Choo-choo! That's the Dragonlance module leaving the station. Prepare yourself for a one-way ride in the land of Krynn as you watch Tanis Half-Elven save the day!
Truth be told, I've never actually played the Dragonlance modules--alas, I missed TSR's heyday--but I've heard that they're legendarily on rails. Good thing we've learned our lessons since then!
Ah. Well. Regardless, I'm here to play the Baatezu's Advocate as I extol the virtues of railroading.
Terminology Check: Railroading, Illusionism, Participationism
Railroad, v. When the player characters get a wild hair and the GM has to corral them back to the adventure he's prepared.
"Railroad" is very vague and not well-defined, so let us instead discuss participationism and illusionism, which are specific modes of play. The definitions of these terms are a bit muddy, and there's not an "official" RPG lexicon, but Darkshire.net has a fairly comprehensive glossary.
Broadly-used term for linear plotting in RPGs. (1) GM behavior when the planned scenario requires a particular sequence of events/scenes leading to a particular ending. The GM ensures that it arrives there by a variety of means. This is generally pejorative, but is sometimes defended as valid as long as it is not overused. (2) On the Forge, a purely negative term for GM behavior that breaks the Social Contract via the GM controlling a player-character's decisions or opportunities for decisions.
The term is used interchangeably to describe moving a scenario from Point A to Point B to Point C, which is typical of adventure path design, and it is likewise used as a descriptor of GM behavior that usurps player agency. In the RPG community at large, this has strongly negative connotations, whereas the other two terms less so. But the basic gist of railroading is that the players don't get to do much beyond what the GM deems acceptable. Because this term spans a host of behaviors and rouses tempers, I don't think it's particularly useful when discussing GMing techniques.
A term for styles where the GM has tight control over the storyline, by a variety of means, and the players do not recognize this control. Coined by Paul Elliot on the Gaming Outpost in January 2001.
The term is used to describe the "magic trick" of providing the illusion of player agency yet performing trickery to ensure that the game proceeds as he has envisioned it.
A term coined by Mike Holmes for play where the GM is fudging results behind the scenes to result in story qualities to the play. This is distinct from Illusionism, however, in that the players are aware and active partners in this process.
It's like illusionism, except the players are in on the illusion. I'm sure we've all had GMs who fudge the dice and script encounters, yet participationism relies on the acceptance and participation of the players. The GM is given the authority to control character actions and deny player agency, within reason. All GMs do this, and it's pretty normal. I'll give an example I'm sure we can all relate to: the party is talking amongst themselves about their next course of action, and they've decided to visit some location, say the Mad Hermit's Hut. They stop and stare at the GM at this point, which is the GM's cue to say, "You travel to the Mad Hermit's Hut, a residence of dubious quality. The thatched roof is half-rotten, and you're uncertain if the crumbling mortar in the stone walls could withstand a heavy rain. (Pause to see if anyone has anything to say. No characters act, so the GM continues.) Entering the structure, you see the Mad Hermit himself..."
Right there, the GM usurped player agency to narrate, yet the players have offered their implicit approval for this activity.
By these definitions, we can talk about good, better, and best gamemastering.
Tools of the Trade
The GM's arsenal is filled with tools to enable illusionist and participationist modes of play. Most of us have done them at some point or another, for good and ill alike.
The act of changing dice outcomes to achieve a preferred result. The characters defeat the climatic end-of-story boss in two rounds? The GM gives him another round or two worth of hit points to keep the tension up. The players missed an important roll by 1 or 2 points? The GM gives them the success anyway. The newbie takes a sneak attack critical hit in the first encounter and is about to be insta-gibbed? Turns out that the natural 20 was really a natural 19 all along.
Try, Try Again
The characters fail an important roll, but this roll is necessary for the story to proceed. Solution: try again until you make it! The single door behind which the MacGuffin resides is locked, and the rogue biffed it. Roll again until that lock comes loose. The clue upon which the investigation hangs is hidden due to a failed Perception or Investigate check? Test again...and again...and again until they find it.
The characters are trying something that isn't supposed to succeed, but it feels like the GM should offer them a roll, even if it's impossible. The solution is to set an arbitrarily high DC so the players feel like they have an opportunity to influence things, but gosh-darnit, they failed anyway. Alternatively, the characters are trying something that is theoretically risky but they're supposed to succeed? Give them a roll so it feels risky, but drop that DC to ground level so success is nigh-guaranteed.
The GM may also simply deny the option to roll when he feels it is vital. The players want to talk down the Evil Overlord into changing his ways and accepting the goodness and light of Pelor? No way, Jose.
Scripting outcomes, events, and plot points is commonly seen in adventure paths, for both good and ill. The NPC refuses to barter with the players. The ritual completes at midnight. The bandits are gambling in their lair and don't notice the player characters. The villain will kidnap the princess when the players return from fighting the dragon. Etc., etc., and etc.
This one's a form of probability manipulation that's not quite the same as Roll Calibration. Similar to Roll Calibration, Dice Fishing seeks to enforce particular outcome, but it uses a multitude of dice rolls to manipulate probability. The GM calls for multiple dice rolls for something, and a single roll (success or failure) shapes the entire skill check. You might have six players rolling, but if one succeeds, they all succeed (or vice-versa). In d20-based games, natural 1s and 20s are exceptionally sought after as they provide justification for the results, as many groups than play these as special successes and failures. (Other games that specifically include critical success and failure mechanics are utilized in a similar capacity.)
Players can also do this, GM permitting, when cycling through characters to succeed at a skill check. If Gandalf fails his Intelligence check to guess the password to enter Moria, Frodo gives it a roll. Related to Try, Try Again, but relies on GM's willingness to permit it.
Accepting the Validity of Illusionism and Participationism
I passionately despise illusionism; my thoughts on participationism are more nuanced, but it's not a game style for me as player or GM.
Regardless of my personal preference, some groups thrive on these styles of gameplay. The players want the GM to act as a storyteller, weaving a story in which their characters stars. They expect him to fudge dice, waive rolls, ignore mechanics, and alter outcomes as a normal part of the game. They relishthe opportunity to playact their characters in a grand narrative and be entertained for a few hours every week. Adventure paths often rely on this. They state unequivocally that A > B > C progression happens, and the players are ushered along this path. It's called an adventure path, not an adventure off-the-beaten-path. Many GMs don't have the time required to plan out detailed adventures with battlemaps and interesting foes. They need the RPG equivalent of a takeaway dish, and adventure paths provide this option. Likewise, many players are more interested in the characters and story than forging their own paths, so they're happy to share the takeaway dish with the GM. They know he's ordering it and serving it, and they're content with that. For many players and GMs, this style of gameplay serves them well, and thus I have managed to reason myself into a conclusion against which my good sense chafes: railroading is a valid playstyle.
Disgusting! Abhorrent! But I can't argue with myself.