"Theoretically?""I punch the big guy standing at the end of the bar." "And I toss my drink in his bigger friend's face."
Mission accomplished. By player agency.
Considering the PCs have a limited perspective compared to the GM, the theoretically is a realistic caveat. Suppose the PC who threw the punch and tossed the drink picked a couple of guys who won't rise to the bait? Maybe they belong to a sect of pacifists. Maybe they're deliberately keeping a low profile because they can't afford to draw official attention. Mission failed.
Theoretically doesn't deny player agency. It's just a recognition that not everything they try may succeed in the stated goal - in this case of starting a bar fight.
In addition to The Shaman's reply that
the fight is already on, I wanted to add this thought: it is possible to have a game in which the players states the goal for his/her PC's action, the GM sets the difficulty, and then the dice are rolled - and if they come up the player's way, the player gets what s/he wants.
Burning Wheel is this sort of game. In my view, 4e is best played in this style also (it's what skill challenges are for). And in that sort of game, not only can the players have their PCs start throwing punches, they can make an Intimidate check, or a Streetwise check, or whatever the appropriate mechanic is,
to make the NPCs fight back.
For an actual play example of 4e played in this way (rather than a bar fight, it was provoking a wizard NPC to attack them during a formal dinner), see
my post here.
In running this session, I cheated, faked, hand-waved, made up stuff, and even re-mapped the dungeon on the fly.
<snip>
I've had many game sessions where I stuck to my core beliefs of a status quo style -- what I feel as a DM and a Player, through many years of gaming, is the best style -- that just completely bombed. But as soon as I do one game session where I break all my personal style rules, I get a great session.
You were telling a story, not playing a game.
Not really all that complicated.
And just to prove that I don't agree with The Shaman on everything - unless I missed something, you ([MENTION=31216]Bullgrit[/MENTION]) didn't cheat - as in, you didn't actually disregard any dice actually rolled as part of action resolution. It just sounds like you framed every situation spontaneously and in response to what had come before, rather than preparing it in advance. And that you made liberal use of "say yes" - as in, only got the action resolution mechanics into play when there was an actual conflict to be resolved.
This isn't "telling a story and not playing a game". This is just GMing "indie style" rather than classical style. (What you describe sounds like it was pretty close to
No Myth.)
EDIT:
Illusionism comes in if the GM has already decided the outcome and ensures that all roads lead there. This negates player choice. From what Bullgrit says, he didn't do that.
Fudging comes in if the GM appears to be using the rules, but ignores the results of dice rolls - such as changing misses to hits, and secretly changes already-encountered reality on the fly, such as arbitrarily lowering or raising hp totals of monsters. That's the stuff that players will object to if they find out (it can be a 'no Santa Claus' moment when they discover their trusted GM is a fudging fudger), and why Bullgrit rightly feels 'dirty' despite a good game overall.
I picked up the absence of illusionism. I missed the hit point fudging, but am not 100% sure whether it's objectionable or not. Bullgrit could easily have decided instead that the monsters surrender, and this would produce an outcome similar to the hit point fudging. But not identical - the players would now have the surrendered foes to negotiate with - and maybe that's enough to show that the hit point fudging
is objectionable.