Hmm. I'd expect that to translate to less than 6.7% black people among registered voters... in which case, CNN getting 10% black voters implies either some sort of sampling bias, or a very high turnout rate for blacks compared to whites (which might be plausible, this time around).