or problems the players have, it only matters if they have freedom to try and solve it any way they want and they can succeed or fail (IOW the end result is not scripted). And it doesn't matter how many choices you give them, if each and every one of them can only end the way you decided it must be solved and nothing they do changes that (IOW the end result is scripted).

    The second one is railroading, and it's not fun for me to play or to run. So I endeavor to provide the first whenever possible.

  7. You're right Peter: suggest because something isn't a sandbox it doesn't imply its all railroad. Non-sandbox adventure structures may have fewer choices for players, but the choices they make can still be meaningful and effect the unfolding of future events accordingly.

  8. The problem is one person's "railroad" is the other guy's great night of adventuring. Lots of people buy and play those Paizo adventure paths and seem to enjoy them, so something seems fishy denigrating the whole lot of them. I had some guy telling me recently the AD&D Ravenloft module was the worst railroad ever written. I don't know which Ravenloft he read

    This shows something I have found, that sandbox have become such a loaded word that it's the only true way to play for some very vocal people. I think our little corner of the blogosphere should try to become a little less obsessed with sandboxes. You know what? If people are having fun, they are doing something right.

    Having the path players take be the right one, with consequences, is a really good suggestion.

    When it comes to horror gaming, breadcrumb style, I think it's kind of inevitable that there will be a hierarchy of clues. The other way might be to have many linear strings of clues to follow, and make sure it's at least somewhat feasible to "jump lines" at some regular interval.

    1. Totally agree with you. (I recently found this blog, heh)

  9. Great post! Some great responses too. Its important to consider what the players want. Some groups, usually less expereinced, will want to be led by the nose through a given scenario, and for them that will absolutely be more fun. Usually more experiences players seek the choice and freedom offered by sandbox play.

    Definetly just have a few worlds prepped, as many as you need for a given session. There is no reason to flesh out a whole map, but once a world has been discovered, it should forever after be at that location. Your campaign will grow organically, as did the original "sandbox" campaigns.

  10. I know I'm way late, but I want to add my thanks for a well written post. addressing many of the issues I had with the whole Quantum Ogre-discussion originally. SOME switches matter, but some don't - sometimes you just take generic material to fill in an option that wasn't prepared, and that'll have to do. If the outcome isn't in anyway meaningful to the ultimate goal of the players, it's not railroading.