That's all pretty fuzzy, and I'm hard put to define the notion in more
specific, structural terms that might be useful during game design, so
I'd like to ask:
Structurally speaking, what puts a game on rails?
To help you research, why don't I point you in the direction of an IF
theory paper by Marnie Parker, titled "The Meta-Puzzle of Interactive Fiction"
at http://members.aol.com/doepage/theory.html .
In it, she talks about how a feeling of immersion/non-immersion, or also
a feeling of freedom/non-freedom depends on how an adventure game engages
various senses and mental faculties of a player; and how each player has
a different set of these senses/faculties that they pay attention to.
To add my own two cents in, consider how the term 'on rails' uses the analogy
of travelling; so, if you find your freedom restricted by having to use a conversation
menu instead of an ASK/TELL system, that would not be 'on rails'; but if you
find the exits around you forcing you through the map in a certain fashion, or you
find that the plot tends to go from A to B without letting you visit A' or B' along
the way, that would probably be 'on rails'.
Now, for the second example, there might be many games that overall force
you to travel from A to B. (If you want to get abstract about it, all games with
a single winning solution will make you do that or die.) To use a metaphor,
a walking path is not on rails, while a train car is. A train car will not permit
you to exit while the train is moving, and you are not allowed to commander the
train and take the route at your speed of choice, or reverse the train's direction.
A path, however, allows you to travel up and down the pathway as you please,
and reach out and manipulate the scenery as you travel. Both the path and
the train ride may be very narrow and lacking in choice of destination (which
would be in IF terms, 'linear'), but only the train ride would be called 'on rails'.
Of course, a more casual use of the term 'on rails' might put both the path
and the train ride in that category. I can understand why the term might
be difficult to pin down.
But then again.. 'on rails' isn't even a native IF term in the first place, but one
borrowed from action games which only allow one direction of travel, like
side-scrolling shooters. Maybe writers should just stop using the square-peg
phrase, or invent a system of terms that mean just what they want it to mean,
rather than formalizing it.
It can also be neutral or even slightly positive, when used to mean the
author has deliberately guided you along a path, getting you from one
crucial scene to another without a lot of wasted time. The ability to
wander endlessly in a garden may be a coding tour de force, but if
nothing happens in there, it can become annoying. As the author, you
know that the action happens at the entrance and exit, so you expect the
player to sit back and enjoy the wonderful descriptions. Unfortunately
the player assumes that since there are fifty locations in the garden,
something big must be happening there. Hours of playing time later, the
descriptions long forgotten, they have a vague unease they have missed
(BTW, an engineering project that is "on rails" is one that seemed to go
exceedingly smoothly from start to finish. A vote that is "on rails" is
one that has been fixed. Ditto for a trial verdict.)
Sent via Deja.com
The following article is taken from the online magazine Critical Miss
(www.criticalmiss.com). It's more aimed at pen and paper RPGs,
but I think it's a good one to read for the folks who want to craft a
What Is Railroading?
How's this for a definition:
Railroading is when the GM creates a scenario whose plot is so linear that
it offers the player characters only one route through. Any
attempt by the player characters to perform an action that doesn't fit into
the GMs preconceived idea of how the scenario should
develop will be blocked out of hand.
Why Do We Railroad?
Because it makes GMing easier. There is no way the GM can create a scenario
which allows full freedom of action to the player
characters, yet also takes account of every possible action that they might
In a railroaded scenario the GM can work out each action that the player
characters might perform and create, in advance, a
descriptive, fully worked-out response.
Is It Always Bad?
So, if railroading can make a GMs job easier, is there any "good" way in
which we can use it?
A railroaded scenario is one in which the player characters can only
perform certain actions. What you need to look at is the key
element that is preventing them from performing the other actions.
If this element is believable, reasonable, matches the motivations of the
characters, and fits neatly within the reality of the setting,
then players will readily accept it.
If this element is arbitrary, makes no sense within the setting, does not
match the motivations of the characters, and is clearly there
to make the GMs job easier, players will resent it.
And if there isn't really any reason at all, and the GM is simply using
emotional blackmail to force the players to make their characters
pick the desired option - the players will probably just get up and leave.
Some Examples Of Bad Railroading
"CURIOROSITY DOESN'T KILL THE CAT - ALRIGHT?"
The player character group is journeying from a small town to the capital
city, carrying a valuable cargo of furs that they have been
contracted to deliver to a trader. As they journey down the road they pass
a mysterious cave...
So what? Why should they care? Why on earth would they divert themselves
from their journey and risk losing their valuable cargo,
just to go poking around a hole in the ground that is most likely nothing
more than a hole in the ground?
Because the GM will keep on repeating his description of the "mysterious
cave" - that's why.
As players, we know that something is significant simply because it is
mentioned. But to our characters it's just one more cave that
we have passed. Give us some reason to go in there, other than mere
THE SLAMMING DOOR
Carefully the thief hammers the stout iron spike into the stone floor. Only
when he is satisfied that the door is securely held open
does he follow his comrades into the room.
Then, err... the spike kindof breaks, and the door, like swings shut. And
This is the worst kind of railroading. The GM has written the scenario on
the assumption that an event will occur (in this case a door
slamming shut behind the players) but this has been foiled by them taking
precautions (in this case spiking the door open). But he
ignores what the players have done, and just makes the event happen anyway.
Some Examples Of Good Railroading
MURDER ON THE ORIENT EXPRESS / AIRSHIP
"Well, I guess I'll wonder around looking for somewhere to get a drink."
"There's only one bar, on the upper deck with views of the ground below."
"I'll go there, order a drink, and see if I can strike up a conversation."
"There is a single other passenger there, a Colonel Drendal... you get
talking... suddenly he stands up
clutching his chest and falls to the floor."
If you set your scenario within a contained location, this will constrain
the actions that the players can take. You don't need to worry
about a character going off to buy some kind of wonder gadget that will
trash your plot. You don't need to worry about them getting
bored and wondering somewhere else, just as your intricate storyline is
about to unfold.
'Cous they've got nowhere to go.
"Patrol Group 3, your orders are to travel immediately to the Deltacron
system at full speed."
If your campaign's theme involves the player character group having some
kind of overall motivation or task, you'll find it much easier
to write scenarios.
The classic example here is the Star Wars roleplaying game (from West End
Games, with a new version out soon from Wizards of the
Coast). Since the player characters are all members of the rebellion, you
can create scenarios that have as a hook the mere fact that
the Empire is doing something bad (which the characters will obviously want
Creating scenarios which are written in such a way that there are only a
finite number of avenues that the players can take, is good.
Nobody likes an ad-libbing GM, umming and ahhing and floundering. But
remember two things:
Give the players some choices.
Put your constraints within the reality of the game world.
Low branching/connectivity of some graph associated to the work.
I think the fuzziness comes from quantifying `low' and which graph.
That is, if you realize the rooms of a work as vertices and the
connections between them as edges, then you get a graph.
A completely rail-roading work would give each room exactly one exit,
and you could not go backwards, while a freer work would have
interconnected areas to explore.
(contrast Photopia / Zork)
Similarly, a conversation menu with only one choice every time offers
no freedom, while an `ASK/TELL/SAY' interface offers the possibility
of many choices (and a menu with several choices would be
Another graph would be the possible states of the work: say, if you
have to complete all tasks in a specified order, the graph of work
states would be linear; conversely if there is no order to the tasks,
the states graph would be very connected.
Also, time contributes to linearity in a work: if after X turns
something happens (maybe depending on the character, maybe not), there
is less free space.
The example another poster used of a train vs. a path is instructive:
in one sense, they are both very linear (the locations are linear),
while from another the path is much less linear (b/c one has space to
There is a balance between freedom and direction:
too much freedom and there is no drive to the story;
too much direction and you're just reading a story -- and of course
different works can draw the line in different places for different
I think railroading is generally a pejorative for too much direction.
Fuzziness in exactly how many possibilities are open is an important
point: when it's not clear exactly what will work, there is the
appearance of greater possibilities -- but if it turns out there is
only one right answer when it seemed there were many, the player will
Fuzziness and the appearance of possibilities is key to ensuring a
feeling of control.
So I'd say: decide how much freedom/direction your work will have (at
the begining, if you like) and be aware of it, and of the
consequences; and make it clear to the reader what kind it is
(the first scene/first few rooms should do this).
If you mislead the player they will likely resent it, though this
could be used for effect.
Lots of direction and not much freedom makes for a more story-like work;
Lots of freedom and not much control makes for a more
Also, note that one can shift gears part-way through (notably Spider
and Web, but so is every cut-scene), and that direction/freedom can
operate at different levels, so a common technique is to have areas
that can be explored, and then the story moves on to the next
area/state (see: Spider and Web, Photopia, The Lesson of the Tortoise,
etc., etc.) -- I visualize this as a string of pearls: each pearl
being a bunch of interconnected possibilities, eventually leading
through an exit/story stage to the next episode.
: "On rails" is a common enough term in IF reviews, and I think it's a
: useful one for describing works that lack a freedom of action. Or
: maybe they resist a freedom of the player's focus. Or maybe it's just
: the *feeling* of freedom of action and/or focus that's important.
We had an extensive discussion on this topic on the MUD a few years back
(1997!) that's still relevant to this topic. It's archived at:
and summarized at:
What I find interesting is that this comes from the opposite side: it
asks how to tell a story.
And the answer is to be (somewhat) linear -- the discussion then
discusses ways to conceal this/alleviate the problems.
To add another angle:
If you fragment a story (so the pieces are out of order, like half of
the scenes is Photopia), you can tell a story non-linearly.
Similarly, if you make your story not have dependencies, you can also
have it develop non-linearly, by revealing different viewpoints on a
story, rather than the extention of one viewpoint (over time).
Simple example: character enters an area with 10 puzzles.
Finishing each one reveals a cut-scene, which is not necessarily
dependent on any other (though to understand the story you need to see
all of them).
Thus, a fragmented story-telling structure is well-suited for
interactive fiction. (though of course it can be used in static
fiction, e.g. Pulp Fiction)
I'd even contend that it hits a sweet-spot, allowing both story
telling and non-linearity (though of course other formats can be very
I haven't given a lot of thought to how this relates, but, I have been
playing Outcast (an Action/Adventure game) lately (again), and
experimenting with the limitations/freedom of the game. Here's what I see
in the game:
First off, there are several goals:
There are two main goals, one given you in the intro, the other given in
exchange for the first, upon your arival in the strange world you are
transported to (later, you learn that the two converge).
This second main goal is divided into several parts:
Sub-goals (puzzles) to solve (in any order, though there are events along
the way that propose a particular path, they do not limit you to that
path); secondary goals that increase the ease with which you can accomplish
the main and sub-goals; small actions which seem like a waste of
time/resources at the time, that end up helping you in your end game.
So, as I see it, you could have something like this:
Note: Hinderance sub-goals are the removal of things which slow your
progress, and may even stop it (dying because of these things, for
Now, in text, the way some of these things would be handled would be
different than in this game, but similar things could be done to keep
things from being railroaded completely, while still having everything
funnel down to a particular path. That is, within reason, allow the
story/game to unfold in pieces, which together point to a particular
path/door, while allowing each area to contain a set of accomplishments
that intertwine with the rest.
I know, some people don't like to carry stuff all over the game map, back
and forth, etc., but, let's face it, sometimes it's necessary. I've read
quite a number of stories where, you want to go north to get to the place
where you fight the monster at the end, but, in order to do that, you have
to go in every other direction, just to find out how to kill the monster,
and then gather the materials necessary, then assemble the weapon, then
find out where the monster really is, then get to him, etc..
At this point, we get back into the discussion of presenting long journeys
as such, without adding tedium to the game. Since that's taken care of in
another thread, I won't add anything to it here.
Not that every story has a monster, or anything like that, but, what I
meant with all of this is, that choices and freedom in sub-goals, while
keeping a main goal in sight, makes it easier for the player to swallow
being limited to "going out the red door".
Now, the question comes, once he goes out that door, can he get back? Is
it even worthwhile to worry about getting back, once you go out that door?
How many game states do you, the author want to concern yourself with in
writing the game (if you use the the difficulty/hinderance/obstacle-removal
goals, then you will have to code things in the game that are dependent
upon them, and it takes on somewhat of a flavor of an RPG, though, these
are not because of things you posess, but rather things you have taken away
from the enemy, so they are not the kinds of advantages you normally think
of in an RPG.).
Ok, so I'm long-winded, I guess I will stop here, and see what comments it
Paul E. Bell | Email and AIM: wd0...@millcomm.com | ifMUD: Helios
IRC: PKodon, DrWho4, and Helios | webpage: members.nbci.com/wd0gcp/
Member: W.A.R.N., Skywarn, ARES, Phoenix Developer Consortium, ...
_____ Pen Name/Arts & Crafts signature:
| | _ \ _ _ |/ _ _(
| | (_X (_/`/\ (_) (_` |\(_) (_) (_|_) (/`
I received a comment about Masquerade saying that on REPLAYING the
player felt the game was on rails (though they specifically stated the
game didn't feel that way the first time around). I wasn't sure if I
should be pleased that I managed to give the player the illusion of
freedom of the first time, or disappointed that on replaying they felt
it was only an illusion.
Kathleen (who hopes to release a post-comp version of Masquerade this
weekend. Really. I know I said that last month... but this time I mean
it! Moral: *NEVER* have 7 Parrot Scrapers to look at your game at one
-- Masquerade - http://baf.wurb.com/if/competition00/inform/mask/
-- The Cove - Best of Landscape, Interactive Fiction Art Show 2000
-- Excuse me while I dance a little jig of despair
You should be pleased. Not only cuz that's a better feeling ;-), but
because the romance genre just NEEDS to pull the illusion off once.
It's the immersion in the environment and the identification of
feelings that count, and Masquerade succeeds elegantly on both counts.
Since it reads very much like a novel, OF COURSE the replaying of it
will suffer. The tension certainly will be reduced. The thrill of
meeting exciting people and visiting interesting people will be gone.
Rather than being "in the moment," the player is much more likely to
see the mechanics of the game--"how did I do this differently last
I think many games fit the "on rails" category, it's just a matter of
degree. And I think that it's neither a good nor bad thing by itself,
it's just another design decision.
> Kathleen (who hopes to release a post-comp version of Masquerade this
> weekend. Really. I know I said that last month... but this time I mean
> it! Moral: *NEVER* have 7 Parrot Scrapers to look at your game at one
I'm looking forward to seeing the time-traveler who comes to the dance
in a "Don Johnson" suit. An "Emily discovers vice" thingy. ;-)