I agree that's not a railroad. It also relies on fairly hard scene-framing - the players aren't really free to have their PCs just wander off and do stuff, as there will be no game unless they plunge headlong into White Plume Mountain (or whatever). What's distinctive about that sort of game, I think, compared to what at least some of us on this thread seem to have had in mind is that plot and story aren't really meaningful issues at all, from the point of view of the players of that game. It's all about the crunch!In a one-off step on up game, you can run a short, challenging adventure (with shallow PC build tools and no PC backstory to go on) where folks just show up with Bobfighter001 etc and just tackle the challenges. There you can have fully legitimized player choice and action resolution that has meaning, affects micro-outcomes and perturbs the macro-outcome wildly (such that all roads don't lead to Rome). I'm not sure I'd call that a railroad.
So in the sort of game you described, fudging the dice would probably be the number one sin on the GM's side of thing.
I'm definitely focused on players. If the characters are free, but everything they can deal with is established solely by the GM, I see that as pretty railroady.I had an epiphany (I hope).
Players or Characters!
Players modify my campaign world all the time. Things change based on input, "geographies" get modified etc, they choose the plot ahead of time. So its a very player driven game.
Characters can only affect stuff within the framework. They deal with the plot. (and their actions have consequence and benefit)
Conversely, if the GM is scene framing fairly hard, but the content of the scenes reflects suggestions/requests/general vibe generated by the whole table, I don't count that as railroading even if the PCs have little or no choice (eg they wake up in a prison cell and have to try and bust their way out).
4e removes a lot of PC capability that earlier versions of D&D had that undermine GM scene-framing authority (eg teleport as the classic one). But for me that is a change that is neutral on the "railroad-ometer", because from knowing only what the PCs can or can't do I haven't learned much about what influence, if any, the players are exercising.
I change my background framework as I go along to respond to player signals. So the players can influence the background and the focus of play other than via the agency of their PCs.Once they have given their inputs, I dont railroad them along a story, but the framework of who, what, why, etc stays the same.
But once a bit of fiction has been established ingame, then it's fixed.
One issue I'm interested in is mechanics that permit the players, via their PCs, to impose finality other than via combat. This is a big part of a game like Burning Wheel (with its Duels of Wits) or Marvel Heroic RP (with its mental and emotional stress tracks, and its rules for complications), but not really a big part of traditional D&D. I run skill challenges as establishing finality - once the challenge is resolved, if the PCs won then the players have imposed their will on the fiction and I, as GM, am bound by that. So, for instance, if the skill challenge resulted in an NPC giving their word, then I have to stick to that. I can't just change my mind and have the NPC go back on its word because I think that would be fun.
On at least one occasion, too, my players have appealed to this principle of finality to help interpret a particular scene - the PCs got in a fight with an enemy in public, after they had goaded him into attacking them (via a successful skill challenge). When the issue came up of how the onlookers responded, one of the players reminded me that part of what the players had achieved via their successful skill challenge was to make it clear that the NPC was the bad guy. This had become part of the established fiction, so I wasn't then able to disregard it and narrate onlooker NPCs who thought the PCs were the bad guys.