Originally Posted by
Frozen_Feet
I, on the other hand, run games at conventions, where I rarely know the players beforehand and typically do not have enough information to actually do the sort of algorithmic response we're talking about. (Even when I use premades, they are not designed to sufficient depth that I could reliably predict how unknown players will play them.)
So I have taken to crafting my games in a vacuum. The thing is that while I've been doing this, I've noticed it's significantly improved my game holding skills. Why? Because by minimizing assumptions of player preference I'm forced to consider wider range of possible actions and design elements.
Tailoring a game to your audience isn't a bad idea, but a lot of GMs plainly do it wrong. That is, they have laser-like focus on their one group and get increasingly trapped in that one box. To the point that they can't deal with any other group or sort of players.
Hence, I hold that the ability to design and evaluate things "in a vacuum" is not just usefull, it is vital for long-term success as a GM and for long-term health of the hobby.