Ah, I have to go through this again. Shame that Jim Goad made the same mistake, but it's an easy one to make.

Suppose blacks & whites were perfectly equally likely to commit rape, and didn't discriminate based on who they raped. Say that 0.1% of each population is a rapist, and commits a rape at random.

Then there are:

• ~50 million blacks * 0.1% = 50,000 black rapists
• ~200 million whites * 0.1% = 200,000 white rapists

The rapes occur at random, with 1/5 of our population black, 4/5 white, so there are:

• 50,000 * 1/5 = 10,000 black-on-black rapes
• 50,000 * 4/5 = 40,000 black-on-white rapes
• 200,000 * 1/5 = 40,000 white-on-black rapes
• 200,000 * 4/5 = 160,000 white-on-white rapes

So assuming blacks & whites are perfectly equal, the total number of interracial rapes is the same for blacks and whites. Therefore there is no need to correct for the population when looking at interracial rape (or other crime) statistics. The raw number is good. i.e. 64 to 1 overstates the case.