I have several thoughts right off the top of my head:
1. This can become complicated in alloys where transformations take place (martensite formation, overtempering of previously Q&T steels, etc.).
2. If all you need is quantification, then measuring microhardness on a cross-sectioned piece will be easier than trying to predict based on chemistry.
3. Predicting based on chemistry will require knowledge of the microstructure as well, since this is have as much of an effect if not more.