This "neat trick" means that if the adventurers wanted to go to Chicago, or Denver, or Kansas City, and end up in New York no matter what, then the players' decisions were ignored by the referee.
That's pretty much the exact opposite of the atmosphere I'm looking for, as a player and especially as a referee.
the problem I see is what is really meant by the players choosing Chicago over New York.
I agree with Janx here - and that doesn't necessarily mean I'm radically disagreeing with The Shaman - but I want to take the notion of "meaning" in another direction.
Janx gives examples where the choice of city isn't meaningful, because it isn't driven by any sort of concern that the GM is undermining by changing the location of an encounter (eg the pickpocket, or the cloaked stranger).
But another sort of meaning is this: maybe the players are choosing Chicago over New York because they want - for whatever reason is important to them or to their PCs - to learn the truth, in the gameworld, about Chicago. If this is their goal, then that goal is not undermined by the GM making up stuff about Chicago on the fly, or moving stuff to Chicago that, in his/her head, the GM had assumed would happen in New York. Because the players are still learning the gameworld truth about Chicago.
I've been thinking about this for the past week or two leading up to and subsequent to running a 4e, quasi-No Myth
exploration scenario. I might start a new thread on this, but in short, I discovered that you can run a scenario in which exploration is the priority - in that the players' main goal is to find out what happened/what is going on in a particular location - without having pre-prepared the answers to their questions. You can make it up as you go along. It still serves the goal of exploration, because the players are still learning the truth about the gameworld.
Now
why the players would care about the truth in the gameworld - be it the truth about Chicago, or the truth about the time-travel scenario I ran - is another question. Maybe one of the players wants her PC to become mayor of Chicago because she herself was born there. Maybe the players have a magic item which they can only use properly if they understand the truth of the gameworld's history. There are any number of reasons why the players might care about this sort of thing, which don't depend upon there being some pre-determined truth that doesn't change, in response to their own choices and exploratory priorities as actually revealed in the course of play, from what was notionally written by the GM in the campaign handbook.
I consider the real problem being the GM not taking "no thanks" as an answer. When the players decline to follow-up on the map, then don't make the sherrif force the party to follow the map to the same dungeon. Move on.
I agree with this, but think that rules in many traditional(-ish) RPGs like D&D could benefit from discussing it a bit more.
For example, they often contain GM advice along the lines of "the players may enjoy having their PCs' backstories incorporated into play from time to time" or "follow up on what seemed to interest the players, by including a connection to it in the next adventure". And I see posts on ENworld from time to time that talk about using PC backgrounds as a basis for sidequests.
But they don't tend to take the next step, which (in my experience) helps reduce the whole need to worry about "no thanks" of, saying "Combine PC backstories with those elements of prior play that interested the players and
make that into the next adventure". Once these things are not peripheral but
the game, then you're less likely to get players saying "no thanks".
Even under this approach, there can still be local issues of sequencing eg if there are two different things the players want to do - let's say, visit the Feywild and also redeem some slaves in a city on the world - then it is likely that the players and not the GM will determine the sequence, meaning you can have issues of unused prep or alternatively needing to improvise. But from my experience, most of your prep isn't wasted because most of it is prep for
excactly the adventures the players want.
(And by way of acknowledgement: a lot of people think the Forge is a waste of time, but this approach to adventure design is something that really crystallised for me after reading an essay by Ron Edwards in which he talked about turning the usual conception of the plot hook on its head: instead of the GM offering a plot hook to the players, the players - by building and playing their PCs - establish plot hooks for the GM. Reading that and reflecting on it, and on actual play examples as well as rules from more non-traditional games, really helped me develop my GMing from the previously embryonic stage it had been in, where I was trying to break out of plot-hook driven simulationism but didn't quite know how.)