I recall hearing a presentation by the State University Soil Fertility specialist, back around the time when Precision Ag was just getting off the ground.  The initial expectation was that the areas in a field that was lowest in nutrients would be the poorest yielding areas in a field. That was not the case. In fact, the best areas in a field was many times the areas with lower nutrient levels.  They concluded that the highest yielding areas pulled out more nutrients (P & K, etc), and that was why those areas tested lower.  He went on to list all the non-fertility reasons that affected yield:  Soil OM, water management, depth to bedrock (or depth to water table), etc, etc, etc.  

It was kind of enlightening presentation at that time.  Made one think about things we don't always think about.