The 1 or 2 HRc point spread in hardness is not that significant, in fact it can be within the margin of error in hardness testing.
How many hardness traverses did you perform for actual comparisons? Also, the surface will still cool more rapidly in comparison to the core for thru-hardening steels. If what you report is correct, it could simply be that the martensite adjacent to the surface contained less retained austenite, in comparison to the martensite formed toward the core.