A couple of things occurred to me the other day as I was reflecting on my last game session: 1) When players ask lots of questions, the the odds of doing something "cool" decrease. 2) If something sounds cool, it is more likely to work in my games.
I don't really think about these things during the game. They are not rules that I created... but this is just how things work out in play with my games.
Here is an example. In the last session, the party arrived at the cliff that separates the road from the keep. The drawbridge was damaged and only one of the two chains were holding it up. Someone asked if they could see the wench that held the remaining chain, and I replied that it was on top of the keep's wall between the two towers on either side of the bridge/gate. Then someone asked if the wizard flew up there would he be able to release it, and would the bridge land properly across the gap. I said he could probably release it, but that because the bridge was askew and damaged, it might break and fall. They ended up going with another plan.
But, had someone said, "I want to blast the wench with a fireball so the bridge falls and we can cross", I'd probably allow that to work (assuming they made the roll to hit).
Thinking about this, I realized that when I GM, lots of "technical" questions get the simulations part of my brain working. So, the answers I give are likely to be based on what makes the most logical sense for that situation. But, if someone states some crazy plan, the more creative side of my brain gets going and I want that to happen.
Of course the players need information before they can formulate crazy plans, so there's a bit of a catch-22 going on. But I think recognizing the ways that I GM and what some of the factors that lead to it helps me and my players.