D&D General The Great Railroad Thread

[bloodtide]

I've always seen railroading as this is the sequence of events A-B-C. Even if you find a plausible way to bypass B and move on to C, B is going to happen to you anyway.

This does not feel like Railroading. Just about all published adventure modules have a couple different settings, like a swamp or a castle, and have 3-10 encounters in each. If the swamp has three encounters, the DM will want to run all of them.

Honestly, this is 95% of railroading to me. Rails aren't a natural feature of the landscape, after all. They're put in place because trains are too clumsy to go anywhere else. Railroading is when the GMing is so clumsy (or uncaring) that it becomes obvious that things are only happening to compel the players to do something. Then my suspension of disbelief weakens. I'm not thinking about the landscape, anymore; I'm thinking about the rails.

I agree. I hope to offer plenty of advice on how not to be the Clumsy DM.

It gets asked often enough of "how to avoid the bad style railroad?" Though the standard answer is just "do the sandbox".

In every example of really egregious railroading I've seen or read about, I as a player would much prefer the GM simply breaking character for a second and going, "Hey, I really only have ___ prepared tonight. Is that okay?" And then I would either decide that __ is okay, and play along, or if it was really an absolute dealbreaker for some reason, I would excuse myself from the table.

I think the problem here is that it is just too common.

Even at a table of good players that are not jerks they will often enough do things that Inexperienced or Casual DMs don't expect. And it just does not really work to ask the players to go another direction every couple minutes. A lot of players would just seee this as a Railroad too. "Oh we can't do plot a or b or c as the DM is not ready...this game has no agency"

Worse here is too many players are enjoy making waves at best, and hostile at worse. There is a whole legion of bad players that love to get the DM all flustered by doing something unexpected. The idea is if you totally disrupt the DM and everything they do, the game will be better somehow.

Then again I watched a YT video recently where the creator proposed that no one actually plays a "sandbox" game because anytime the GM imputs anything in to the narrative they are denying players agency by determining details of the narrative without player consent. It was a strange take to say the least.

I find plenty of players that agree with this idea. This is the big problem with player agency: many players think they should be free to do anything at any time. And that does not make for any sort of story plot.

Then again I also recently watched a video where the creator stated that a PC dying without express player permission was also denying player agency as the player wasn't given the choice as to whether or not they wanted the PC death to occur, even though the system and dice decided that outcome.

I find this one common too. Though I'm also a Killer DM that lets the dice roll where they may. So many times a PC will die and a player will get very upset. To the point where they yell and demand that their PC did not and does not die. It does not phase me much, as I'll just reply "your PC is dead in my game. The End."

I'm old so at this point the whole dichotomy of "railroad" vs "sandbox" seems a little silly as I believe most games are a combination of both.

True, It really comes down to where you draw the lines.

2. "Suppose there is a final match with the main villain. The DM decided that, when defeated, he will not immediately die, but instead he will first deliver an essential final piece of information to the characters. Suppose this happens and, while he starts chanting off his final piece of evil speech, one of the players have his character cast "disintegrate" on the villain. According to the rules, the villain will become dust. This would ruin the epilogue of the game to everyone. A sensible DM will work around this event somehow, such as granting the villain a ring of counterspell with a disintegrate, even if it wasn't supposed to have it. Only the Master know the villain full equipment in details, and he can use this point to his advantage."

This is, of course, more Metagaming by a player. As again, the player(s) must recognizance story plot drama things like monologues, speeches and final words. As, yet, again, this is a game. Even if the character would act, the player should not. For the sake of the game. This is the Big Problem with Player Agency: many players think they should be allowed to do anything. For no real reason. Ask such a player if it is Okay to ruin a DMs planned monologue and they will happily say yes. Ask them if it is Okay for the DM to ruin their PCs monologue and they will cry "No, Never!". Why? What is the difference?

This is railroading here: the DM is changing things in the game to deny the actions of a player. Though note this is good railroading. The first thought is the above: a player should not deny the DMs Agency and fun. Again this works both ways. To kill the NPC to stop the speech is just a 100% pure jerk move. The next point is that there are other players. Should one player get to decide for them if they get to hear the speech? Should one player be allowed to dominate the game? The answer is, of course, no.

Of course, the above is also Clumsy Dming, and does not need to happen in the first place.

Pro Tip-When you as DM need something to happen, you should Player Proof it. Really this is simple enough. Player Proofing is simple enough, as it just making things in the game that the characters can't effect with "wild wacky actions". If your a DM with high Game and Rule mastery, then by all means use them, if not you might have to Add to the Game. This is a concept that has been oddly lost for a lot of modern DMs. They have the idea that The Game is only what is in the official rulebooks and they are forbidden from adding anything to the game. This idea is silly. And easy way to prevent lots of Railroading is to simply make it unnecessary. And you don't need to add artifacts to blow up the moon, often very, very simple things are all that is needed. A perfect one for #2 is have the villain’s ghost hang around for a minute to tell the information. So no matter what, the ghost will give out the information.

At Least Three Things: For just the “essential final piece of information”,or any such important detail, you should never ever only have just one source for it. This is Metagaming, but no matter how "secret" something is you should always have at least three ways for the PCs to find it out. Have it on a scroll. Or have a minion or two of the main villain know it. Have a rescued prisoner of that villain know it. That sort of thing. So if something does happen to one of the sources, there are still the other two.

 

[payn]

Well, my personal take is probably odd, but as far as I can tell by what is described as what actually happens at the table, is that a "railroad" game and a "linear adventure" play out the same.

Not always, and I wish folks wouldn't make such blanket assumptions.

The GM creates a predetermined plotline filled with specific scenes and encounters and the players follow the GMs prompts from one to the next until they reach the concluding scene/encounter. The real difference is that in a "linear adventure" the players are happy to do that, whereas in a "railroad" game the players want to do other things but are somehow (straight up or by illusionism) forced to follow the predetermined sequence of events.

A linear adventure can encompass far more than that. Often, there are preconceived notions of what the player's troupe is. For example, the adventure path might be one where the PCs are a group charged by a church to search out restless undead souls and give them peace. A pretty specific metagoal to follow session to session. So, it it not sandbox in the sense that folks can just stop being ghost hunters and become pirates or dungeon splelunkers instead. Though, that doesnt mean ghost hunting is done from point A to B to C in sequential order. You can encompass all kinds of sandbox atributed styles such as map and key, mystery investigation, etc..

I think it's the perceived negative light that the "railroad" classification has attached to it that birthed the much newer "linear adventure" classification. Either way, at the table, both playstyles play out exactly the same, the difference being player buy in. I'm not sure why people don't just champion being "railroad" players and shout from the rooftops about how having tracks and predetermined scenes/encounters makes for a better game, as the "linear adventure" crowd often does.

Again, you are conflating two things that are not the same. Railroading is a pejorative in which the game play is bad and not enjoyable becasue the GM has taken all agency away. A linear game is a compromise in which the players agree to some constraints but are still allowed agency in the adventure. The GM may set up an antagonist, or a situation, and the players can still investigate as they see fit. No, A to B, to C even if there are only three items to investigate. The players may investigate as they see fit. The sitaution can unfold based on their actions, in which the antagonists can become reactive or even proactive changing the very nature of A, B, C just like a sandbox or non-linear adventure.

What if sandbox was called litterbox pejoratively instead? Where there is no direction, no point, but to fancy a deuce in the GMs world as you see fit. Would you champion "litterbox" style of play? Or would you prefer folks not conflate litterboxing with non-linear adventure style RPGs?

Then again I watched a YT video recently where the creator proposed that no one actually plays a "sandbox" game because anytime the GM imputs anything in to the narrative they are denying players agency by determining details of the narrative without player consent. It was a strange take to say the least. Then again I also recently watched a video where the creator stated that a PC dying without express player permission was also denying player agency as the player wasn't given the choice as to whether or not they wanted the PC death to occur, even though the system and dice decided that outcome.

Exactly, even a "sandbox" has walls. There isnt really such thing as total agency. Every game group is going to have their own ideas on who controls which aspects of the game. Both GM and player will have some buy in. However, a lot of folks want to boil it down to total agency or zero agency when really its a spectrum.

I'm old so at this point the whole dichotomy of "railroad" vs "sandbox" seems a little silly as I believe most games are a combination of both. I claim to run a "sandbox" game because I want the players to be the primary drivers of the narrative and create a story that emerges through play. I'm also seriously bad at guessing what the players will do and instead rely on real time prompts during play to inform me on what elements I should be adding to the narrative. The again, as a player I will happily sit back and wait for the next quest marker to pop up so I can dutifully follow it as I am usually interested to see where the GM is going with the story.

So yeah, does that mean I'm a "sandbox" GM but a "railroad" player?

It's entirely possible to GM in a certain style, and perfer to play in another. Which is why I think its important to be respectful of diverse opinions.

 

[bloodtide]

It's one part. Another is that, even when two people completely agree about what "railroading" means, one will say it is a completely good and wonderful thing, and the other will say it is a horrible awful thing. I tend toward the latter camp myself.

I'm in the other camp, of course.

"Railroading occurs when the GM (or author) forbids a player's character to take an action that is reasonable, warranted, and within the system's scope, simply because the GM (or author) does not wish the character to take that action." This, of course, depends on "reasonable" and "warranted" (I should hope that "within the system's scope" is reasonably clear on its own?), but I think we can appeal to a common-sense use of these terms. A reasonable action is one that, if you suggested it to some random bystander, said bystander would agree that it has a sensible, well-founded reason for being done, even if that bystander wouldn't do it themselves. A warranted action is one that follows from the situation at hand and the information available--so a character defending herself against what looks like a threatening monster is warranted, but a character brazenly and unexpectedly attacking a king who is surrounded by his retinue, in his throne room, is probably not warranted.

I would just add to this Metagaming , as I think it's important that a player should not be able to ruin a game for everyone just as they are a jerk.

Now, it's possible that the GM might know things the player doesn't that make an action unreasonable. (I don't think that's possible for being unwarranted--knowledge the player doesn't, or shouldn't, have is specifically what makes many actions unwarranted, e.g. "metagame" actions like knowing that trolls are weak to fire.) If so, the onus is on the GM to explain why the player is mistaken. If they cannot do so in a satisfactory way, that's still railroading, even if there "really is" an explanation. It is quite possible to railroad while trying not to!

Wait, you lost me. It's Railroading if a player is clueless? This is sort of saying like 95% of all players are railroading themselves?

Only in principle. In practice, the GM is the origin of all consequences, and thus it is quite possible for "having consequences" to be VERY VERY much railroading, if the GM is adjudicating in a biased or deceptive manner.

True.

Not in my experience. Rather, I find that many players have seen too many GMs who are not very good writers, and thus said GMs have to resort to a lot of force, arbitrary/capricious/deceptive GMing tactics, because the GM forgot to close several major loopholes.

I agree. I put this as one of the Top DM problems: not knowing how to write. And it is not just "writing a novel", it is "writing an RPG scene, encounter, plot or story.

Plot hooks may or may not be railroading. That's all in the execution. But if the GM is "resist"ing, solely because of the effort she invested, that tells me she's getting a little too precious about her time investment and maybe needs to reevaluate some things. Especially "why are you doing so much prep work for things that have a high chance of failure/breakdown and then blaming your players for not reading your mind?"

This again falls under bad writing, or even bad game craft.

"Clumsy Dming" as you call it, is the cause of a great deal of railroading, yes. A lot of GMs are clumsy. That's part of what makes them mediocre, rather than good or bad. (Few outright bad GMs are clumsy; bad GMing generally requires more skill than that. The irony, that to do the greatest bad, you must have some virtue!)

Agreed. Most clumsy DMs are just inexperienced or have never really been taught how to DM. Though I do count Casual DMs as bad because of their "whatever" attitude.

I would not use such a limited definition of "metagame". As noted, a player knowing that trolls are weak to fire would be metagame knowledge, but it has nothing to do with things in real life affecting the game. Likewise, the "character creation metagame" is entirely within the game, it just isn't within this specific world until the character actually progresses.

Metagame fits though, right? It is what you call something done Within the game for Real Life reasons.

For example, no matter what each PC is as their personality and backstory they Must be a group of adventures together, because it would be wrong to tell a player "all the other PCs hate your character so you can't play that character".

Same way it is metagaming when a player has their PC like an other PC that is a hag or harpy just because they find the player of that PC attractive and want to flirt.

When a DM make a clue easy to find it is metagaming.

First: Use something like "progress step" or "event" rather than "plot point". "Plots" almost always get GMs thinking too hard in railroad-y directions. Creating steps or events, on the other hand, helps to keep GMs focused on what matters: there is a situation in the world, which invites the players to respond. A good GM prepares situations--"frames scenes"--that are inviting in ways that the players will innately, spontaneously want to respond. They don't have to be led by the nose, they do it because it excites them, or intrigues them, or challenges them, or whatever.

Plot Step? Event Point?

Again, this goes back to good writing......and more to the point good RPG design.

Second: Remember the important exception: "...unless you have the players' buy-in out of character." This is most obvious with a campaign pitch. In most cases, the very very very start of a campaign kind of needs something to just be declared to be true, or needs to just spontaneously happen around the PCs in a way that gets things rolling. Out of the Abyss, for example, needs characters born or sold into slavery under the Drow. Can't really do the campaign if that doesn't happen. Similarly, if a homebrew game ends one adventure and the GM says, "Hey, how do you guys feel about doing a swashbuckling pirate adventure? I just got a cool module we could run", that's not railroading if the players explicitly buy into it. It would be railroading to force the players to get on a ship and thus have a pirate adventure--but it isn't railraoding to have an adult conversation with your players and ask if they think a pirate adventure would be interesting to play. In other words, don't do things that rely on the PCs' unprompted participation--either make something that they'll innately just WANT to do, or tell them what you're aiming for, so they're going in eyes-open, understanding your goals.

I do second having such conversations.

So, amended, I would say this pro tip reads...

Pro Tip: Never make an inciting incident, or really any major progress step, dependent on the PCs' actions, unless you have their buy-in out of character. In this example, don't have a riot that depends on the PCs' unprompted participation to develop into an actual adventure.

Agreed. Though I would add on the even more "don't do it even with an agreed by-in". A lot of players don't like the buy-in concept. They will do it, but endlessly complain "guess we agreed to do it so we must whatever" and then sort of half-play as they are unhappy.

It is best to have Inciting Incidents that really hook the players....hook, line and sinker. Make a good one, and the players will jump on it very willingly.

At Least Three Things comes into play here too. It is often best to have three or more reasons for the PCs to do any task, adventure, plot or story. The easy way is to have all the PC be Lawmen, then it is simple enough for them to go after criminal. More complex is to give each PC a unique reason. One of the nice things about 5E is the bit of the focus on affiliations and factions. This gives PCs built in reasons to do things....assuming the players role play deeply.

 

[Jacob Lewis]

Here's my attempt to describe a "limit case" for this sort of play:

The GM maps and keys a dungeon that is a single corridor, with doors along it and rooms leading from it. The players can travel the corridor, using ESP and Detection spells (and perhaps other exploration techniques) to get a sense of what is behind the doors, and then come up with a strategy about how to tackle the rooms, in the manner Gygax describes in his PHB.

The GM, in keying the rooms, includes opportunities that reward one approach rather than another, and hence reward good player planning: eg it's good to get the dragon-slaying sword from the Ogre room before you open the door to the dragon room; and it's good to get the hammer of giant-slaying from a bargain with the Dwarves before you tackle the Ogre room; and maybe there's another room, across the way from the Ogres, with some info about the Dwarves like Balin's tomb in Moria; etc.

This would not be the most exciting dungeon of all time, and its map is fairly boring, but I don't think playing through this would have to be a railroad-y experience.

CODA:
Classic Traveller is published in 1977. The Traveller Book (1982) compiles, and adds a bit to, the 1981 revision. There is nothing in the 1977 edition that suggests railroading by the referee. Whereas there is that sort of thing in the Traveller Book. To me, this change seems consistent with a broader hobby-wide change that began in the early 80s and was cemented by the end of the 80s. Which is the change to a GM-centrism that is also very railroad-y in advocated approach.

I think your dungeon example illustrates the distinction I was trying to get at. Yes, players have meaningful choices within the map-and-key framework—where to go, what to risk, which tools to secure first. But those choices are all still contingent on the GM having designed the structure, keyed the rooms, and decided what information can be learned through exploration. That’s not necessarily a bad thing, but it is exactly what I mean by “GM-centric.” The authority is still centralized, even if it’s exercised in a way that avoids a railroad-y feel.

Which is the point: the discourse about railroading has always been about the experience of constrained agency, not just the surface techniques. In the 70s it was framed around “do my choices as a player actually affect outcomes?” In the 80s and 90s it leaned into “am I being forced through a prewritten story?” More recently it’s “am I allowed to help shape the fiction at all?” The term evolves with the expectations. Your corridor-dungeon isn’t railroading in one sense, but in another context, it might be read as exactly that.

 

[zarionofarabel]

What if sandbox was called litterbox pejoratively instead? Where there is no direction, no point, but to fancy a deuce in the GMs world as you see fit. Would you champion "litterbox" style of play? Or would you prefer folks not conflate litterboxing with non-linear adventure style RPGs?

Yes! Yes I would! I would also champion players trying to take the biggest duce they possibly could!

I often see the argument that "sandbox" games are meandering chaotic random messes that are ultimately meaningless. Yet I myself have neither ran nor participated in a so called "sandbox" game that was any of those things. I have however participated in a few "railroad" games where as soon as the players tried to do things the GM hadn't prepared for, the game either shut down completely or the GM had a really tough time improvising things to continue and it often turned in to a meandering chaotic mess. I don't think either playstyle is better or worse, and I admit that having the GM be the primary narrative driver and plot creator seems to be far more common than the onus for those things be in the hands of the players. I guess I'm just a lazy GM and I'm bad at figuring out what kind of storyline the players will be keen on experiencing, so I leave it up to the players to point me in the right direction. I also don't prepare plotlines, scenes, or encounters, preferring to improvise those things in real time at the table. I do engage in world building (or use published settings) but I do think that "adventure prep" and "world building" are two distinct things.

Of course, I think I'm also weird in that I love it when the players take a big duce in my litterbox and I have to scramble to deal with the consequences. Having to deal with crazy players doing crazy things is half the fun!

Exactly, even a "sandbox" has walls. There isnt really such thing as total agency. Every game group is going to have their own ideas on who controls which aspects of the game. Both GM and player will have some buy in. However, a lot of folks want to boil it down to total agency or zero agency when really its a spectrum.

That I can definitely agree with. I find it very confusing that many people seem to consider the GM doing anything with the narrative as somehow removing player agency. If the GM can't manipulate the narrative in any way without it being seen as them impeding on the players agency, how would the GM continue to have any role in the game? I think at this point the definitions of "sandbox" and "railroad" playstyles has been stretched to the point where they have become absurd to the point of being effectively meaningless. Long live the "sandroad" and "railbox" playstyles!

It's entirely possible to GM in a certain style, and perfer to play in another. Which is why I think its important to be respectful of diverse opinions.

Well, you know what they say about opinions...mine included!

 

[pemerton]

I think your dungeon example illustrates the distinction I was trying to get at. Yes, players have meaningful choices within the map-and-key framework—where to go, what to risk, which tools to secure first. But those choices are all still contingent on the GM having designed the structure, keyed the rooms, and decided what information can be learned through exploration. That’s not necessarily a bad thing, but it is exactly what I mean by “GM-centric.” The authority is still centralized, even if it’s exercised in a way that avoids a railroad-y feel.

Which is the point: the discourse about railroading has always been about the experience of constrained agency, not just the surface techniques. In the 70s it was framed around “do my choices as a player actually affect outcomes?” In the 80s and 90s it leaned into “am I being forced through a prewritten story?” More recently it’s “am I allowed to help shape the fiction at all?” The term evolves with the expectations. Your corridor-dungeon isn’t railroading in one sense, but in another context, it might be read as exactly that.

My own take (and it's not really mine - it's pretty influenced by Robin Laws) is that RPGing pretty quickly expanded to include a whole lot of participants who weren't really looking for a problem-solving/wargaming experience. This was encouraged by the single-character focus, the tropes and colour, the use of figleaf backstory even in the most Gygaxian of dungeons, etc.

And as soon as participants want (something like) "story" or "character", then the Gygaxian model of play stops working. And this feeds into some of those generational changes you point to.

 

[bloodtide]

Well, we seem to be on the same page.

Great!

This is called a "Rowboat World". You can go anywhere but you have to put in all the effort and the vast majority of what you find will be just empty, meaningless and featureless. It's an example of a dysfunctional sandbox the way a railroad is an example of a dysfunctional adventure path.

I've never heard this one. It is what I often see in a Sandbox game.

The defining trait of a sandbox is that you prepare far more content than you intend to use. Sandboxing is really hard and requires tremendous effort on the part of the GM. It's basically only possible if you are spending as much effort on the game as a full time job, which is why I haven't been in one since college and have never really tried to run one. I do have relatively open areas "a city", "a dungeon", "a valley to explore", or "an island" but these are really "tiny worlds" with a few well chosen things to do and a plot hook which focuses the players on trying to find something or solve some problem or mystery within the open area.

Odd, most people I know seem to think the Sandbox game is super easy and all just effortlessly improv. The players do whatever, the DM does whatever, and the whole game is whatever.

The Dragonlance adventure path was the first time I remember this becoming a big debate, precisely because by the text the adventures had a lot of very heavy handed railroading advice. Railroads had occurred before in D&D, but not campaign length ones and not with just heavy handed devices for keeping the players on track and certainly never before with stage direction telling the player how to play the character.

Of course, before Dragonlance, few D&D adventure modules had any plot or story.

Many people said, "These are terrible, it's not even real D&D" and other people said "Well, the guard rails are there just for newbs, you don't have to force things on to the rails if you don't want to."

And some people were like "dragonlance is so cool....I want to be a Kender Red Wizard Dragonrider!"

 

[Celebrim]

Odd, most people I know seem to think the Sandbox game is super easy and all just effortlessly improv. The players do whatever, the DM does whatever, and the whole game is whatever.

I just don't even bother to play with people who tell me that the improv a sandbox anymore. Honestly, at this point, if someone tells me how good they are at improv, it's just a massive red flag. The whole game is likely to be, "Whatever."

Of course, before Dragonlance, few D&D adventure modules had any plot or story.

Certainly, quests weren't at the forefront of play until Tracy Hickman, but I think that that is a bit of an exaggeration plot or story was rare or some new innovation. GDQ was the original adventure path, and it has a story - fight your way through the minions to get to the powerful beings behind the giant's attacks. And you even see something like railroading in A3 and A4 where the intention is for the PCs to be inevitably captured and then have to break out of prison in the beginning of A4. Plus modules like U1 and UK1 very much are plot heavy and if not in their execution at least in their conception would stand up today.

 

[overgeeked]

I could be mistaken, but I think it’s a Forge term. One of many they came up with for various solutions to what Ron Edwards called “the Impossible Thing Before Breakfast” - the inherently self-contradictory notion we nonetheless take for granted about RPGs, that the GM has complete control over the story, and the players have complete control over their characters, who are the main characters of the story.

The way he viewed it, these things can’t both be true simultaneously, and there are a lot of different techniques that have been employed, usually unconsciously, to try to resolve this by limiting either the GM’s control over the story, or the players’ control over their characters, or both, in a way that will be tolerable to everyone in the group.

Like most Forge nonsense, it’s based on a half-baked faulty premise. The referee does not have complete control over the story. The story, such as it is, emerges from the interplay between the referee’s prep, the roll of the dice, and the players’ choices. There’s no Gordian Knot to cut. It’s an illusion.

 

[cbwjm]

This does not feel like Railroading. Just about all published adventure modules have a couple different settings, like a swamp or a castle, and have 3-10 encounters in each. If the swamp has three encounters, the DM will want to run all of them.

And if the players do something that plausibly allows them to bypass those swamp encounters but the DM forces them through them anyway, that seems a good definition of railroading to me.