I want new and innovative mechanics because mechanics influence and direct the flow of the game. If they weren't important, we wouldn't have multiple systems.
Nobody said they weren't important. But "important" does not equal "must be new". If the desired flow of the game can be achieved with fairly well-known techniques, why reinvent the wheel?
I also note - you can get new flows of the game with old mechanics, combined in new ways. For example, take out alignment, and bolt a WoD-style morality system into the game instead. With different results for many stock D&D behaviors, the game would play much differently, though there's nothing really new there.
I also want the actual gameplay to bring entertainment over and beyond the narrative. I want it to be fun as a game in and of itself.
And for those who don't actually want to have to deal with game rules? What happens to them? Go a long way to bake the "fun and entertaining game" into the system, and you'll find many of those who are seeking immersion, which is broken by having to worry so much about the details of rules, get turned off. So, there's a choice there - go for one end, the other, or some compromise middle-ground?
And again, in order to be a fun game in and of itself, they must be new an innovative mechanics?
But over and above, I want innovation because it has to do something new and better than the games I already have. Otherwise I don't see why I would want it.
Well, it potentially can do things better than before, with modification and rearrangement of well-known mechanics. Sometimes, good engineering comes from retuning designs that are already proven in practice, rather than starting from the drawing board.
Not that I mind new and innovative either, mind you. I honestly don't care that much. But I think "it must be new" as a barrier to entry can be self-defeating. When the system is done, look at it, and see what it does. Don't worry about what bits are new and what aren't - see what it accomplishes, new or old.