Question Your Assumptions
When you’re writing your game, there’s a tendency toward tunnel vision, to assume players will do what you expect. Take a step back and think about what assumptions you’re making:
gentleman’s agreement – You expect the players will do something based on the type of game you are playing. If you all agree to play a mystery, and you present the players with a mystery, it is reasonable to expect the players to demonstrate curiosity and investigate the mystery. If you are playing a modern spy game, it is reasonable to expect the players to make spy characters, not leprechauns. If you have a gentleman’s agreement problem, it is less likely to be an oversight in design than a disagreement or misunderstanding between you and the players about what kind of game you are playing.
the obvious choice – You presume the players are going to do the predictable thing. Often this involves a choice that is genre-appropriate (the heroes agree to help defend the village against rampaging giants, because that’s what heroes do) but this does not mean that the alternative which the players choose is necessarily out of genre. Sometimes the problem is that the GM only foresees one genre-appropriate response but the players see others.
falling for it – You expect the players to fall for some trick. In fact your plot hinges on it. They will not realize that the kindly old man who hired them is really the head of the assassins guild, they won’t make the connection between the full moon and the rash of murders, etc. Very good players will sometimes see the trick and fall for it intentionally to save the game, but you shouldn’t depend on them doing so. Note that some “falling for it” assumptions are really gentlemen’s agreements, and thus acceptable, but in those cases the players are collaborators and only the characters are falling for it. Having the players know their characters are being duped and going along with it so they can play out the righteous revenge later on is good stuff.
mental leap – You want the players to figure something out, make a connection, etc but you haven’t necessarily given them enough to go on. It is easy to make things obvious by providing a literal roadmap of clues, equally easy to make something completely inscrutable by never revealing any of the pertinent information (so only a wild guess would locate your plot) but the tricky part is finding that comfortable spot in between where the players have to put on their thinking caps but still can get the answer after some satisfyingly difficult rumination. You often don’t realize you’ve assumed a mental leap until the middle of the game when the players are sitting around trying to piece things together. For hours.