I’ve been GMing/DMing/facilitating roleplaying games for most of my life, but I only recently began to feel like I was getting to be good at it. The secret ingredient that I’d been missing all those years? A willingness to surrender control of the story to the players.
I’m a big fan of the Karma/Drama/Fortune (K/D/F) model for understanding how the efforts of roleplayers direct the resolution of events. My relationship with the K/D/F model is a story for another blog post, but I’ll be using it as as a framework here in the meantime.
I could write forever about the way I interpret the K/D/F model, but for now we’ll suffice to have a quick primer on terminology. When talking about resolution (how it’s decided what happens next in the story), the K/D/F model describes the relationship between three forces: Karma (player choices), Drama (story needs) and Fortune (luck, e.g. dice rolls).
When the lich king comes to theregion to provide a villainous plot hook, that’s Drama. When his lucky critical hit kills an ally of the adventurers, that’s Fortune. When one of the heroes comes up with a clever way to misuse a magical artefacts from a previous quest to defeat him, that’s Karma. And all along their quest they probably saw many other resolutions lean more-heavily on one or another of the three parts of the model.
Some days, a critical hit is just enough. Other days, you should’ve just stayed at the inn and got drunk.
For most of those many years of gamemastering, I saw my role as providing the “drama” part of the K/D/F model, and that approach creates an inevitable trend towards railroading: “forcing” players down a particular path. If a gamemaster has an inflexible and excessively concrete idea of the direction that a story must go, then whether or not they mean to they become unable to see the narrative through any other lens. The players are deprotagonised and the adventure just becomes a series of set pieces, connected by the gamemaster’s idea of how things should play out.
What if Bilbo and his party escaped from the wood elves by land, heading directly East to Erebor instead of via Esgaroth? What if he failed to determine Smaug’s weakness, or chose not to steal from him? What if the dwarves successfully fought off the goblins and didn’t need rescue? The difference between an adventure story and an adventure game should be that in a game, nobody – not even the author – can be certain ahead of time of the answers to all the questions.
I’ve railroaded players to some degree or another on many ocassions. For example, and in the spirit of learning from my mistakes:

Quantum Ogre
Scenario: In a short-lived high fantasy GURPS campaign, I wanted the party to inevitably meet a band of gypsies and for at least one of them to have their fortune told, in order to foreshadow other parts of the story yet to come.
What I did: I pulled a quantum ogre (magician’s choice) on them: whether they travelled by road, or water, or hacked their way through the forest, they were always going to meet the gypsies: their choice of route didn’t really matter.
Why that was wrong: I’d elevated the value of the encounter higher than the importance of player agency.
What I could have done instead: Reassessed the importance of the cncounter. Found other ways to foreshadow the plot that didn’t undermine player choices. Been more-flexible about my set pieces.

Fudging
Scenario: In a Spirit of the Century one-shot the plot as-written demanded that the antagonist successfully kidnap a NPC from aboard an oceanbound ship but to my surprise – with some very lucky rolls – the players foiled the plot!
What I did: I used a fudge – an exploit based on the fact that in most games the gamemaster controls both the plot and the hidden variables of the game mechanics – to facilitate the antagonist kidnapping a different NPC, and adapted the story to this new reality.
Why that was wrong: It made the players feel like their choices didn’t matter.
What I could have done instead: Used the failed attack as a precursor to a later renewed offensive by a villain who’s now got a personal interest in seeing the party fail. Moved towards a different story.

Ex Machina
Scenario: In a long-running Warhammer Fantasy Roleplay (1st edition!) campaign, a series of ill-advised choices and some terrible luck left the party trapped and unable to survive the onslaught of a literal army of bloodthirsty orcs.
What I did: I whipped out a deus spiritus ex machina, having a ghost the party had befriended turn up and basically solve for them a useful puzzle they’d been struggling with, allowing them to escape alive (albeit with the quest truly failed).
Why that was wrong: It deprotagonised the adventurers, making them unimportant in their own stories.
What I could have done instead: TPK (sometimes it’s the right thing to allow it to happen)! Pivot the plot to facilitate their capture (e.g. the arch-nemesis can’t solve the puzzle either and wants to coerce them into helping).

…
Let all gamemasters strive to be as prepared as Abed.
[diagram showing “stages” of a quest typical to a published sourcebook]
…
In my defence, many professionally published roleplaying adventures are written as a series of scenes connected by the assumption that the author knows exactly how the players will proceed from each, and don’t teach gamemasters how to handle any deviation.
…
…
…