Why doesn't everything effect everything?

How is it that there are phenomena which are independant of all but a very small set of conditions?

Not sure where i read this, which is funny because i read it yesterday. OK found it: Wigner: The Unreasonable Effectiveness of Mathematics in the Natural Sciences.

It turns the whole holism/ecology/complexity thing around nicely. The point is usually 'Hey guys, listen, your models are very pretty, but they only account for ten percent of the activity we see, maybe they should be more complicated'. Computers allow for more complicated models that take more into account. Pollsters and ecologists are constantly facing strange unexpected confounds on what 'should have been' straight forward results. And now we look in from the other end of the telescope and get confused: How could it be that somethings aren't affected by everything? How is it that a simple equation can explain a pendulum with maybe 3 significant digits and accounting for friction can give you another two? Why isn't the movement of a pendulum sensitive to social whims? it should be obvious, but the phrasing of that above quote has me really confused.

Some things are explained 99% by one 'cause' .9% by the second .09 by another and on down. With three parameters you can account for 99.99% of behavior. Other systems are influenced 1% by one, 1% by another, and on, so that with 99 parameters you wouldn't have the same predictive power. So what kinds of systems are more prone to be affected by 'everything'? What kind of systems are predisposed to having pretty much just one 'cause'. it sounds like the philosophy of causation gets involved quick, which is too bad because I've tried reading that shit before and its damn dry. I wonder if scale matters...