A prediction fulfilled!
But, jokes aside, this is becoming a significant waste of time.
After us both boasting about being genii, you just went back to the high-school level per-cap analysis that you used to start the conversation. You did not follow it with the 2nd, target-based, analysis which is often used - and so got a 10x disparity in B:W murders out of the tiny OG set of 246: 566 cases.
First, counselor Branca's original claim was a FIFTY to one disparity. This is just not real. And, you yourself just confirmed that the OVERALL BCR, even for violent Index felonies, is 2-3x the WCR in per capita terms and not anything like "12.4x" (2.2:1 per the BJS-NCVS; you used FBI data and got 3:1). That disparity, of course, is the one that exists before any adjustments for age, region, etc. These are pretty major mistakes, now on the record.
But, also, my point about target selection remains valid. In the integrated environment where INTER-racial crime is even possible, if 76% of potential crime victims are white or Hispanic Caucasians, and 12% are Black, you would expect an neutral criminal of any race to target WH 76% of the time and B 12% of the time. What you would predict ACROSS group lines is B=(.12 x 76) while WH = (.76 x 12).
You would not expect the smaller group with 6x as many targets to commit only 1/6 as many INTER-racial crimes as the larger group with 1/6 as many targets. Instead, expected crime numbers would be identical or almost identical: this model has run in dozens of papers and is not something I am "confused" about. This is why I noted that multiplying the starting 246:566 ratio BY 5-6, to reflect smaller Black criminal numbers, should be followed by dividing the sum by 5-6 to reflect larger white potential victim numbers.
May mute here to cook, at least temporarily. Keep seeking!