John Derbyshire explains here the invalid reasoning in the Tim Wise link. Derbyshire's whole article is worth reading, but the takeaway is that the 2.8 in "any given black person in the United States is about 2.8 times more likely to be killed by a white person than any given white person is to be murdered by a black person" is meaningless because it needs to be compared to the number that would be expected if blacks and whites were equally homicidal. Derbyshire works out that number to be about 4.7. Since the 2.8 < 4.7, the implication is that blacks are more homicidal than whites. Tim Wise conveniently skips the comparison.