How do you assign probabilities to expected lengths of inferential differencies needed to technically understand clearly 'cascading' issues? I mean, the grass is green. I know it has to do with light being selectively reflected off it, because of photosynthesis, because of the plant needing energy to live. I estimate the ID in the range of c. 5 steps (the level at which I'd feel comfortable in predicting outcomes of adding fertilizers to vegetable garden) to 50 (answering an exam question). Yet, I happen to need knowledge of photosynthesis now and again, a semi-technical one. And I usually feel, well, like neither the '5' nor the '50' resolution fits.

I know about 'emergent properties' and such, but - can you share your own eхperiences? Maybe there's some explainable variation between fields.