At the time of it's release, the Storyteller system had some great ideas. Some really pushed the potential of roleplaying, some were just codified instructions for what a lot of us did around the table anyway, and some opened the door for later innovations.
The idea of multiple degrees of success was always an interesting one; coupled with the concept of the "botch", it gave a fun progression of linear outcomes to a task resolution. When I was developing my Vector Theory of roleplaying back in 2010, I said that the story parts of an RPG session were generally straight lines (vectors), while the game elements were "node" points where the narrative might change speed or direction. In a traditional railroaded game, these nodes might act as traffic lights, stopping and starting the flow of the narrative depending on whether the characters failed or succeeded. In some of the more interesting traditional games, there might be switch points along the railroad, where players t get the opportunity to take different paths to the same end goal. Games like D&D, Pathfinder, or pretty much anything OSR generally work in this way. A scenario has a bunch of gates, and characters are stopped by those gates if they fail a skill test, if they work out a way to come at the gate from a different angle with new skills or new information, then they might be able to challenge the gate again, but it is distinctly possible to completely fail the sceneario because every gate has been "sealed shut" by poor rolls.
The Storyteller system doesn't just have a binary pass/fail outcome in it's rolls. Instead it may deflect the characters in an unexpected direction (with a botch), it may impede them in a traditional sense (as described above), it may allow them to proceed with a low velocity (with a low level success), or it may shoot them forward with high velocity (through multiple degrees of success). The system really doesn't address the direction of the story with it's mechanisms, this is all handled through the metagame of description and roleplay.
Mage goes a little further down the narrative path compared to other Storyteller games, and certain more than most other mainstream RPGs of its era. The nature of coincidental magick invites players to engage in the description of the world, creating elements of the narrative that can be manipulated, either by their coincidental presence, or a manipulation of magick to invoke them into existence. But the task based nature of the system and the wider metagame context of the game still meant that such manipulations of the narrative were arbitrarily determined by the GM/Storyteller for their effectiveness. Similarly, the nature of a story's redirection due to a botch is also determined by the GM/Storyteller, sotoo the manifestations of paradox in the game. Deliberately or otherwise, the "person in charge" is designated the Storyteller, and everyone else are merely players portraying characters in that person's "Story".
It's almost there, but not quite.
I think that one of the more interesting developments of the Story Games movement has been the idea of allowing the player to determine how something succeeds. In one version of this, an action might have a few possible outcomes and the player gets to choose which of these outcomes are prioritised; in another version, a player with a near success might choose to accept a minor sacrifice that could deviate the story or have future ramifications in exchange for the success now. Putting more choice back into the hands of the players has a tendency to increase their investment in the story. That's the one thing I'd really want added into the core system of Mage to bring it up to date with modern gaming. Rather than simply making the node points where stories change speed and direction purely determined by a combination or rules, randomness and Storyteller, it would give the players more of an option to determine where they want to go.