The overall DxOMark score is obviously useless where medium format is concerned - by including low-light capability, they are testing something that the MF makers never CLAIM to provide. It would be like a car magazine testing the off-road capability of a Ferrari, or an audio magazine testing the home theater credentials of a two-channel tube amplifier, then including that in a bogus overall ranking. Even the individual numbers surprise me, and I think there's something wrong. I bought my D3x after careful evaluation against medium format, and even as a very happy D3x shooter, I don't believe that it has MORE dynamic range than a Hasselblad - prints just don't bear that out!
DxO claims that the D3x has about 1.5-2 stops more DR than most older DSLRs like the Canon 1DsII - fine, I believe that - I'm putting ink on paper that shows a whole bunch of extra shadow detail, and holding some extra highlights as well. I haven't shot another DSLR that does anywhere near as well. I don't believe DxO's absolute DR numbers on ANY DSLR, but their relative numbers seem pretty good (they are always optimistic by about two stops, but their 13 stop D3x has ~11, while their ~11 stop 1DsmkII really has 9).
Hasselblad claims about 12 stops of real DR, and I can see that in prints they showed me - the prints look like they have slightly broader DR than even a D3x, and much broader than older DSLRs. I don't have real-world experience, only tests in controlled situations, but I'm inclined to take Hasselblad's (and the other MF makers) claims pretty much at face value.
DxO comes in claiming that the Hasselblad has half a stop LESS DR than the D3x... That doesn't seem right (12+ stops for the Hasselblad seems fine, but when you subtract the 2 stops that DxO routinely overestimates by, 10+ stops doesn't seem plausible).
The only thing I can think of is whether MF might have much lower shadow noise than anything else? In that case, the absolute DR (until black clipping) that DxO measures might be much closer to real-world photographic DR than from other cameras - the 2 stop "DxO correction factor" might not be needed...
-Dan