In order to compare the free energy landscape for a protein under two conformations- substrate vs inhibitor bound, will g_sham give me the required FE landscape?
What g_sham does is a trivial calculation based on histograms. The real key to a useful result is in identifying two variables along which a suitable free energy surface can be created.
Free energy calculations are called "the holy Grail" of computational biophysics. It's challenging task and Justin is right - well described questions require more than just a standard solutions.
Histogram approach may give useful informations but everything depends on several questions and - at least in this context - good enough space sampling. IMHO Justin's opinion is the best way to consider free energy changes in most cases, cos we should narrow our research to somehow well described problems.