Honestly, I think the Alexandrian's article on the 'Railroading Manifesto' and subsequent articles should cover everything about railroading vs linear design, how to avoid railroading and why it's best to do so, etc.
It's worth a full read, but there's a lot of juicy quotes:
The rest also has some fantastic advice.
Note: I haven't actually ran anything yet, but I keep these articles in mind when I do design stuff and when watching or playing in campaigns.
It's worth a full read, but there's a lot of juicy quotes:
Linear scenarios are built around a predetermined sequence of events and/or outcomes.
Consider a simple mystery:
Scene 1: The PCs come home and discover that their house has been broken into and an arcane relic stolen from their safe. They need to figure out who did it, which they can do by analyzing fingerprints, looking at their neighbor’s surveillance camera, asking questions around town to see who took the job, or casting a divination spell.
Scene 2: Having discovered that Jimmy “Fast-Fingers” Hall was responsible for the break-in, the PCs track him down. They need to figure out who hired him, which they can do by interrogating him, following him, analyzing his bank statements to figure out who paid him, or hacking his e-mail.
Scene 3: Having discovered that Bobby Churchill, a local mob boss, was the guy who hired Jimmy, the PCs need to get their relic back. They can do that by beating Bobby up, agreeing to do a job for him, or staging a covert heist to get it out of his vault.
That’s a fairly linear scenario: House to Jimmy to Bobby. But because we used the Three Clue Rule to provide a multitude of paths from one event to the next, it’s very unlikely that a GM running this scenario will need to railroad his players. The sequence of events is predetermined, but the outcome of each scene is not.
Non-linear scenarios do not require specific outcomes or events, allowing freedom of player choice.
Linear scenario design and non-linear scenario design exist on a spectrum. Generally speaking, requiring specific events (“you meet an ogre in the woods”) is less restrictive than requiring specific outcomes (“you meet an ogre in the woods and you have to fight him”). And the more specific the outcome required, the more likely it is that the GM will have to railroad the players to make it happen (“you meet an ogre in the woods and you have to fight him and the killing blow has to be delivered by the Rose Spear of Vallundria so that the ogre’s ghost can come back and serve the PC at the Black Gates of Goblin Doom”).
With that being said, it’s often quite trivial for an experienced GM to safely assume that a specific event or outcome is going to happen. For example, if a typical group of heroic PCs are riding along a road and they see a young boy being chased by goblins it’s probably a pretty safe bet that they’ll take action to rescue the boy. The more likely a particular outcome is, the more secure you are in simply assuming that it will happen. That doesn’t mean your scenario is railroaded, it just means you’re engaging in smart prep.
My point here is that you can’t let fear of a potential railroad make you throw away your common sense when it comes to prioritizing your prep. This, by the way, leads to one of the most potent tools in the GM’s arsenal:
What are you planning to do next session?
It’s a simple question, but the answer obviously gives you certainty. It lets you focus your prep with extreme accuracy because you can make very specific predictions about what your players are going to do and those predictions will also be incredibly likely to happen.
The rest also has some fantastic advice.
Note: I haven't actually ran anything yet, but I keep these articles in mind when I do design stuff and when watching or playing in campaigns.