The only way to make sure that every prop in a prediction program is accurate is to test each one and define a loading constant and static thrust constant, corrected for atmosphere. Then, you have to decide if one sample is enough or if you will test more than one of each to take into account manufacturing tolerances. That is probably more of an issue with wood props. This would be thousands of propellers and thousands of hours. I've considered doing it for props I use on a regular basis, but have not found the time to do so yet.