“We estimate 22% savings” is a claim. “22% savings via a model calibrated to within 8% of billed consumption, inputs in Appendix B” is evidence — and the difference between those two sentences is this document. Five recognized methods with when each is appropriate and what must be disclosed for it, calibration expectations for simulation stated as ASHRAE Guideline 14 conventions, and a per-measure disclosure form. Two traps get named directly: summed independent estimates presented as a package total with no note, which is the most common overstatement in the field, and unstated escalation assumptions, which turn a defensible analysis into a disputed one the moment a reviewer rebuilds your figure at flat rates.
