This question has no clear answer in its current form. The strip method is a lower bound approach (meaning that you approach the optimum solution from the safe side) based on beam analogy. So in theory (given endless ductility, etc.) you could arrange your strips in any way you want and you would always end up with a safe (but possibly horrible in all other aspects) solution. The recommendations given for the application of the strip method are to keep redistribution needs and service state-behavior in check as well as finding economical solutions. There's a lot of different recommendations in the strip method that vary between cases, corner angles for load dividers, moment ratios, widths of column strips vs mid-strips, widths of edge strips, reinforcement distributions within the strips, etc.
If you want to learn more about the strip method, and if you're going to apply it you really should, Hillerborg's book "Strip Method Design Handbook" is a good place to start. I personally find both the strip method and the strut and tie method to be very simple and elegant manifestations of applied engineering judgment.