Regarding your last post, you just need to let go of conventions and try to come up with radical improvements on existing games. I mean really, really go wild.

Here's an axiom for you: you can just take any action-oriented game, and create an extremely innovative experience just by twisting the combat (conflict resolution) mechanics around a lot. Messing with other mechanics usually results in nice eye candy, but nothing more.

Often this means detailing a part of the game mechanics that previously was approximated by a "crude" system, like a random decision, rock-paper-scissors systems, linear progress bars, etc. Detail it, add new elements, remove others. Don't worry about unbalance or what "should be right".

For example, take a run-of-the-mill medieval RPG, make weapon reach important (sword hits 10 times before dagger gets close enough to hit), and a shield only gives directional protection (fighters are unprotected if attacked from behind). Now you've got a host of strategic elements that just don't exist in regular RPGs. I'd say that game would be innovative. The emphasis here is in the strategy that arises from positions / formations, and to have that I needed to twist a few mechanics that are usually modeled as entirely linear (and predictable) systems.

[edit] Oops, I got carried away and forgot to congratulate you on starting a new project! I hope this can be of some use in designing the mechanics for this new game ;)

Thanks a lot for the feedback!

Ill definately keep that in mind and you made a great point.
Thats exactly what I plan on doing, taking familar concepts and twisting them.

If you have any more advice/comments, let me know!
Thanks :)