I also came across a similar idea while studying industrial design at university: there's some collection of established design problems and potential solutions. Always wondered if this could work for games.
I've seen various design patterns collections for game design, although now that I think of it they're not really doing what A Pattern Language or the Gang of Four Design Patterns book did. They were more lists of game mechanics, not really solutions to specific problems. So yeah, finding the problems ain't so bad a way to start.
Nidre, may I suggest a methodology that might be use to you? If you assemble of corpus of discussion about games (whether that's reviews, forums, etc.), you could make a tag cloud or something that lets you see the most frequent important words, and pick out a variety of negative terms ("bored", "frustrating", "quit", "unfair", etc.) Those can be turned into questions:
"Describe a recent episode in which you felt bored while playing a game. What made you feel like that?"
"What kind of game was it?"
"Is this a common problem in games of this type, or is it unique to this game?"
You could make a gamer version (the above) and also a dev version ("Have you ever had a complaint that an aspect of your game was boring? What aspect was it?").
(Also, if you have such a game discussion corpus, you can do things like run correlations on genre terms and negative words, to see what people are negative about when they talk about different genres.)