1) System development and selecting the results.
"The "top of the hill" should be broad - this tells you that you are not picking just one lucky combo of stop/target or other parameters you are testing for."
How often do you update the analysis?
How are you choosing your path?
Are you...