The reported hashing power only matters if the same algorithm is being used. Clearly SC2 either doesn't use the same one, or (more likely) has creative reporting of the hash power.
I tested SC2 and Art's CPUminer against each other at the same clock speed:
CPUMiner Scrypt: 16-17kh/s
So divide the reported SC2 hash by three to get something comparable to the various scrypt coins.
That still leaves rather more power there then elsewhere, maybe CH/RS rented a botnet for a while so he didn't have to pump'n'dump SC2 like he did SC1?
I wonder if hash rate estimates are adjusted for every even block being difficulty 1, or if it simply the same networkhashps getinfo calculation (which assumes constant difficulty blocks) as the client it's derived from? If the later then the actual mining power can be divided by 2 again. Still, that would still yield double the power of litecoin and tbx combined. I'm still not buying it.
It would be cheaper to simply dump money and paint the charts on btce than rent even a small botnet for days, IMO.