This morning I came across a post by David Varadi on the (futile?) quest for simplicity. He writes:
The most optimistic quantitative researcher knows deep down that an unexplained noise dominates the data that mysteriously eludes linear models such as regression.
Markets are chaotic systems characterized by feedback. If they were neat and orderly (and predictable), our collective efforts to exploit those opportunities would heighten feedback and propel the system back to chaos.
But chaos can be described. The separation of signal from noise is certainly not easy. In fact it may be so challenging that we can only rely on rough first or second-order approximations. But to quote George Box (again), "Essentially, all models are wrong, but some are useful."