In the vein of offering some practical advice that I myself am having some difficulty accepting...
Good writers read. They read a lot. When they're standing in line at the grocery store, they read. While they're waiting for their car to fill up with gas, they read. Great chefs eat. They sample everything, no matter how crazy. Everyone isn't a prodigy. We don't all take to this medium like Maga, Zarf, Emily, Graham, Mike, et cetera ad nauseum. While they all have different "strengths", I'm going to assume that they all have one thing in common.
They play a lot of IF.
Play a lot of IF. Note those things that you tried and didn't work. Note those things that you tried and didn't work like you wanted. Read transcripts. Club Floyd is both educational and hilarious. Figure out those things that you can write a universal reaction to and respond to those. Figure out those things that require special attention and respond to those. Hopefully you get the idea.
You can't just let the player run wild through your world. You can't possible anticipate or program everything they might do. You have to guide them gently and as invisibly as possible. Take the following example.
You're standing on the front lawn. An autumn sun lends some warmth to what would be a chilly breeze as it brushes gently across your face, disturbing the overgrown grass. The flowers that you planted last year for your mother have bloomed beautifully, seemingly in reverence to her passing. The flower beds are ringed in large pieces of sand stone and filled with pea gravel. The front door of the house is north, offering the bittersweet recollection of a time you haven't reflected on in months.
Not a great example. But what can we reasonably expect the player to do here?
- Examine flowers
- Examine sand stone
- Examine gravel
- Get gravel/stone
And what are they probably not going to care less about?
- Examine grass
- Examine sun
- Sniff breeze
- Get grass
- Scratch arm
- Check watch
Your problem isn't what your players might possibly do or try. Sometimes they're going to get a nonsensical response from the interpreter. Your problem is writing prose that makes sense, that leads the character in the direction that you want them to go while at the same time accomodating the obvious things that the player might try to do. There will always be people out there who try to do stupid things, (hopefully) knowing that they won't work.
I'm going to go ahead and promote Aaron's book again here. He discusses, encourages and provides examples of what to emphasize and when. The way you write is going to dictate a lot of what the player does. Not their whims.
The truth of the matter is that people want to play your game. They want to get through the story. They want to see what all you have to offer them in the way of a playing experience. If they try to do something that makes sense, something that it seems viable they should be able to do and can't do it, they may complain. Otherwise, you can't stress it.
You see a brick wall.
I'm not going to examine the bricks. I'm not going to count them. I'm not going to examine the mortar to make sure they used appropriately sized spacers for the size of the bricks and the intent of the application when they were doing the masonry. I'm going to check my inventory for a rope or some dynamite. I might poke the wall a couple of times to make sure there isn't a secret passage. Then I'm gonna find something better to do.
So again... How you present it dictates what the player will do far more than your anticipating their every whim. Play some games. Run some beta tests. Find out what players are likely to try and what they aren't.
Go from there.
[Edited to get a cooking reference in there.]