Why I Dislike the term Railroading

[pemerton]

the flow of an adventure, of the structure (is this what is meant by scene-setting?) or let them 'skip to the end'.

By "scene-setting" I've been meaning the opening and closing of "episodes" or "encounters" in the game - scene-framing and scene-closure.

This is separate from structure, I think, although related to it.

When you talk about "flow" or "structure" I think of A > B > C vs A > C > B (or maybe B is now redundant, if the players have "skipped to the end"). Ron Edwards describes the control of this as "plot authority". If it's all under the GM's control - if nothing the players do can influence the sequence in which these events happen - then I think we have a railroad by most people's understanding.

Of course, what counts as an event, in the relevant sense, is game and group relative - eg for some D&D players shopping at the fair is an event, and for many D&D players overland travel is an event, whereas for most superhero players shopping is not an event, and for many travel may not be either. In my view some of the disagreement in this thread is a result of people not recognising that what counts as an event in the unfolding story of the PCs is not independent of game systems and group preferences.

Scene framing is different - it's getting to specify the starting parameters of the events as they occur (eg is it night or day, is the Duke in a good or bad mood, does he have 5 lvl 2 bodyguards or 10 lvl 8 ones, etc).

In typical D&D play plot authority and scene-setting authority blend together, however, because the action-resolution mechanics are so open-ended - the mechanics are mostly task-resolution, and the players are always allowed to have their PCs attempt a task. And conversely, the official rules have very little to say about how we might move from one episode to another other than by the players describing the tasks that their PCs undertake.

If you want to play a game in which the players have a lot of plot authority but the GM has a lot of scene-framing authority, then you may need to constrain the action-resolution mechanics in some way, either by convention at the table, or by building it into the rules themselves (perhaps by switching to conflict-resolution mechanics). Skill challenges in 4e are the first example I can think of in D&D that create the possibility, at the purely mechanical level, for a separation of plot authority from scene framing - players can, in principle, choose whether to attempt to resolve situation X or Y or Z (and thus exercise plot authority) but the GM gets to establish the initial parameters for each skill challenge, and the skill challenge mechanics themselves tell us when the scene is resolved (see below for more on this).

What I'm describing might be a bit closer to narrow-wide-wide, where the start of an adventure is fixed but both the methods and conclusion are more open.

This is a pretty typical way to play D&D - the GM describes the opening situation in the first session of the campaign, and then it expands from there. Is that railroading? I don't think so - it's just "playing D&D".

On the other hand, in Sorcerer it would be outrageous railroading, even cheating, by the GM, because in Sorcerer the players specify the starting situation, by building a so-called "kicker" into their PC description.

I have seen railroaded used in a non-pejorative sense fairly often also, just to add to the confusion over the term. So railroaded doesn't necessarily mean bad, though it usually does.

I'm going to suggest an alternative definition:
The GM in an rpg has a lot of power. But the GM's presence is normally invisible. He acts thru NPCs, the environment, the world. The players can suspend their disbelief and believe that they are living in a secondary world. Railroading occurs when the GM's normally invisible hand becomes visible.

<snip>

By this definition, railroading is not necessarily a bad thing. Okay, it's bad that the players can no longer suspend disbelief, but that might be the lesser of two evils. It can be beneficial for the players to know where the plot is, to know where the GM's prepared material is, to be directed toward the adventure. This is where the visible hand, pushing, is a good thing, because it becomes perfectly clear what the PCs should do.

Interesting.

I tend to follow the Forge in my terminology, and so prefer to use "railroading" only for the GM using force to deprotagonise the players. But as I've already said, what counts as deprotagonising is game and group relative.

As to the GM's visible hand - after reading Edwards' discussion of ouija board roleplaying, and noticing that that is applicable to some of the sessions I've GMed, I've become a bit more ready for the GM to use force, rather than faff around hoping the action resolution mechanics will get the game to an interesting place. But I don't like to deprotagonise my players. I've therefore become more attracted to games that impose a clearer limit on action resolution mechanics (see 4e's encounter powers, milestones etc), to give the GM a bit more freedom in scene framing, and that at the same time give the players more authority over the plot, by enabling the players to force encounters to a conclusion - most fantasy RPG rules have this feature when it comes to combat (conclusion = foes at 0 hp) but 4e also has this feature in noncombat situations, via skill challenges - after at most 14 primary skill checks we'll all know the outcome one way or another! The GM can't just keep stringing the players along.

Here's a couple of examples of railroading, one good, one bad, that I experienced as a player.

Good:
Penultimate session of a fixed length campaign.

<snip>

My character was amazingly physically capable and I felt she had finally been properly tested by a physical challenge and got to show what she could do. I never felt that I must be able to succeed for the encounter to be fun. I was fine with following the rules of 'story'. I mean, what's one more tier of rules in an rpg? There are already so many.


Bad:
The PCs were defending an underwater city from attack by raiders.

<snip>

It was a setup, there was no way we could have successfully defended the city.

<snip>

Why is the second one bad and the first one good? In the former we still got to be big damn heroes, it was only delayed by one session. We didn't majorly fail. In the latter we had a huge appalling, terrible failure that had all been planned by the GM, for no particularly good reason imo, and, maybe worse, it was completely implausible. How could the city have lasted for so long if its force field went down to one blast from raiders they've been contending with for years?

There was a sense, with the first GM, of essential benevolence, of using the railroading for the purpose of telling an entertaining story. Whereas with the second GM I get the sense he was saying, "Hah! Gotcha! You thought you were heroes? Well you just lost a city, not so heroic now, eh?" Can't be good, that.

I don't want to be too forward in analysing sessions of play that you participated in and I didn't - but in the first, it seems to me that what was at stake was not only beating the bad guy, but succeeding in expressing your PC in a way that made her look heroic rather than useless. (Edwards has interesting things to say about the importance of this - ie of the mechanics allowing for PCs to fail at their overall goal without the player having to envisage his/her PC as a failure.) This latter was achieved. And then you got to beat the bad guy pretty soon after. Overall, very little or no deprotagonisation.

Whereas in the second scenario, you are forced to play out your PCs failures in a way that doesn't at all express your conception of your PC.

I can think of at least two ways to handle the "fall of Atlantis" which are in principle viable, although both will misfire if the players aren't interested in playing a game that deals with the consequences of the fall of Atlantis. One is to simply skip over the battle - the GM explains to the players that the forces are overwheliming in number and firepower, and despite the PCs best efforts as defenders the city falls. The question for the game now becomes - What do you do, having lost the battle? Approached this way the fall of Atlantis is not an encounter at all, it's just a prelude to actual play - so protagonism and deprotagonism are not even on the table.

The second is to play out the battle, but have the interesting aspects of play not being the military stuff - which the PCs can't win - but the social and emotional aspects of the downfall - who do the PCs choose to take with them in the escape pod, for example, and who do they leave to die or be taken prisoner? This would help set up some further context for playing out the consequences of the downfall, and would allow the players to express their PCs in ways that aren't predetermined by the GM.

But obviously neither of these approaches would work for a typical fantasy or science-fantasy game where the whole point of the game, from the players' point of view, is to be heroic defenders of Atlantis. In that sort of game it is obvious railroading for the GM to build in a defeat from the get-go.

This also confirms that railroading is not just about action resolution mechanics but also about scene-setting. In the fall of Atlantis, as you described it, the GM didn't fudge any die rolls in the action resolution - it's just that he framed the scene in such a way that victory for the PCs was impossible (a bit like the notorious 1st level PCs vs Ancient Red Dragon). Another reason I like 4e D&D is that has robust encounter-building guidelines that give me, as GM, a much better handle on how to frame scenes that will preserve player agency during combat rather than steamroll over it with game-mechanically-unbeatable monsters.

 

[pemerton]

Here's the transformation I've, personally, seen over the past 5-6 years.

<snip>

Before: Railroading is bad gaming. It's a degenerate state of affairs where a DM basically plans both the adventure and the PCs' reactions to it ahead of time. Most problems only have one solution, and there's only one right path - and deviation from that path is severely punished. Stuff like mandatory capture scenes, unavoidable ambushes, NPCs you can't attack, and so on are hallmarks. Railroads are best defined by their inflexibility during play, the lack of player agency over their characters' actions, and the arbitrariness of the DM's decisions. It's to be dreaded as a player, and avoided as a DM.

After: "Railroading" can include anything from linear adventure structures (where the PCs actions still aren't constrained by fiat), to branching flowchart adventures, to even presenting an adventure plot to your group up-front rather than having them pick one from a list of rumors.

<snip>

"It's okay if you like to be railroaded" is more than a bit condescending; the term has historical negative connotations, and pretending they don't exist anymore is a dodge, at best. I think it's more often used in a "my-sandbox-gaming-is-better-than-your-story-gaming" manner. And, frankly, I dislike that this redefinition has been allowed to happen without a fight.

Your diagnosis seems plausible.

I think part of what has allowed the "redefinition" to take place may be that many players who emphasise sandboxing are mostly familiar with games like D&D, Runequest and Traveller, or at least look to those games as paradigms of RPGing. In most versions of D&D PCs begin play as blank slates, and become defined as characters through the course of play. In this sort of game, too much GM constraint will be experienced as railroading, because the players won't get the opportunities, ingame, to define their characters. (Of course, for those players who don't particularly care for a character focused game, but rather are happy to try and win at the encounters the GM provides while enjoying the story along the way, there won't be a problem - I assume quite a bit of adventure path play is like this.)

In Traveller or RQ, PCs start with randomly determined characteristics that reflect the gameworld rather than player metagaming. Not blank slates, but nevertheless it is only in the course of play that a player gets to put his/her mark on the PC. Again, for players who care about this, they will need to be provided with some choices by the GM. Otherwise it is the GM who is making those marks - again, classic deprotagonising railroading.

But in a game in which the players get to metagame their PC creation and make their mark on their PCs from the start - a lot of indie games like HeroWars or The Burning Wheel, but also traditional games with open-ended point buy like Rolemaster and Hero (assuming the GM doesn't impose "realism" or "gameworld reality" constraints on character building), fit this description - then the player, in building his/her PC, is more-or-less telling the GM what sorts of situations s/he wants to play out. A GM who doesn't provide, and instead just plops the PCs down in the sandbox, is making those character creation choices meaningless - which may be just as deprotagonising as a railroad can be.

One of the weirdest thing to me, then, about the "redefinition" of railroading, is that these sorts of character-driven games- in which it is the GM's job not to build a sandbox in which the PCs wander around, but to cut straight to the chase by introducing situations into the game that speak to the attributes and relationships the player has built into his/her PC - get redescribed as railroads. Whether or not this is condescending, it's just radical misdescription.

 

[Ariosto]

"Railroading' is a phenomenon, a behavior. To apply it at every single decision point would an extreme. To apply it less often is not the same as not applying it at all.

All I know of Paizo's Adventure Paths is what I read of them at second hand. I'm looking at Wikipedia entries.

Statements that the PCs "do" this, or "must do" that I will take with a grain of salt. A certain amount of presumption is in the nature of the beast, at least as presented to the prospective DM. There tends to be a bit more in event based scenarios, regardless of how much railroading they may entail. The impossibility of covering all possibilities is part of what often allows the site based approach to get more mileage out of so many words.

It's up to the DM to infer, extrapolate, elaborate, modify, and otherwise adapt the material to the course of actual play -- however it is presented.

When I see it related that, "In this chapter, the heroes must find a missing paladin," and that in the next the players are "following their warning from a dying paladin", then I suspect an extra helping of assuming going on.

"The heroes must march on the Cagewrights' headquarters, in the heart of a volcano, before their ritual is complete and a permanent gate to Carceri is opened above Cauldron." If they don't then I'm guessing that the next/final two installments of "Shackled City" would require revision.

However, for the most part, that "path" does not appear (from my limited information) to have chapters dependent on certain outcomes in previous chapters. There might be possible situations, especially in a more wide-open campaign, that would make some chapters extremely implausible. (There is often a big practical difference between requiring a very particular condition and requiring anything but such a peculiarity.)

That's pretty typical of published event-driven scenarios, though, and one reason I don't find them much worth my while as a DM. Something like that could introduce the temptation to 'railroad' events, even though the scenario in itself does not suggest any such thing.

As to the structures of individual chapters, I really have no clue. They could require as careful rigging as an M. Night Shyamalan movie, or be as wide open as a basketball match.

 

[pemerton]

it's really only a railroad when the DM is willfully removing the ability of the player to change a situation they should be able to, and are actively trying to, change.

And the "should" and the "trying" here are both game and group relative.

In the case of the enemy getting there first:

If the PCs have done nothing to try to prevent them from getting there first, then simply writing "the enemy gets there first" in the adventure wouldn't be a railroad.

It would BECOME a railroad though if the players actively try to prevent that scenario, but the DM simply decides none of what they do has any effect purely to force the enemy getting there first.

It's interesting that, in this paragraph, you move from PCs to players. From the mere fact that the PCs are trying to stop the enemy getting there first, we can't tell much about the play epxerience if the GM determines that they will automatically fail.

Suppose the players say "We try to stop the enemy getting there first - we mount our horses, load our biggest wands and crossbow, and head off". If the GM simply says "Despite your best efforts, you fail - the enemy is there first - but make skill checks to see how exhausted you become in the chase" then we have pretty aggressive scene-framing - maybe too agressive for many players! - but the GM hasn't wasted the players time and hasn't strung them along.

Has the GM cheated? This depends on what the rules of the game say about action declaration and action resolution. In most versions of D&D this would be at least a mild cheat, because the action resolution mechanics are all about miles per hour of speed, chances to hit with crossbows and wands, etc, and none of that has been brought into play in resolving the situation. In games which give the GM more liberty in setting DCs, based on considerations not of ingame causation but narrative flow, and which also give the GM more express scene-framing power, this may not be cheating at all. (Robin Laws, in chapter 1 of DMG2, tries to introduce some of these techniques into 4e. I personally don't think he succeeds very well, precisely because he fails to reconcile them with other, inconsistent, aspects of D&D's action resolution mechanics.)

Suppose, however, that instead of doing what I've described above the GM lets the scenario play out, speeds are calculated, maps looked up, dice rolls made to hit, damage rolls made, and so on - but the GM just stipulates (without actually applying the rules) "You don't do enough hits/damage to stop your enemy, and at the last minute he outflies you to arrive their first" then you have something that looks to me much closer to Doug's story of the Fall of Atlantis - the GM is wasting the players time, letting them go through the motions of playing the game when in fact s/he has no intention of actually treating it as part of the gameplay at all. This is illusionism at best (if the players don't know because they choose to be deceived) and railroading - or, if you prefer, cheating - at worst.

 

[pemerton]

Any adventure that destroys the campaign world if the Pc's don't snap to is a railroad.

Why? The players, after all, can choose to have their PCs (i) join with the cultists, or (ii) teleport to another plane that won't be destroyed, or (iii) try to stop the destruction, or (iv) recongise the inevitability of the end of all things and spend the final days trying to comfort their loved ones/resolve any remaining grudges/die with dignity rather than in terror/etc.

If the players have signed on for an open-ended D&D campaign where the basic premise of play is that their PCs will keep getting stronger and stronger until everyone gets sick of it and decides to try something else, then maybe the end-of-the-world scenario is a railroad. But a lot of games aren't like that - including 4e, which has an endgame (lvl 30 and the Destiny Quest) built in.

 

[TheAuldGrump]

Wait... So the PCs would be fully capable of hypothetically freeing the prisoners on Day 1 and stop the Evil Vizier from transporting them to Sunburst Vale? (Perhaps forcing him to kidnap more victims on Day 3, and similarly postponing the rest of his plans unless they catch him first.)

I'm confused. Where, exactly, is the linearity in that design? To my eyes you appear to be describing a non-linear scenario. That's not a plot, it's a situation.

I guess we need to come at this from the opposite angle. What would a non-linear version of this scenario look like for you?

*Shrug* I wrote a steampunk adventure that begins with a contact of the PCs getting murdered after sending them a message to meet him the next morning - but I also put in what would happen if they decided to keep a watch on his house immediately after getting his message. (Yes, they could, and did, save his life.)

His death is scripted, but I allowed the option to change the script. Even if he gets saved he may, or may not, be too injured to give them the information that their organization has been infiltrated.

I prefer branching trees to linear stories, it gives both the players and myself more room for change. My starting point when writing an adventure varies - sometimes I do a few timelines for what the villains do if uninterrupted, sometimes I do a flowchart, sometimes both. Both leave the PCs (and villains) a lot of wiggle room.

The flowchart acts as a decision tree, with a general guide of 'if the players do W then X happens, if the players do Y then Z happens. And of course the team sometimes to go with something off the charts, and these can be my favorite moments.

In the steampunk example the assassin tries to get away, whether he succeeded or failed. Sadly, when he breaks into a not yet finished station
of the Underground he encounters a large number of the goblins from London Below - the PCs can intervene, but both times that I have run the scenario they have instead stopped, and watched as what look like street urchins with glowing eyes slaughtered the assassin, not wanting to risk what might happen when fighting the creatures. (Both times the party split up - so it would have been two or three heroes against a fairly serious number of goblins.)

Again, while the assassin's death is scripted, it can be prevented - and chop a good length of the adventure off short, since he has much of the information that they would otherwise need to go hunting for.

Important clues can sometimes be gained in multiple ways, and other clues may or may not be needed.

In your example, perhaps the Vizier changes his hunting grounds, or, if he knows who thwarted his attempt, he may kidnap a contact/relative/friend/love interest of the PCs.

The Auld Grump, losing coherence, need sleeeeep!

 

[pemerton]

the teleport nerfing and the like isn't inherently bad. it's "kryptonite" -- sometimes you need a plot device to promote a certain kind of play experience.

But promoting play experience A when the players all signed up for play experience B is railroading.

Plus, it ultimately goes back to the Player-DM sparring issue. Teleport was put in the game to allow PCs to fast travel, not bypass the adventure. When players started using it to bypass the adventure (probably the second time it was cast) the DM, having done all that prepatory work, said, "No way, Jose" or some Gygaxian equivalent.

The real issue here seems to be that D&D has traditionally included game elements which militate against the play experience that many who sign on to play the game actually want. It's not alone (Rolemaster has worse scry-buff-teleport issues than 3E D&D does, for example).

One way to resolve this problem is to introduce yet more game elements - like kryptonite - that solve the problem. When you read the original AD&D rulebooks you see that this was one of Gygax's preferred techniques - listening at doors getting boring? ear seeker! or, cleric players getting to powerful with their spells? the GM, as god, refuses to grant them!, or, players acting like mass murderers and robbers because this is an easy source of XP? they're violating their alignment and lose levels!, etc etc.

My preference is actually to wind back the problematic game elements. Only put elements into the game that promote, rather than detract from, the desired play experience. Doing it this way also reduces the sense of player vs GM sparring which ingame nerfs tend to lead to.

 

[pemerton]

IMHO, railroading is "the usurpation of player agency by the GM". The term "usurpation" does mean that the GM is making decisions which are legitimately the players', in terms of the ruleset used and the social contract at the table.

Fully agreed. And for this reason, it's always a poor way of playing the game. But it's also highly table-relative as to whether or not any given episode of play is a railroad or not.

EDIT: For a certain type of D&D play, player agency is almost entirely focused on making tactical choices during combat. For these sorts of players, even the most linear adventure path would not be a railroad, provided that the GM actually lets the players play out the combats, and provided that, if the AP depends upon a certain NPC surviving, the GM simply makes it clear that there is no combat going on in relation to that NPC.

 

[Ariosto]

By "scene-setting" I've been meaning ...

If all that works for you, then that's groovy.

Me? I feel like Achilles with a tranquilizer dart in the keester just reading that. It's tortoises all the way down!

I just play the game, following the instructions, and it works out.

Of course, the game doesn't mention any "scenes" to "frame", or "plot" over which to have "authority". There is space, and there is time, and events have pretty definite dimensions in the game's frame of reference. How much real time we're going to spend on this fight, or that negotiation, or the ride from A to B ... or Anne's half-orc jokes, or what happened to Carl in a game last week ... is something we settle as a group.

The big difference with something such as Rise and Decline of the Third Reich or Settlers of Catan is that there are relatively few options and the level of resolution is set. Everything is stereotyped.

That does not keep it from being possible to want to "hurry along" or "spend more time considering this move", or to spend more or less time looking up and debating rules.

Gaming is always a social engagement, to the extent that other people are involved!

 

[Ariosto]

I think part of what has allowed the "redefinition" to take place ...

I'm not following your logic. We're in the context of D&D. There's a definition from players "mostly familiar with games like D&D, Runequest and Traveller". Those are all from the earliest years of FRP: 1974-78.

Then there's a definition from ... what? Sorcerer? Hero Wars? Burning Wheel? And that's supposedly the older one?

Or are you seriously suggesting that if a situation in one of those games gets called 'railroading', then all of a sudden it doesn't mean what it meant for decades? That there's some kind of double standard in which the same behavior would be called 'railroading' by the AD&Der if it were somehow associated with Hero Wars, but not if it were labeled as AD&D?

That I have not seen. People can judge whether they would consider something 'railroading' in their old-style games without needing to know the brand name of the rules set used in the case.