So you start out with large rocks.
Then they become round (so they can roll more often).
Then you prop it up in the air with something, which means the previous phase of your "evolving mousetrap" wasn't really necessary, because the rock is standing propped up in the air all day waiting for the mice to come past.
Oops, the mouse detection system doesn't come part & parcel with thing propping it in the air, so you have a bunch of round rocks standing around propped up in the air and doing nothing.
At least the ones rolling around still kill mice occationally, so they get to live happily ever after.
You end up with nothing close to a mousetrap and nothing that is really effective either.
Adding the prop without the detector was actually detremental to the success of the design, although it was a step in the right direction. It's the same with the IC systems that Behe talks about. If you only have half the components assembled, the system will sit there and be very complicated, but utterly useless. I would like to hear an explanation on how a FLAGELLUM got put together in an evolutionary way. If such an explanation can't be given by an intelligent human being who knows what the outcome should be, how can it be that random and unguided events could "invent" such an efficient motor?